Health Services Research (HSR) Methods
Print Page

Full Glossary List

Absolute Risk Reduction (ARR) [EQ]
Compares the probability of an outcome in the control group with the treatment group.

An individual's ability to obtain appropriate health care services. Barriers to access can be financial (insufficient monetary resources), geographic (distance to providers), organizational (lack of available providers) and sociological (e.g., discrimination, language barriers). Efforts to improve access often focus on providing/improving health coverage.

The degree of conformity of a measure to a standard or true value.

Active Duty
In the context of legal research, a duty to take some definite action or engage in a continued course of action.
Synonyms: Positive Duty

Actual Authority
Authority intentionally granted to an agent or authority that an agent reasonably believes he or she has as a result of dealings with the principal.

Actual Damages
Money awarded to repay an actual loss or to compensate for a proven injury or loss.

Acute Care
Medical treatment rendered to individuals whose illnesses or health problems are of a short-term or episodic nature. Acute care facilities are those hospitals that mainly serve persons with short-term health problems.

Acute Disease
A disease that is characterized by a single episode of a relatively short duration from which the patient returns to his/her normal or previous level of activity. While acute diseases are frequently distinguished from chronic diseases, there is no standard definition or distinction. It is worth noting that an acute episode of a chronic disease (for example, an episode of diabetic coma in a patient with diabetes) is often treated as an acute disease.

Additional Damages
Damages provided by the statute in addition to direct damages, including expenses related to the injury, compensatory damages and damages.

Activities of Daily Living (ADL)
An index or scale which measures a patient's degree of independence in bathing, dressing, using the toilet, eating, and moving from one place to another.

Adverse Event
In a medical context, an injury resulting from a medical intervention.

Adverse Event Reporting Systems
The Adverse Event Reporting System (AERS) is a computerized information database designed to support the Food and Drug Administration's (FDA) post-marketing safety surveillance program for approved drug and therapeutic biologic products. Adverse drug reaction reports are submitted by manufactures, health care professionals, and consumers through the MedWatch program. The FDA uses this system to make regulatory actions to improve product safety and protect the public health, such as updating a product’s labeling information. Reports are compiled by the Food and Drug Administration’s Center for Drug Evaluation and Research as well as their Center for Biologics Evaluation and Research.

Advisory Opinion
An opinion issued by a court at the request of an interested party suggesting how the court would rule in a matter should a case or controversy arise. Advisory opinions are not binding.

A sworn declaration of fact. Generally an affidavit is used to present evidence in court.

A person authorized by another person (generally referred to as a "principle") to act on his or her behalf.

Agency for Healthcare Research and Quality (AHRQ)
AHRQ was created in December 1989 as the Agency for Health Care Policy and Research (AHCPR), a Public Health Service agency within the U.S. Department of Health and Human Services reporting to the Secretary. The agency was reauthorized December 1999, as the Agency for Healthcare Research and Quality. AHRQ's mission is to support research designed to improve the outcomes and quality of health care, reduce its costs, address patient safety and medical errors, and broaden access to effective services. The research sponsored, conducted, and disseminated by AHRQ provides information that helps people make better decisions about health care.
Synonyms: Agency for Health Care Policy and Research (AHCPR)

Allowable Costs
Items or elements of an institution's costs that are reimbursable under a payment formula. Both Medicare and Medicaid reimburse hospitals on the basis of only certain costs, usually based on medical necessity. Allowable costs may exclude, for example, luxury accommodations, costs that the insurer deems as unreasonable expenditures (or not substantiated in medical records as medically necessary), or that the insurer deems unnecessary for the efficient delivery of health services to persons covered under the program in question.

Alternative Hypothesis
The alternative hypothesis, H1, is a statement about a specific relationship between the parameters of interest. If the null-hypothesis is rejected, the alternative hypothesis is supported; however, the alternative hypothesis can never be proved.

Amicus Curiae
Latin for "friend of the court." Describes person not party to a lawsuit but with interests in the case.

Amicus Curiae Brief
A brief filed by those not party to a lawsuit but with interests in the case. The brief is generally filed with the intent of influencing the court.

Ancillary Services
Supplemental services, including laboratory, radiology, physical therapy, and inhalation therapy that are provided in conjunction with medical or hospital care.

Anonymous Reporting
An error reporting method used to protect the identity of those individuals who report medical errors so that their reports cannot be easily used in civil lawsuits against them. Under anonymous reporting, data that could identify the reporter are emitted from the report.

Antecedent [EX]
Preceding a result or outcome.

Apparent Authority
Authority that a third party reasonably believes an agent possesses, as based on the third party's dealings with the principle, even if the authority was not granted by the principal. Under theories of agency, apparent authority is often sufficient to bind the principal for the agent's actions.

A written request to a higher court to modify or reverse the judgment of a lower court (such as trial court or an intermediate appellate court).

Appeal de Novo
An appeal in which the appellate court uses the record of the lower court (trial court) but reviews the evidence without regard for the lower courts rulings.

Appellate Brief
A brief submitted to an appeals court.

Appellate Court
A higher court that reviews the decision of a lower court when an appeal is filed. Traditionally, new evidence may not be entered and new arguments may not be raised at an appellate court. The appellate court takes all facts presented at the original trial, and accepted by the original judge or jury, as true. The appellate court generally only rules on whether there was a mistake in the application or interpretation of law.

The party of a lawsuit against whom an appeal is taken. Often referred to as a "respondent" in a court of appeals, the status of an appellee does not correlate to a status as a defendant or plaintiff in the trial court.

Appropriate health care is care for which the expected health benefit exceeds the expected negative consequences by a wide enough margin to justify treatment.

Artifactual Association
An apparent but false association between two variables that results from chance occurrence or bias in a study's methods.

A term signifying a relationship between two or more events or variables. Events are said to be associated when they occur more frequently together than one would expect by chance. Association does not necessarily imply a causal relationship. Statistical significance testing enables a researcher to determine the likelihood of observing the sample relationship by chance if in fact no association exists in the population that was sampled. The terms "association" and "relationship" are often used interchangeably.

Attributable Risk [EQ]
A measure that explains the proportion of disease incidence (or other outcome) that can be attributed to a specific exposure. Attributable risk is calculated by subtracting the risk of disease in the non-exposed group from the risk in the exposed group.

Attrition Bias
A type of threat to internal validity that refers to the differential, systematic loss of study participants from the control and treatment groups.
Synonyms: Attrition

The right or permission to legally act on another's behalf.

Correlation between the error terms of observations. Autocorrelation can be cross-sectional, arising, for example, from observing multiple members of the same family; or time-series resulting, for example, from seasonal trends in the data. Tests for autocorrelation include the Durbin-Watson test, the Durbin M test, Godfrey-Breusch Lagrange Multiplier test, and the Box-Pierce test.

Autoregression [EQ]
A regression model that uses a prior, or lagged, value of the dependent variable as an independent variable. Common instances are pre-post intervention, adaptive expectation, geometric or distributed lag, and partial adjustment models. Inclusion of a lagged value of the dependent variable among the explanatory variables can result in biased coefficients.

Average Wholesale Price (AWP) of Prescription Drugs [EQ]
The average wholesale price of a drug relates to the price that wholesalers charge pharmacies, and is often used by pharmacists to price prescriptions. Drug manufacturers and labelers commonly publish suggested wholesale prices for their products. Price surveys of wholesalers are also available.

Bar Chart [FIG]
A chart made up of vertical or horizontal bars that represent the relative values of a variable.

Bayesian Approach
A statistical approach that uses Bayes' Theorem to integrate related data and a-priori belief with observed data to estimate unknown parameters. This represents an alternative to the traditional frequentist approach that attempts to establish confidence intervals around parameters, and/or falsify a-priori null-hypotheses.
Synonyms: Bayesian Analysis

Before-After Design [EX]; [FIG]
A study design in which the dependent variable (such as a clinical outcome) is measured both before and after an intervention in the same group of individuals. Comparison of the outcomes is made before and after the intervention to assess the effect of treatment.

The process of setting organizational goals and measures in comparison to those of other organizations or products. Most often, comparison is made to products, services, and practices recognized as leaders in the field or industry. The benchmarking process is a quality assurance effort designed to identify best practices that will lead to improvements in one's own products and services.

Between Group Design [EX]; [FIG]
A study design in which two or more groups subject to different experiences or treatments are compared. The purpose is to make statistical comparisons between two or more groups and demonstrate a causal relationship between the independent variable and outcome of interest.

Between-group designs use a separate sample of individuals for each treatment condition. By comparison, a within- subject design (such as a before-after design) uses the same individuals for each treatment condition.

Bias [EQ]
A systematic (consistent, non-random) error that results in over or under-estimation of a parameter. Bias reflects the degree to which the statistic inaccurately measures the parameter that it is intended to estimate.

Binding Precedent
A that the court must follow (e.g., the lower courts must follow the precedent of higher courts in the jurisdiction where they sit).

Binomial Distribution [EQ]
The probability distribution of a binomial random variable, which represents the number of successes in "n" independent trials, each of which has exactly two mutually exclusive outcomes. For each trial "p" represents the probability of success.
Synonyms: Bernoulli Distribution

Binomial Logistic Regression
A model for a binary outcome that represents the probability of an even in the form of the natural log of the odds, based on a linear function of independent variables.
Synonyms: Logit

Binomial Probit Regression
A model for a binary that represents the probability of an even in the form of the standard normal distribution, based on a linear function of independent variables.

Blinded Study [FIG]
Blinding conceals a study participant's group assignment (to either the treatment or control group) from either the participant and/or the investigators in a study. In a blinded study, participants do not know which medicine is being used in the study and are therefore able to describe their experiences without bias, and investigators are unable to influence the results of the study. Blinding is not always practical, but is highly desirable whenever possible in order to minimize study bias.
Synonyms: Single Blind Study; Single Blinded Study; Blinding

A statistical technique used to estimate the sampling distribution of an estimator or sample statistic, usually in order to calculate a robust standard error or confidence interval or when a closed form solution is not available.

Bootstrapping involves drawing many random sub-samples from the original data set and computing the statistic of interest from each in the same way. After sampling, each sub-sample is returned to the data set, a process known as sampling with replacement.

Box Plot [FIG]
A visual display of information used to identify patterns in quantitative data. Box plots divide the data into four quartiles. The quartiles are represented by a box that extends from the first quartile (Q1) to the third quartile (Q3) and contains a vertical line drawn at the second quartile (Q2 -- the median). Two lines called whiskers extend horizontally from the front and back of the box. The front whisker extends from Q1 to the smallest non-outlier in the data set, and the back whisker goes from Q3 to the largest non-outlier. Outliers are plotted as separate points on the chart.

A written statement describing the legal contentions of a particular party. This document serves as the basis for a party's case or controversy.

Fixed or durable non-labor inputs or factors used in the production of goods and services, the value of such factors, or the money specifically allocated for their acquisition or development. Capital costs include, for example, the buildings, beds, and equipment used in the provision of hospital services. Capital assets are usually thought of as permanent and durable as distinguished from consumables such as supplies.

Capital Expenditure Review
A review of proposed capital expenditures of hospitals and/or other health facilities to determine the need for, and appropriateness of, the proposed expenditures. The review is done by a designated regulatory agency and has a sanction attached that prevents or discourages unneeded expenditures.

Case Law
Legal precedents created by judges, not legislatures. Created through the interpretation of statutory and common law.

Case Mix
A measure of the mix of cases being treated by a particular health care provider that is intended to reflect the patients' different needs for resources. Case mix is generally established by estimating the relative frequency of various types of patients seen the provider in question during a given time period and may be measured by factors such as diagnosis, severity of illness, utilization of services, and provider characteristics.

Case Report Study
Study design in which data is collected as detailed, narrative information reported by the cases, rather than quantitative scores measured or observed through statistical sampling.

Case Reporters
A collection of the courts' published decisions and opinions. (e.g., Supreme Court Reporter, the Atlantic Reporter).

Case Severity
A measure of intensity or gravity of a given condition or diagnosis for a patient.

Case Study
A qualitative study design in which a single case or a group of cases is examined in an in-depth, systematic manner. The objective of case studies is to provide a detailed comprehensive narrative rather than providing a large sample of quantitative data.
Synonyms: One Shot Case Study; Simple Case Study

Categorical Missingness
A special problem of missing or incomplete data that involves variables measured with categorical rather than continuous values. Many approaches for estimating and imputing values for missing observations rely on numerical data and produce only continuous values. In instances where the problem of missing data pertains mostly to variables measured with categorical values, an alternative approach to dealing with missing data is necessary.

Relating causes to the effects they produce. Most of epidemiology concerns causality, and several types of causes can be distinguished. A cause is termed "necessary" when a particular variable must always precede an effect. This effect need not be the sole result of the one variable.

A cause is termed "sufficient" when a particular variable inevitably initiates or produces an effect. Any given cause may be necessary, sufficient, neither, or both.

Join a discussion on the definition and use of the term causality. 

Cause [EX]
A person or thing that is responsible for a result.

Cause of Action
A specific legal claim seeking relief from another. Includes medical malpractice, negligence, breach of contract, etc.

Centers for Disease Control and Prevention (CDC)
The Centers for Disease Control and Prevention, based in Atlanta, Georgia, is charged with protecting the nation's public health by providing direction in the prevention and control of communicable and other diseases and responding to public health emergencies. Within the U.S. Public Health Service, CDC is the agency that led efforts to prevent such diseases as malaria, polio, smallpox, toxic shock syndrome, Legionnaire's disease and, more recently, acquired immunodeficiency syndrome (AIDS), and tuberculosis. CDC's responsibilities evolve as the agency addresses contemporary threats to health, such as injury, environmental and occupational hazards, behavioral risks, and chronic diseases.

Central Limit Theorem [EQ]
For a random sample of size n taken from any distribution with mean (µ) and variance (σ2), the sample mean (x-bar) will be approximately normally distributed with mean (µ) and variance (σ2/n). In general, the larger the value of n, the better the approximation to the normal.

Certificate of Need (CON)
A certificate issued by a governmental body to an individual or organization proposing to construct or modify a health facility, acquire major new medical equipment, modify a health facility, or offer a new or different health service. Such issuance recognizes that a facility or service, when available, will meet the needs of those for whom it is intended. CON is intended to control expansion of facilities and services by preventing excessive or duplicative development of facilities and services.

Latin for "to be more fully informed."

Ceteris Paribus
Latin term that is literally translated as "with other things [being] the same." Generally used to mean "all other things being equal." In econometric modeling, ceteris paribus refers to a situation in which the effects of variable other then the variable of primary interest are held constant.

Stands for Code of Federal Regulations. The codification of the general and permanent rules published in the Federal Register by the executive departments and agencies of the Federal Government.

Chi-Squared Goodness of Fit Test [EQ]
A test to compare observed sample data with a hypothesized distribution.

Chronic Disease
A disease that has one or more of the following characteristics: is permanent; leaves residual disability even if well managed; is caused by nonreversible pathological alternation; requires special training of the patient for rehabilitation; or may be expected to require a long period of supervision, observation, or care. Data on disease states are often categorized by chronic or acute disease conditions.

Circuit Court
A court with jurisdiction over several counties, districts or states.

Civil Court
A court with jurisdiction over non-criminal cases.

Civil Law
The body of law that relates to disputes or rights between two parties, in contrast to criminal law.

Civil Liability
Liability under civil law instead of criminal law.

Records of patients' receipt of health care services and payment for those services. Such data are typically used for administrative purposes, such as utilization review, billing, or reimbursement.
Synonyms: Claims Analysis

Claims Data
Administrative databases maintained by health plans and other third-party payers of health care. These data typically include person-level information, including demographic characteristics, as well as details of the health care services each patient received, such as procedures performed, medical diagnoses, and provider charges. For research purposes, claims data are a fairly inexpensive way to track large number of patients over time, but because the data are collected for purposes other than research, they are likely to lack certain elements of particular interest.

Class Action
A lawsuit involving numerous parties with similar legal claims. These various parties join together in a group/class in order to sue someone such as a large company or organization.

Clinical Guidelines
A set of steps or activities that have been systematically developed to assist practitioner and patient decisions about health care. Guidelines typically describe the activities that can be undertaken in response to one or more clinical conditions.

Clinical Performance Measures
Instruments that estimate the extent to which a health care provider: delivers clinical services that are appropriate for each patient's condition; provides them safely, competently, and in an appropriate time frame; and achieves desired outcomes in terms of those aspects of patient health and patient satisfaction that can be affected by clinical services.

Cluster Sampling [EX]; [FIG]
A sampling strategy where the population of interest is divided into representative "clusters" of individuals, among whom a random selection of subjects is drawn. Cluster sampling is often conducted when it is impossible or impractical to draw a simple random sample or stratified sample because the researcher cannot get a complete list of members of the population.

Centers for Medicare and Medicaid Services (CMS)
The government agency within the Department of Health and Human Services that directs the Medicare and Medicaid programs (Titles XVIII and XIX of the Social Security Act) and conducts research to support those programs. Formerly the Health Care Financing Administration (HCFA).
Synonyms: Health Care Financing Administration (HCFA)

A document describing the location of variables in an electronic file as well as the numerical codes representing attributes/response options for each variable.

Codifying Statute
A restatement or classification of the law on a particular topic, including case law and legislative provisions (i.e. common and statutory law) of the United States, a single state, or a local jurisdiction.

The process of assigning numbers to response categories in a questionnaire, where each category represents a unique attribute or response.

Coefficient Alpha Reliability
Measures the correlation between responses to two or more items relating to the same concept.

Cognitive Map
A cognitive map is the development of knowledge, decisions and interpretations in the human mind. In qualitative research, Cognitive Mapping refers to a type of interview in which researchers ask a participant a question and then ask follow-up questions to understand the way the participant interprets the questions and develops their answers.

Group of persons (or other organizational units) that experience a significant event or have a specific characteristic in common (i.e., birth year).

Cohort Study [FIG]
In a cohort study, individuals exposed and not exposed (to suspected risk factors) are followed and compared to assess the extent to which each group experiences an outcome of interest -- often illness or death.
Synonyms: Cluster Randomization; Panel Data Study; Longitudinal Study

Colddeck Imputation
An imputation approach in which a constant value from an external source is substituted for all missing values in a dataset.

The number of ways in which objects drawn can be drawn from a group.

Common Law
Law that is derived from judicial decisions instead of from statutes.

1) A group of people who have common characteristics; communities can be defined by location, race, ethnicity, age, occupation, interest in particular problems or outcomes, or other common bonds (Turnock 2004).
2) Individuals with a shared affinity, and perhaps geography, who organize around an issue, with collective discussion, decision making, and action (Labonte 1988).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

The Community Tracking Study (CTS)
The Community Tracking Study (CTS) is a longitudinal study which tracks the perceptions of the quality of healthcare in a cohort of American communities. The CTS was established in 1996 with sponsorship by the Robert Wood Johnson Foundation (RWJF) and conducted by the Center for Studying Health System Change (HSC). The CTS is a collaboration of three studies within a cohort of randomly selected communities. The Household Survey, Physician Survey, and Employer survey are each conducted separately under contract by private organizations, but employ the same method of data collection, through biennial site-visits and in-person interviews of the same community.

Comparative Effectiveness
A comparison of the effectiveness, safety or outcomes of different options of diagnosing or treating a given medical condition for a particular set of patients.

Comparative Effectiveness Research
Research studies that compare one or more diagnostic or treatment options to evaluate effectiveness, safety or outcomes. Comparative effectiveness research includes primary and secondary research.

Comparative Effectiveness Review (CER)
A systematic review of the evidence from multiple clinical studies about the comparative effectiveness of different diagnostic or treatment options. A CER may synthesize direct comparator (head-to-head) studies as well as studies where the treatments were not compared directly with one another. In the latter case, treatments may be compared to a placebo. A CER may also non-comparative studies that describe in detail the impact of a treatment on a particular group of patients or a particular outcome. Systematic reviews are especially suitable for critiquing and summarizing a body of evidence relevant to focused questions about diagnostic, prognostic, or therapeutic clinical practices.
Synonyms: Secondary Clinical Effectiveness Research

Comparison Group
A group of individuals (sometimes called a control group) that are selected to be as similar as possible to the experimental group. Comparison groups can be created by random assignment of treatment, and in some senses this is the best kind.

Comparison Group Posttest Only Design [FIG]
A study design in which two groups, a study group and a control group, are measured at one point in time following an intervention or experiment. The study group experiences an intervention or experiment while the control group does not.
Synonyms: Static Group Comparison

Compensatory Damages
Money awarded sufficient to compensate an individual for a loss suffered and nothing more than the loss. In other words, compensatory damages do not include punitive or other types of damages.

Compiled Statute
Laws that have been arranged by subject matter but have not been substantively changed.

Complement [FIG]
The complement of an event is the event not occurring.

Complete Case Analysis
Analysis based on a set of cases that do not have any missing values for the variables of interest.

Computer Assisted Telephone Interview (CATI)
A computer-assisted telephone interview protocol in which interviewers use a computer to display survey questions and record individual responses.

Computerized Physician Order Entry (CPOE)
Electronic systems in which physicians enter and transmit medication orders as well as orders for radiology, lab work, and other ancillary services. Physician order entry systems help catch and prevent errors by checking physician orders against potential drug to drug interactions, normal dosages, and diagnostic or therapeutic guidelines. Physician order entry systems also prevent medical errors due to misreading of hand-written orders.

Conceptual Framework
A comprehensive framework that is generally used to guide formulation of a research question or hypotheses. Conceptual frameworks integrate the previous literature, theory and methodological approaches that characterize a particular discipline.

Concurrent Cohort Study [FIG]
A study design in which a population at risk is examined to identify diseased as well as non-diseased individuals. These sub-groups are categorized as exposed or not exposed to the suspected factor and then followed for a period to identify who develops the disease and who does not within that period.

The purpose is to determine the association between a risk factor and the disease. This type of cohort study may be described as both longitudinal and prospective. Some special cases of cohort studies are retrospective rather than concurrent.

Concurrent Non-Randomized Control
An experimental study design in which patients are not assigned to control and treatment groups in a randomized fashion, but in some other way. This may be done for investigative convenience or based on a preset sampling plan that is not random.

Concurrent Validity
A form of criterion validity analysis that measures the extent to which a measure/question corresponds to the criterion measure ("gold standard") at the same point in time.

Conditional Logitistic Regression
A model used for a nominal outcome with two or more values where the determinants of the choice probabilities are based on characteristics of the choices.
Synonyms: Conditional Logistic Regression; McFaddens Logit

The assurance given to research participants that the information disclosed as part of the study is only accessible to authorized entities.

A condition or state that is associated with both an explanatory variable and the dependent variable but is itself not caused by the explanatory variable, e.g. not a consequence of treatment.
Synonyms: Confounding Variable

Confounding [FIG]
Confounding occurs when there is a relationship between an exposure, the outcome of interest,  and a third factor called a "counfounder" or a "confounding variable".  Due to the presence of the confounder, it is not possible to accurately assess the relationship between the exposure and outcome of interest.  A counfounder must: 
1) be related to the outcome and the exposure;
2) have a different distribution between exposure and non-exposure.   

Confounding may be accounted for by randomizing study participants, matching participants by likely confounders (such as age and sex), or controlling the effect of the confounder by stratifying the analysis.

Conjoint Analysis
A method used in marketing research to determine the ideal or most preferred configuration of a new product or service. In a conjoint analysis, respondents are asked to rank different combinations of product attributes. A computer program is then used to find the value (utility) associated with each attribute and orders the respondent's preferences accordingly.

Computerized Needs-oriented Quality Measurement Evaluation System (CONQUEST)
CONQUEST was developed by the Agency for Healthcare Research and Quality (AHRQ) as a tool that permits users to collect and evaluate health care quality measures to find those suited to or adaptable to their needs. CONQUEST has interlocking databases describing "Measures" and clinical "Conditions."

The Measure Database contains information on clinical performance measurements to assess the quality of the health care delivered by providers. The Condition database contains information on incidence, prevalence, cost and utilization, co-morbidities, risk factors, treatments, and guidelines. These databases link by codes for clinical services and health outcomes related to specific measures and conditions.

Consolidated Appeal
A statute that collects the legislative provisions on a particular subject and embodies them in a single statue with minor amendments and drafting improvements.

Construct Validity
The extent to which relationships between measures/items capture the hypothesized relationships predicted by theory.

The philosophy that learning is a process of constructing new ideas based on past experiences and events. In other words, new learning is constructed based on what individuals already know. In qualitative research, constructivist observations are made to observe the way different stakeholders in a study develop their beliefs in a specific social setting.

Content Analysis
A set of procedures for collecting and organizing non-structured information. This approach makes it easier for researchers to systematically and objectively analyze the data and make inferences about the population of interest.

Content Validity
The extent to which measures/items adequately represent an underlying concept or construct.

Contingent Valuation
A survey elicitation method used to elicit willingness to pay for specified mortality or morbidity reductions, or willingness to accept compensation for specified mortality or morbidity increases.

Contractual Duty
A duty arising to accordance with a contract.

Indication of a specific circumstance in which use of an otherwise appropriate intervention could be harmful to an individual.

Individuals who do not receive an intervention.

In case control studies, it is ideal to select a random sample of the entire population; however this is often not possible. An alternative is to select study participants based on whether or not they experience an outcome of interest (cases) and compare them to a group of otherwise similar patients who do not have the outcome of interest (controls).
Synonyms: Controls; Control Group

Control Group
In randomized experimental research, the control group is a group of patients that receives no treatment. Outcomes for this group may then be compared to the "active" or "standard" treatment group in order to assess the impact of treatment. In order for the comparison to be valid, both the control and treatment groups should resemble each other as much as possible.

Control Variable
A condition or state not experimentally manipulated by the scientist that is associated with the outcome, but unassociated with treatment, which is included in the regression to improve overall model fit to the data and precision of a treatment effect estimate.

Convenience Sampling
Generates a sample from available subjects. Convenience sampling does not result in a sample that is representative of the entire population of interest.
Synonyms: Grab-bag Sampling; Haphazard Sampling; Volunteer Sampling; Judgemental Sampling

Convergent Validity
A form of construct validity concept that measures whether a hypothesized association between a measure/item and an alternative measure of the same concept is confirmed.

A statistic that indicates the association between two variables. Correlation may be positive, zero, or negative. Positive correlation reflects a relationship where, as one variable increases, the other also increases. Negative correlation reflects a relationship where, as one variable increases, the other decreases. The correlation between two variables is zero if no systematic relationship exists between the two variables.

Correlation Coefficient [FIG]
Measure of the degree to which two variables are associated, based on the degree of their linear relationship. Correlation measures range between -1.0 and +1.0, where each extreme represents a perfect linear relationship and 0 represents no relationship.

A correlation coefficient equal to 0 means there is no linear relationship between the variables.

Correlational Studies
Surveys conducted to define bi-variate relationships between factors of interest.

Cost Benefit Analysis (CBA)
An analysis of the net costs of implementing an intervention, and the net benefits that arise from the intervention. In CBA, both costs and benefits are expressed as monetary units. The result of the analysis is the difference between the net benefits and net costs, or net benefit, expressed in monetary units.

Cost Consequence Analysis (CCA)
A form of cost effectiveness analysis comparing alternative interventions or programs in which the components of incremental costs (e.g., additional therapies, hospitalization) and consequences (e.g., health outcomes, adverse effects) are computed and listed, without aggregating these results (e.g., into a cost-effectiveness ratio).

Cost-Effectiveness Acceptability Curve
A graph showing the probability that an intervention is incrementally cost-effective, as a function of the cost-effectiveness threshold. Cost-effectiveness acceptability curves are used as a summary result for probabilistic sensitivity analyses, in which the parameters of cost-effectiveness analysis are varied according to specified probability distributions.

Cost Effectiveness Analysis (CEA)
An analysis of the incremental costs and incremental health consequences of an intervention compared to an alternative. In CEA, the effects or consequences (generally health effects) of all programs being compared are measured in common units, such as premature deaths averted, years of life gained, or quality-adjusted life years (QALYs) gained. The result of the analysis is a ratio of incremental costs to the incremental effect of an intervention (e.g., cost per QALY).

Cost-Effectiveness Ratio
The difference in the cost of using two treatments relative to gaining a unit increase in the effect (e.g. dollars spent per year of life gained).

Cost-Effectiveness Threshold
A value of cost per year of life gained or cost per quality-adjusted life year (QALY) gained, below which an incremental cost-effectiveness ratio is considered acceptable.

Cost Minimization Analysis (CMA)
When the effectiveness and safety of alternative interventions are considered to be equivalent, CMA determines the least costly alternative among a set of interventions.

Cost of Illness Analysis (COI)
Study that provides a monetary estimate of the total economic impact (including direct and indirect costs) of a particular disease or condition. Generally, one of two approaches is taken: the prevalence-based approach, which considers all expenditures attributable to the disease or condition in a specific period of time; and the incidence-based approach, which considers the lifetime cost of those who develop the condition during a specific period.

Cost Utility Analysis
A special type of cost-effectiveness analysis, in which the measure of health consequences is quality-adjusted life expectancy. QALYs can be used to determine the relative value of health care alternatives because they are applicable to a wide range of health conditions and interventions.

A conceptual notion of what would have happened to a group of study participants who receive an experimental treatment if they had not received the treatment. Since this group of study participants cannot simultaneously lead two different lives (one in which they receive the treatment and one in which they do not), the counterfactual is a hypothetical condition that can never actually happen. Nevertheless, it is a useful way of thinking about the causal effects of treatment because it involves comparing, in theory, two groups of participants who are alike in every way except for whether or not they received the treatment.

Covered Entity
Refers to three types of entities that must comply with Federal health information privacy regulations (i.e., HIPAA Privacy Rule): health care providers, health plans, and health care clearinghouses. For these purposes, health care providers include hospitals, physicians, and other caregivers, as well as researchers, who provide health care and receive, access, or generate individually identifiable health care information.

Criterion Validity
The extent to which a measure/item predicts or agrees with the "true" value (or "gold standard") for the measure.

Cross Sectional Study [FIG]
Studies that conduct measurements on a group of subjects at one point in time. Cross-sectional studies look at both exposure and outcomes at one point in time and are designed to evaluate associations between risk factors and outcomes in a specific population.

Crossover Design
A study designed to compare two or more treatments by ensuring that study participants receive all treatment options.

This is typically done by randomly assigning each participant to a treatment and then, upon completion of the initial treatment, switching them to the other after a washout period.

Each participant's outcome in one period is compared with his/her outcome in another period so each individual serves as his or her own control.

Crossover Trial
A study design in which each participant serves as their own control over a two-period study. In the first period of the study, participants receive either the treatment or control. Then, after a "washout" period to minimize the effect of the first period, the participant switches to receiving the control or treatment, depending on which they received in the first period. Randomization is used to assign the order in which the treatment and control conditions are administered.

A phenomenon whereby new public programs or expansions of existing public programs designed to extend coverage to the uninsured prompt some employers to no longer provide coverage or privately insured persons to drop their private coverage and take advantage of the expanded public subsidy.

Crude Rate
A proportion representing the number of events in a given population at a given time. Often measured as events per 1,000 population or per 100,000 population.

Called a "crude" rate because the statistic has not been adjusted to account for related circumstances or data.

Cumulative Incidence
Calculation of the number of new cases (incidence) of an outcome divided by the number of individuals in the population considered to be at risk for the outcome.

The Current Population Survey
The Current Population Survey (CPS) was established in 1940 as a household survey used to determine the national employment rate and earnings for different population groups. The March supplement of the CPS provides information about health insurance information and income. The CPS is conducted monthly through cooperation between the Bureau of Labor Statistics and the U.S. Census Bureau. Data is collected through telephone and in-person interviews of non-institutionalized persons 16 years and over.

Current Procedural Terminology, Fourth Edition (CPT-4)
A manual that assigns five digit codes to medical services and procedures to standardize claims processing and data analysis. 

Money awarded to a person as compensation for loss or injury; money ordered to be paid.

Data Cleaning
Process of checking for and correcting errors in a data set.

Data Use Agreement (DUA)
A Data Use Agreement (DUA):
1) The requirements to obtain approval for a research project. A DUA is the permission and disclosure of a "limited data set" given by the researchers of such data.

2) Upon request for information contained in a "limited data set" from a covered entity, a DUA must be obtained in order to disclose or use the data for research, public health, or health care operations purposes.

De Facto
Latin for "in point of fact." Actual; existing in fact.

A process whereby information that could identify the clinician, the reporter, the health care institution, or another organization involved in a medical error are removed from an error report after it is received. This process is used to maintain records of factors that could cause errors, it assure those who report errors that their reports will not be used in civil lawsuits against them.

De Jure
Latin for "as a matter of law." Existing by right or according to the law.

De Novo
Latin for "anew."

Quantiles that divide the distribution into ten equal, consecutive groups.

Decision Analysis
A quantitative approach to modeling decision-making under uncertainty. These models typically take the form of a decision tree or an influence diagram that illustrate the multiple pathways or strategies for a particular outcome. Estimates of the probability of particular events and outcomes are drawn from existing literature or experts.

Deductive Process [FIG]
The process of using theory to guide research drawing inferences regarding specific applications from general principles or phenomena.

Degrees of Freedom
The number of data values (scores) that are free to vary when calculating a statistic. That is, the total number of observations (n) minus the number of restrictions placed on the data. In many cases, the degrees of freedom are equal to the number of observations minus one.

Delphi Panel Method
A structured process designed to elicit information from a group of experts, often with the goal of achieving consensus. Individuals are asked their opinions on an issue, and the summary results of that process are then given to the group without identifying the source of the comments. Individuals are then asked to revise their opinions in light of the group results, with or without discussion. This process is then repeated until consensus has been reached.
Synonyms: Delphi Technique; Delphi Study

For a rate or proportion, the number of individuals, or person-years of experience, at risk for the outcome counted in the numerator. For a performance measure based on the proportion, the sample of cases that will be observed (e.g., the number of patients discharged alive with a confirmed diagnosis of acute myocardial infarction, excluding patients with bleeding, or other specified conditions).

Dependent Variable
The outcome or criterion of interest. In a causal model the cause typically is an explanatory variable and the effect is the dependent variable. Dependent variables often are denoted as Y in regression models.

Descriptive Statistics
Used to describe data. Includes summarizing via statistical or qualitative means, organizing, graphing.
Synonyms: Exploratory Data Analysis

A determinant influences the nature of a result.

Non-random. Refers to relationships among variables in which chance or randomness does not contribute to an outcome.
Synonyms: Mathematical Identity

Dichotomous Variable
A variable with two levels or categories (i.e., yes and no; male and female).

Difference in Differences Approach
An estimation approach in which changes in the dependent variable from the pre-intervention period to the post-intervention period for subjects in a treatment group are compared to changes over a similar period for subjects in a control group.

Diffusion Bias
A type of threat to internal validity that results from the unintended exposure of the control group to the intervention (or a similar intervention) intended for the treatment group.
Synonyms: Imitation of Treatment

One aspect of a concept. Each dimension represents a single aspect of a concept.

Direct Costs
The value of resources consumed in providing the health care service or intervention in question. Includes physical goods, as well as labor whether paid or unpaid. Includes resources provided by patients, such as transportation, as well as resources provided by health-care providers. Also includes costs associated with the consequences of the intervention, such as adverse effects, additional tests and treatments, and future clinical events.

The release of information from one entity to another.

Discourse Analysis
A method in which researchers analyze conversations in order to gain insight in participant interaction rather than the actual content of the narrative. This then allows researchers to develop theories concerning the rules and structure of communication. Ethnographers performing discourse analysis not only focus on the conversation but also the cultural context of the conversation.
Synonyms: Conversation Analysis

Discretionary Damages
Damages that are not directly quantifiable but are subjectively determined by a jury for suffering such as mental anguish or pain. Also called indeterminate damages.

Discriminant Validity
A form of construct validity analysis that measures whether hypothesized association between a measure/question and an alternative measure of a different concept is confirmed.

Dissenting Opinion
An opinion by one or more judges disagreeing with the majority opinion.

District Court
The name of the main trial court (in federal court and in most, but not all, states). States may also group their appellate courts into districts.

Document Review
A technique of content analysis involving the examination of existing records or documents.

In cost-effectiveness analysis, a case in which an alternative or treatment can be ruled out because it would not be cost-effective at any cost-effectiveness threshold. Strong dominance applies when the alternative is both more effective and less costly than at least one other alternative. Weak dominance (also known as extended dominance) applies when an alternative is less costly than another, but has a higher incremental cost-effectiveness ratio.

Dot Plot
A visual display of information in which dots are used to display frequency counts of data within groups or categories.

Double-Blind Study
A study in which neither subjects nor investigators know whether the subject is receiving an active treatment or a placebo (or control).

Dummy Variable
A dummy variable is a binary form of a categorical or non-numeric variable. By transforming the initial variables into dummy variables (typically with values equal to 0 or 1) the effects of each of the levels of the original variable may be included in a regression model.

A legal obligation owed to another.

Duty of Good Faith
A duty that requires two parties to deal with each other fairly.

Ecologic Fallacy
Refers to the fact that group level patterns do not take into account variability between individuals. As a result, it may not be appropriate to draw conclusions about individuals based on group or population-level characteristics.
Synonyms: Simpson's Paradox

Ecologic Studies
Studies in which data is collected on group characteristics only and do not include individual exposure and outcome data.

Ecological Validity
Closely related to "external validity". Refers to the ability to generalize research findings obtained in a laboratory setting to the real world.

Effective Health Care Program
An AHRQ program focused on the comparative effectiveness of different treatments and clinical practices. The Effective Health Care Program has three main approaches to research on the comparative effectiveness:
1) Review and synthesis of published and unpublished scientific evidence.
2) Promotion and generation of new scientific evidence and analytic tools in an accelerated and practical format.
3) Compilation of findings and translation of findings into a variety of useful formats for stakeholders.

The effect of a treatment in a "real" clinical practice setting. Effects or outcomes may be measured in terms of quality-adjusted life years, cases of disease (symptoms) prevented, or years of life saved.

The effect of a treatment in a controlled setting such as a clinical trial. In studies to assess efficacy, all conditions are controlled to maximize the effect of the treatment.

The construct or cognitive map of a social setting as defined by members of the group.

Empirical Data
Information based on direct evidence, experiences, or observations rather than on reasoning, theory, or arguments.

A disease that occurs regularly or continually in a specific geographic area or population.

Endogeneity [EQ]
A source of bias in a multiple regression model that results from correlation between one or more of the independent variables and the error term. This bias may result, for example, from an omitted variable, reverse causality, measurement error, or in some cases, a lagged value of the dependent variable used as an explanatory variable.

The occurrence of illness within a specific geographic area or population that exceeds normal expectancy.

The study of the patterns of determinants and antecedents of disease in human populations. Epidemiology utilizes biology, clinical medicine, and statistics in an effort to understand the etiology (causes) of illness and/or disease. The ultimate goal of the epidemiologist is not merely to identify underlying causes of a disease but to apply findings to disease prevention and health promotion.

Theory of human knowledge of reality. How knowledge is acquired and limitations to the validity and reliability of knowledge.

Error Term [EQ]
An expression of the unmeasured independent variables that are related to the outcome.
Synonyms: Disturbance Term

Errors in Variables
A situation in which one of more of the independent or dependent variables has been observed with error. This may result from poor or imprecise measurement techniques or from the use of a proxy variable in place of the true variable that belongs in the model. Even purely random measurement errors in the explanatory variables result in coefficients on the explanatory variables that are biased towards zero.

A mathematical function used to estimate the value of an unknown population parameter from a sample of data. If the expected value of the estimator is equal to the population parameter, the estimator is said to be unbiased. When the estimator has the smallest sampling variance among all possible estimators, it is said to be efficient. An estimator expressed as a single value is known as a point estimate or effect size; when the estimate is a range of values, it is known as an interval estimate, or confidence interval.

The study of a culture through observation to understand group member's emic, or shared sense of reality. Ethnography includes three steps:
1) Observation
2) Examination
3) Discussion and interpretation

The construct or cognitive map of a social setting based on the perspective of outsiders (researchers) performing a study. In essence, an outsider's view of a culture.

Cause. Used by epidemiologists

Evidence-based Decision Making
In a health policy context, evidence-based decision making is the application of the best available scientific evidence to policy decisions about specific medical treatments or changes in the delivery system. The goals of evidence-based decision making are to improve the quality of care, crease the efficiency of care delivery, and prove the allocation of health care resources.

Evidence-Based Medicine
Use of evidence from scientific research to make decisions regarding individual, patient care. Combines clinical expertise and experience, patient preferences, and review of clinical research.

Expected Value [EQ]
The mean of a random variable. The expected value of X is noted as E(X).

Any planned process that results in the collection of data. In many cases, experiments are those situations in which the researcher has control over some of the conditions under which the experiment is conducted.

Experimental Data
Numerical information produced by controlled experiments. Researchers collecting experimental data are typically able to modify all factors related to the association being studied, including randomly assigning subjects to the treatment and control groups. Experimental data are not commonly used in the social sciences because they are costly to obtain and often raise moral and ethical concerns.

Experimental Group
Individuals or units of analysis that receive a treatment, intervention, or program.
Synonyms: Treatment; Treatment Group

Experimental Quantitative Research
Studies that make use of both experimental approaches and statistical analyses of quantitative data. This includes comparison of experimental and control groups, and formal, systematic measurement of quantities, with the aim of determining the relationship between variables.

Experimental Study
A set of observations, conducted under controlled circumstances, in which the scientist manipulates the values of a variable to ascertain the effect of that manipulation on the outcomes of interest.
Synonyms: Randomized Study; Randomized Trial

Explanatory Model [EX]
A research technique designed to elicit causes and effects of social and empirical phenomena, which is then used to offer predictions on how a response might change or vary relative to other factors.

Explanatory Variable
In the context of regression equations, a variable that appears on the right hand side of a regression equation. The term "explanatory variable" is more inclusive than the term "independent variable," though the two often are used interchangeably.

Note: There is a literature in philosophy regarding the requirements for a legitimate "explanation."

External Validity
An aspect of research validity. Specifically, external validity assesses the extent to which the entire study is generalizable to the population of interest.

Face Validity
An aspect of measurement validity. Face validity assesses the extent to which experts or researchers believe that the instrument appropriately measures the construct of interest. Also used to describe how respondents may view the questionnaire.

A qualitative independent variable whose levels are set by the researcher. For example, a study with three treatment "arms" -- where participants are randomized across the study arms, would be a three factor study design.
Synonyms: Arms; Study Arms

Factor Analysis
A statistical technique that summarizes multivariate data by condensing it to a smaller set of underlying, related dimensions, or factors.

Factor analysis relies on correlation analysis to assess the degree of similarity among measures. This technique is often used to develop ratio scales and questionnaires in order to reduce the number of items and avoid duplication of concepts.

Factorial Design [FIG]
A complex study design used to test treatments (e.g. two independent variables). Must assume that the two treatments are different and independent of one another.

Favorable Selection
A tendency for utilization of health services in a population group to be lower than expected or estimated.

Federal Poverty Level (FPL)
The amount of income determined by the federal Department of Health and Human Services to provide a bare minimum for food, clothing, transportation, shelter, and other necessities. FPL is reported annually and varies according to family size (e.g., for a family of three in 1999, the FPL was $13,880, or $1,157 per month). Public assistance programs usually define income limits in relation to FPL.

Federal Register
The daily publication for rules, proposed rules, and notices of executive departments and federal agencies as well as executive orders.

A person who has a duty to act for the benefit of another individual within the confines of their relationship. For example, a fiduciary is typically someone legally obligated to act with good faith, loyalty, and care on behalf of another.

Fiduciary Duty
A duty of good faith, trust, confidence and candor owed by a fiduciary to beneficiary.

Fixed Costs
Costs that do not vary with the volume of output that is produced. Generally applicable only in the short term, because firms may vary all costs in the long run with sufficient time to modify production or output processes.

Focus Group
Qualitative study design in which formal, structured discussions are conducted with non-randomly selected groups of individuals who are brought together to talk about their attitudes, experiences, and, perspectives on a topic. Focus groups help researchers obtain qualitative data to inform subsequent quantitative phases of a research project.

The unit of analysis in focus groups is the entire focus group rather than individual participants because of the emphasis on group dynamics.

Follow-up [FIG]
The process of tracking and observing a group of subjects over time in order to document changes and monitor the effect of earlier activities or interventions.

May also refer to the amount of time an individual participates in a study.

Using econometric models to predict future changes in the value of a dependent variable. In econometrics, forecasting usually refers to theoretical models estimated for the sole purpose of making accurate predictions of the dependent variable, rather than estimating causal relationships. For example, an economic forecasting model may use only current and past values of the dependent variable to generate future estimates, omitting theoretically relevant explanatory variables.

Formative Research
A technique typically used in evaluation research to investigate processes and program operations in order to make refinements or improvements. 

Frequentist Approach [EQ]
A statistical approach focusing on drawing a sample from the population, computing parameter estimates, constructing confidence intervals, and testing hypotheses.

The name "frequentist" comes from the central assumption that the probability of an event is the same as the relative frequency of occurrence when a same experiment is repeated many times.
Synonyms: Frequentist Analysis

The extent to which the findings of a particular study can be extrapolated to other populations, programs, or organizations.

The study of genomes, which includes gene mapping, gene sequencing, and gene function.

Gibbs Sampling
A statistically-efficient method for drawing samples from a complex joint density function used in Markov Chain Monte Carlo (MCMC) estimation.

Global Budgeting
A method of cost containment in which participating hospitals must share a prospectively set budget. Methods for allocating funds among hospitals may vary but the key is that the participating hospitals agree to an aggregate cap on revenues that they will receive each year. Global budgeting may also be mandated under a universal health insurance system.

Granger Causality
In econometrics, Granger causality is present when past values of an explanatory variable (X) are helpful in predicting the current values of a dependent variable (Y) after controlling for past values of Y, but future values of X are not.

Gray Literature
Literature or research reports that are not peer reviewed (non-indexed). Examples include government agency documents, unpublished company or organization reports.
Synonyms: Grey Literature

Grounded Theory
A systematic method used to develop theory from data. Concepts are acquired based on repeated tests in order to make classifications which are then refined to generate a concrete theory.

Growth-Curve Models
A model of development in which individuals are assumed to all follow the same typical pattern of development.

Growth-Mixture Models [FIG]
An extension of growth-curve modeling which assumes that individual developmental patterns may follow any number of discrete growth trajectories (as opposed to traditional growth-curve modeling, in which individuals are assumed to all follow the same typical pattern of development).

Growth-mixture models identify unobserved sub-samples, or classes, of individuals, who follow distinct trajectories of development. Since each sub-sample is likely to be more homogenous than the sample as a whole, the factors associated with membership in each class can also be estimated.

Hausman Test for Endogeniety [EQ]
A statistical test to check for correlation between a suspected endogenous explanatory variable and the error term. The test works by comparing the estimates obtained from ordinary least squares (OLS) and two-stage least squares (2SLS) regression equations. The test requires at least one variable that is causally related to the suspected endogenous explanatory variable, but otherwise unrelated to the dependent variable.

Hawthorne Effect
A positive chance in the outcome variable in a study that may result from the participants' knowledge that they are being observed. A type of threat to external validity.

Healthcare Cost and Utilization Project Quality Indicators (HCUP QIs)
HCUP QIs comprise a set of 33 clinical performance measures that inform hospitals' self-assessments of inpatient quality of care as well as State and community assessments of access to primary care. Developed by the Agency for Healthcare Research and Quality (AHRQ) as a quick and easy-to-use screening tool, HCUP QIs are intended as a starting point in identifying clinical areas appropriate for further, more in-depth study and analysis.

HCUP QIs span three dimensions of care:
(1) potentially avoidable adverse hospital outcomes
(2) potentially inappropriate utilization of hospital procedures
(3) potentially avoidable hospital admissions.

1) The state of complete physical, mental, and social well-being and not merely the absence of disease or infirmity (WHO 1994).

2) The extent to which an individual or group is able to realize aspirations and satisfy needs, and to change or cope with the environment. Health is a resource for everyday life, not the object of living; it is a positive concept, emphasizing social and personal resources as well as physical capabilities (WHO 1984).

3) A state characterized by anatomical, physiological, and psychological integrity; an ability to perform personally valued family, work, and community roles; an ability to deal with physical, biological, and psychological stress; a feeling of well-being; and freedom from the risk of disease and untimely death (Stokes, Noren, and Shindell 1982).

4) A state of equilibrium between humans and the physical, biological and social environment, compatible with full functional activity (Last 1997).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

The Health and Retirement Study (HRS)
The Health and Retirement Study (HRS) is conducted biennially through telephone surveys by the University of Michigan to collect information on the health, employment, family and retirement planning among a nationally representative sample of adults 50 and over. The HRS was first conducted in 1992 and currently includes a sample of older adults from the Assets and Health Dynamics Among the Oldest Old (AHEAD) study.

Health Inequality
1) A generic term designating differences, variations, and disparities in the health of individuals and groups (Kawachi, Subramanian, and Almeida-Filho 2002).
2) Those inequalities in health deemed to be unfair or to stem from some form of injustice. The dimensions of being avoidable or unnecessary have often been added to this concept (Kawachi, Subramanian, and Almeida-Filho 2002).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Health Insurance
Financial protection against the health care costs arising from disease or accidental bodily injury. Such insurance usually covers all or part of the costs of treating the disease or injury. Insurance may be obtained on either an individual or a group basis. Although the term is often used by policymakers to refer to comprehensive coverage, insurers and regulators use it also to refer to other forms of coverage such as long term care insurance, supplemental insurance, specified disease policies, and accidental death and dismemberment insurance.

Health Insurance Flexibility and Accountability (HIFA)
The primary goal of the HIFA demonstration initiative is to encourage new comprehensive State approaches that will increase the number of individuals with health insurance coverage within current-level Medicaid and State Children's Health Insurance Program (SCHIP) resources. The program utilizes CMS Section 1115 waiver authority and emphasizes broad statewide approaches that maximize private health insurance coverage options and target Medicaid and SCHIP resources to populations with incomes below 200 percent of the Federal poverty level (FPL).

Health Manpower Shortage Area (HMSA)
An area or group which the U.S. Department of Health and Human Services designates as having an inadequate supply of health care providers.

HMSAs can include:
(1) an urban or rural geographic area
(2) a population group for which access barriers can be demonstrated to prevent members of the group from using local providers
(3) medium and maximum-security correctional institutions and public or non-profit private residential facilities

Health-Related Quality of Life
The impact of the health aspects of an individual’s life on his or her quality of life or overall well-being (Gold et al. 1996).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Health Risk Factors
Chemical, psychological, physiological, or genetic factors and conditions that predispose an individual to the development of a disease.

Health Service Area
Geographic area designated on the basis of such factors as geography, political boundaries, population, and health resources, for the effective planning and development of health services.

Health Services Research
Health services research is the multidisciplinary field of scientific investigation that studies how social factors, financing systems, organizational structures and processes, health technologies, and personal behaviors affect access to health care, the quality and cost of health care, and ultimately our health and well-being. Its research domains are individuals, families, organizations, institutions, communities, and populations.

Health State/Health Status
The health of an individual at any point in time (Gold, Stevenson, and Fryback 2002).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Health Status
The state of health of a specified individual, group, or population. It may be measured by obtaining proxies such as people's subjective assessments of their health; by one or more indicators of mortality and morbidity in the population, such as longevity or maternal and infant mortality; or by using the incidence or prevalence of major diseases (communicable, chronic, or nutritional).

Conceptually, health status is the proper outcome measure for the effectiveness of a specific population's medical care system, although attempts to relate effects of available medical care to variations in health status have proved difficult.

Healthy Years Equivalent (HYE)
The hypothetical number of years in good health that is equivalent to an individual's current health status or a defined state of ill-health. The HYE allows for ordering preferences of health states.

Health Plan Employer Data and Information Set (HEDIS)
A set of performance measures for health plans developed for the National Committee for Quality Assurance (NCQA) that provides purchasers with information on effectiveness of care, plan finances and costs, and other measures of plan performance and quality.

A specific system (such as a method or theory) used to interpret text, and systems of reading. May also be understood as the process of researching individual parts or perspectives to develop theory.

Heteroscedascity [FIG]
The variance of the error term is not the same for all observations. For example, the variance of the error term might increase with the mean of the dependent variable as is often the case with health expenditures. Or the variance might be larger for observations in one group, for example, one health plan, than in another group. The presence of heteroscedascity does not bias the estimators, but it does bias estimates of their standard errors. Common tests to detect the presence of heteroscedascity include the White test, the Breusch-Pagan test, and the Goldfeld-Quandt test.

High-risk Pool
A subsidized health insurance pool organized by some States as an alternative for individuals who have been denied health insurance because of a medical condition, or whose premiums are rated significantly higher than the average due to health status or claims experience. Commonly operated through an association composed of all health insurers in a State. HIPAA allows States to use high-risk pools as an "acceptable alternative mechanism" that satisfies the statutory requirements for ensuring access to health insurance coverage for certain individuals.

Hindsight Bias
A bias in investigating the cause of a medical error or accident where in retrospect the reviewer simplifies the cause of the error to a single element, overlooking multiple contributing factors. The hindsight bias makes it easy to arrive at a simple solution or to blame an individual, but often makes it difficult to determine the true cause(s) of the error or propose systematic solutions.

Histogram [FIG]
A visual display of quantitative data in the form of a bar chart in which the height of the columns is determined by frequency counts within each of the possible values or specified ranges of a variable.

Historical Controls
A group of patients chosen from a sample who were observed in the past. In this type of study, patients received new treatment form the experimental group and patients who received standard treatment serve as "controls."

Selection bias might be a concern in this approach because subjects receiving new treatment might not be comparable to subjects that received traditional treatment.

A type of threat to internal validity referring to the possibility that events or incidents that take place between periods of observation may influence the study outcome in unintended ways that cannot be controlled by the researcher.

Health Maintenance Organization (HMO)
An entity with four essential attributes: (1) an organized system providing health care in a geographic area, which accepts the responsibility to provide or otherwise assure the delivery of (2) an agreed-upon set of basic and supplemental health maintenance and treatment services to (3) a voluntarily enrolled group of persons and (4) for which services the entity is reimbursed through a predetermined fixed, periodic prepayment made by, or on behalf of, each person or family unit enrolled. The payment is fixed without regard to the amounts of actual services provided to an individual enrollee. Individual practice associations involving groups or independent physicians can be included under the definition.

Horizontal Integration
Merging of two or more firms at the same level of production in some formal, legal relationship.

Hotdeck Imputation
An imputation approach that uses 'donor' responses from units with similar characteristics and substitutes these responses for missing values. Donors may be selected at random from a pool of observations with a similar set of characteristics (e.g. age, sex, and occupation).

Health Resources and Services Administration (HRSA)
One of the eight agencies of the U.S. Public Health Service, HRSA has responsibility for addressing resource issues relating to access, equity, and quality of health care, particularly to the disadvantaged and underserved. HRSA provides leadership to assure the support and delivery of primary health care services, particularly in under-served areas, and the development of qualified primary care health professionals and facilities to meet the health needs of the nation. HRSA focuses on support of states and communities in their efforts to plan, organize, and deliver primary health care, as well as strengthen the overall public health system.

The process of substituting an estimated answer into a field or data record that has missing data, or has an implausible or incorrect value.

Incidence [EQ]
The number of new cases of disease (or other event) in a defined population and time period. Number of new cases in one year is frequently used as a measure of incidence.

Incidence Rate
The ratio of new cases of a disease or condition occurring (in a specific population over a specified period of time) compared to the population of individuals who were at risk for the disease. Generally expressed in person-years.

Incremental Cost
The additional cost of one intervention compared to another.

Incremental Cost-Effectiveness Ratio (ICER)
The ratio of the difference in the cost of two treatments or alternatives divided by the difference in outcome of the two treatments. If there are more than two treatments to compare, a table of incremental cost-effectiveness ratios is often presented, with the dominated alternatives excluded and the remaining alternatives ordered from least to most costly and effective.

A duty to make good any loss, damage, or liability incurred by another. (i.e. the obligation to "make someone whole" after a loss).

Independent [EQ]; [FIG]
Two events (or variables) are independent if they have no influence on one another. That is, the occurrence of one event does not affect the probability that the other will occur.

Mathematically, events A and B are independent if the unconditional probability of A is the same as the conditional probability of B given A, and vice-versa.

Independent Practice Association (IPA)
An organized form of prepaid medical practice in which participating physicians remain in their independent office settings, seeing both enrollees of the IPA and private-pay patients. Participating physicians may be reimbursed by the IPA on a fee-for-service basis or a capitation basis.

Independent Variable
A measurement (or multiple measurements) of individuals, populations, organizations or other relevant unit of analysis (possibly experimentally manipulated) that explains variation in the outcome).

In an experiment, the independent variable is the variable that is manipulated in order to determine its relationship to an outcome of interest. In a causal model the cause is the independent variable. Often denoted as X in regression models. In the context of regression equations, the terms "independent" and "explanatory" variables often are used interchangeably. However, explanatory variables need not be independent (of each other) in systems of simultaneous equations. Thus, the term "explanatory variable" is more general than the term "independent variable."
Synonyms: Explanatory Variable; Predictor Variable; Regressor Variable; Regressor; Covariate; Exogenous Variable; Right Hand Variable; Manipulated Variable

Indirect Costs
A term that is best to avoid because it has two or more incompatible definitions.

1) In cost-accounting, costs that are shared among many services or interventions are called "indirect costs," or "overhead." Indirect costs are allocated among interventions in proportion to some often arbitrarily chosen "basis," and then combined with direct costs to derive total costs of an intervention. For example, maintenance costs for a hospital facility are often allocated in proportion to the number of patient bed-days, or in proportion to the total direct costs per case. 

"Overhead costs" is a preferred term.

2) In cost-benefit analysis and cost-of-illness analysis, the monetary value of productivity losses attributable to an illness or intervention. Includes morbidity costs (goods and services not produced due to illness). In cost-of-illness analysis, may also include mortality costs (goods and services that could have been produced were it not for premature death). 

"Productivity costs" is a preferred term.

Individual Authorization
A research participants' authorization to use or disclose their protected health information for research purposes. One of the requirements to receive human subject approvals for research. It is common for individual authorization to be combined with consent to participate in the actual study.

Individually Identifiable Health Information
Information that is a subset of health information, including demographic information collected from an individual:

(1) is created or received by a health care provider, health plan, employer, or health care clearinghouse
(2) relates to the past, present, or future physical or mental health or condition of an individual; the provision of health care to an individual; or the past, present, or future payment for the provision of health care to an individual
    (a) that identifies the individual
    (b) or with respect to which there is a reasonable basis to believe the information can be used to identify the individual

Inductive Process [FIG]
The process of using empirical observations to guide development of theory on inferring general principles from specific observations.

The process of making conclusions about certain attributes of population based on the data from taken from a representative sample. For example, the interpretation of the results of a regression analysis or hypothesis test.

Inferential Statistics
Statistics that allow researchers to draw conclusions, or inferences about a population.

Information Bias
Type of bias that results if the quality of the information obtained for two groups in a study (e.g. cases v. controls or exposed v. unexposed) is systematically different. This can occur when the means for obtaining information about the subjects in the study differs so that the data on each of the groups is also incorrect.

Informed Consent
The process of a participant agreeing to participate in a study, which involves informing a research participant of key aspects, procedures, risks, and benefits of a study and treatment as well as formal agreement to participate in the study.

Informed consent can be a written document, oral agreement, or legal waiver.

Institutional Review Board (IRB)
A group that follows Federal regulations, State laws, and institutional policy to review, monitor, and approve research in order to protect the ethical rights and privacy of the subjects involved.

IRB approval is generally required for all research involving human subjects.

1. In Survey Research, a data collection tool (e.g., a survey “instrument”).

2. In Econometrics, a variable that changes the value of an endogenous explanatory variable of interest, but has no direct effect on the dependent variable of interest after controlling for the effects of other variables in that equation. Instruments may be factors that the researcher controls (e.g., randomization to the treatment and control groups) or factors that occur naturally in the data.
Synonyms: Manipulator; Instrumental Variable

Instrumental Variables (IV)
A statistical technique for making causal inferences when one of more of the explanatory variables is thought to be correlated with the error term.

An IV must have two properties:
(1) It must be correlated with the suspected endogenous explanatory variable, preferably highly so.
(2) It must not be correlated with the error term, e.g., it must not affect the dependent variable in any way except through the endogenous explanatory variable. If a variable is a valid instrument, the coefficients on the explanatory variables obtained from IV estimation will be unbiased.

Instrumental Variables Regression [EQ]
A methodological technique for producing unbiased estimates of parameter coefficients when one or more of the independent variables is correlated with the error term. The most common computational approaches for estimating an instrumental variables regression are two- and three-stage least squares (2SLS and 3SLS).

A type of threat to internal validity that refers to changes within a study participant over time. Similar to history threats may influence the study outcome in unintended ways that cannot be controlled by the researcher.
Synonyms: Instrument Decay

Inter Alia
Latin for "among other things."

Inter Alios
Latin for "among other persons."

An interaction is assumed to be present when the incidence rate of two or more factors differs from the incidence rate of each of the factors individually. If the joint effect of the two factors is greater than the individual effects, this is called positive interaction or synergism. If the joint effect of the two factors is less than the individual effects, this is called negative interaction or antagonism.
Synonyms: Effect Modification

Interaction Effect [EX]
1. In statistics, an interaction effect is a situation where the two or more independent variables, in combination, have a distinct effect on the dependent variable, and the joint effect differs from the individual effect of either independent variable.

2. In epidemiology, an interaction effect may refer to a biological interaction, the interdependent operation of two or more factors that cause or prevent a disease. Biological interaction occurs when the magnitude of the effects that the causal agents have on a disease varies with the level of a third factor.
Synonyms: 1, in Statistics: Effect Modification; Conditioning Effect; Moderating Effect; Statistical Interaction 2.in Epidemiology: Biological Interaction; Synergism

Interaction Term
An independent variable that causes the magnitude of effect of an explanatory variable on an outcome to differ according to the level of this independent variable.
Synonyms: Moderator; Moderator Variable

Internal Consistency Reliability
A form of reliability analysis used to evaluate the homogeneity of items for developing a multiple-item scale to measure a given concept.
Synonyms: Cronbach's Alpha; Alpha

Internal Validity
An aspect of research validity. The degree to which the study is conducted properly without major methodological problems and therefore allows the researcher to draw valid conclusions about the relationships between the variables of interest.

International Classification of Diseases (ICD)
A publication of the World Health Organization (WHO), revised periodically and now in its 10th Revision, dated 1994. The full title is International Statistical Classification of Diseases and Related Health Problems. This classification, which originated for use in deaths, is used world-wide for that purpose. In addition, it has been used widely the United States for hospital diagnosis classification since about 1955 through adaptations and modifications made in the United States of the 8th, and 9th Revisions.

The philosophy that our objective reality is socially constructed. Researchers observe groups and individuals to interpret the meaning that people give to this objective reality.

Interrater Reliability
Measure of agreement between two or more judges assessing the same target (e.g. patient, episode of behavior). High correlation between scores assigned to an individual indicates that the scores are reliable. Measures of agreement may include: Cohen's Kappa, correlation, interclass correlation coefficients (ICC).

Interrupted Time Series Design
Study design in which outcomes are measured repeatedly in a single group of participants both before and after a manipulation or a natural event.
Synonyms: Simple Interrupted Time Series Design; The Time Series Experiment

A defining principle of naturalistic observation. Naturalistic observation seeks to detect invariants, which are aspects of, or patterns that exist in the real world.

Ipso Facto
Latin for "by the fact itself."

Jackknife [EQ]
A statistical technique used to estimate the variance of an estimator or sample statistic, usually in order to calculate a robust standard error or confidence interval or when a closed form solution is not available. The jackknife method produces this estimate by omitting a small group of observations in the sample at a time (or each of the observations one by one) and recalculating the variance following each omission.

The Joint Canada/United States Survey of Health
The Joint Canada/United States Survey of Health is a collaborative project undertaken by Statistics Canada and the National Center for Health Statistics (NCHS) of the U.S. Center for Disease Control and Prevention. A one time telephone survey, which began in November 4, 2002 and ended March 31, 2003, represents the ten Canadian provinces, all 50 United States, and the District of Columbia. The analytical report in this publication compares Canadians and U.S. residents on a broad range of health dimensions including self-perceived health, chronic conditions, functional status, life-style factors such as smoking and obesity, health care utilization and satisfaction with health care services.

The Joint Commission
A national private, nonprofit organization whose purpose is to encourage the attainment of uniformly high standards of institutional medical care. Establishes guidelines for the operation of hospitals and other health facilities and conducts survey and accreditation programs.
Synonyms: Formerly the Joint Commission on Accreditation of Healthcare Organizations (JCAHO)

Kappa Statistic [EQ]
Quantifies the extent to which the observed agreement exceeds that which would be expected by chance alone.

Kappa greater than 0.75 represents excellent agreement while kappa less than 0.4 represents poor agreement.
Synonyms: Kappa

Kendall's Tau
A nonparametric measure of association between two ordinal variables, most often for ranked, paired data. Kendall's Tau tests independence as well as measures the strength of dependence between variables. Often used in tests for trends.

Key Informant
A person who is extensively interviewed by researcher because he/she is an insider to a group and can share valuable insights to qualitative and descriptive data.

Kruskal-Wallis Test
A non-parametric alternative to a one-way ANOVA. Used for studies with one independent variable and two or more levels. Tests the null hypothesis that the populations studied have identical distribution functions.

Kuder-Richardson Formula [EQ]
A special case of the coefficient alpha measure of reliability for questions with dichotomous (i.e. yes/no) outcomes rather than multi-level response categories.

Large Simple trials
A prospective, randomized controlled trial that includes a large number of participants (typically several thousand) and is conducted over a long period of time. This method features broad patient eligibility criteria, simple enrollment procedures, and minimal data collection on each patient.

The goal of large trials is to improve external validity and gain data on effectiveness, in particular to detect small or modest treatment effects.

Law of Large Numbers
The idea that the probability of an event is roughly equivalent to its long-run relative frequency. The relative frequency of an event is the number of times an event occurs, divided by the total number of trials. The law of large numbers states that as the number of trials or replications of an event increases, the relative frequency of an event will converge on the probability of the event.

Law Review
A scholarly journal or periodical that is produced by a law school and typically written by judges, lawyers and law students. In most law schools, only the top members of the class participate in the law review.

Legal Opinion
An official opinion provided by a representative of the government (such as an attorney general or general counsel to an agency) or an attorney in which his or her understanding of the law is applied to described facts.

Legal Portals
Sites that act as gateways to information on the internet. These include: FindLaw, Jurist and the Cornell Legal Information Institute.

Length-Biased Sampling [EX]
A statistical artifact that occurs in survival analysis when the probability of including an individual observational unit in a sample is related to its survival time. The consequence is that the sample mean becomes a biased estimator of the population mean.

Lexis Nexis
A service that compiles legal opinions, news articles, and other scholarly resources, and other types of information in a variety of areas. There are over five billion searchable documents from legal, news and business sources, which are generally available through a paid subscription.

A legal obligation or responsibility.

Likelihood Ratio [EQ]
In the evaluation of screening tests, the likelihood ration refers to the ratio of the True Positive rate to False Positive Rate. In other words, the ratio of likelihoods that a positive diagnostic test result is expected in a diseased person to the likelihood that a positive result is expected in a non-diseased person.

Likert-Type Rating Scale
A survey response scale that asks respondents to indicate their attitude by rating their level of satisfaction or degree of agreement with a statement. The scale usually ranges from 1-5; 1-7; or 1-9 points. Degrees of agreement many be substituted for numeric points on a scale.

Limted Data Set
Censored protected health information that excludes direct identifiers of the study participants. These datasets are created for research purposes and excludes information such as names, addresses, phone numbers, and other personal identification material.

Literature Review
Type of study design to provide a summary and interpretation of previously published literature in a specific area. May be systematically conducted or provide an unstructured qualitative review by a single author.

Log-Rank Test
A statistical test used in survival analysis to compare survival across groups over a period of time. In general, the Mantel-Haenszel test is an inference test for an unpaired design with a nominal dependent variable that is not affected by time and a nominal independent variable.
Synonyms: Mantel-Haenszel; Mantel-Cox test

Logit [EQ]; [FIG]
A mathematical transformation of a quantity between 0 and 1.
Synonyms: Logistic regression; Binomial logistic regression

Longitudinal Study [FIG]
Studies that conduct two or more repeated measurements per subject in a group of subjects over two or more time periods.
Synonyms: Panel Data Study; Longitudinal Analysis; Longitudinal Study

Majority Opinion
An opinion held by more than half of the judges in a given case.

Managed Care
The body of clinical, financial and organizational activities designed to ensure the provision of appropriate health care services in a cost-efficient manner. Managed care techniques are most often practiced by organizations and professionals that assume risk for a defined population (e.g., health maintenance organizations).

Mandatory Reporting
A system under which physicians or other health professionals are required by law to inform health authorities when a specified event occurs (e.g., a medical error or the diagnosis of a certain disease).

Manipulation [FIG]
Changing the value of a variable. For example, in instrumental variable estimation, we look for an instrument that “manipulates” the value of an endogenous explanatory variable, but has no direct effect on the dependent variable after controlling for other variables in the equation of interest.

Mann- Whitney U Test
A nonparametric statistical method used to determine whether two populations are statistically different by comparing whether both populations have the same medians. Often substituted for a t-test when the dependent variable for two independent samples is ordinal.

The Mann-Whitney U Test is conducted by combining the two samples into a single group, and ranking the observations. Then, the ranks of each of the two populations are summed and compared with one another.
Synonyms: Mann-Whitney-Wilcoxon (MWW) test; the Wilcoxon rank-sum test

Markov Chain Monte Carlo (MCMC) [FIG]
An algorithm used in Bayesian statistics to derive marginal distribution from complex joint distributions of parameters. MCMC models cycle among the parameters of the model, drawing sample values of each parameter from the conditional distribution of the parameter holding other parameters constant. After a sufficient number of these sample values have been drawn, the resulting empirically derived marginal distributions of the parameters are a good approximation to the true parameter distributions. The model, the mean, the median, or the mode of these parameters can be used as point estimates.
Synonyms: Monte-Carlo Markov Chain

Matched Pair Study
An attempt to assign case-control among study subjects who are identical with respect to relevant factors in a study. Generally, once pairs are established, treatments are randomized to members of the pair so that one member of the pair receives Treatment A and the other receives Treatment B, suggesting that the only difference in the pairs is the treatment.

Case-control pairs should not be matched on variables of scientific interest.

Matched Pairs T-Test
A hypothesis test to compare the sample means of two populations and assess whether there is a significant mean difference between two sets of paired data. Assumes the data are two measurements on the same individuals, e.g. before and after the intervention.

Matched Propensity Score Analysis
A statistical technique designed to construct a comparison group similar to the treatment group with respect to the observable characteristics of subjects. The propensity score expresses the probability of being in one group as opposed to another, based on observable characteristics prior to an intervention. Assuming that all variables associated with assignment to treatment are observed and measured, the propensity score can be used to match individual observations with the match(es) that are closest, while eliminating those that are the worst. This approach weeds out the irrelevant observations that are poor matches for members of the treatment group more efficiently than attempting to construct a relevant comparison group by matching on one or two observable characteristics.

Mathematical Identity
A mathematical equation without a stochastic error term.

A type of threat to internal validity that refers to changes within a study participant over time. Similar to history maturational threats may influence the study outcome in unintended ways that cannot be controlled by the researcher.

Maximum Likelihood Estimation
An approach to estimating parameter values which calculates the values of parameters of interest (e.g., regression coefficients) that are most likely to have produced the observed data. As long as the distribution is specified correctly, the maximum likelihood estimator is asymptotically consistent, efficient, and normally distributed.

McCarran-Ferguson Act
A 1945 Act of Congress exempting insurance businesses from Federal commerce laws and delegating regulatory authority to the states.

Mean [EQ]
A measure of central tendency in a set of data commonly referred as the "average". The arithmetic mean is derived by summing all values then dividing by the number of observations.
Synonyms: Average; Arithmetic Mean

Mean Imputation
An imputation approach in which the average response from all responders with the same characteristics is substituted for all missing values.

Measurement Validity
The extent to which there are systematic differences between the information obtained in response to a question relative to:

(1) The full meaning of the concept it was intended to express (content validity)
(2) Related questions about the same concept (criterion validity)
(3) Theories or hypotheses about the relationship of the question to other concepts (construct validity).
Synonyms: Construct Validity

The value half-way through an ordered data set, below which 50% of the scores fall. A good measure of central tendency that works well with skewed (non-normal) data.

An independent variable that lies on the causal pathway between the treatment and the outcome.
Synonyms: Intervening Variable

Mediator Variable
A variable that lies on the causal pathway between an explanatory variable and a dependent variable and represents one of the ways in which the explanatory variable caused the dependent variable. For example, tar is a mediator between cigarette smoking and lung cancer.

Medicaid (Title XIX)
A Federally aided, State-operated and administered program that provides medical benefits for certain indigent or low-income persons in need of health and medical care. The program, authorized by Title XIX of the Social Security Act, is basically for the poor. It does not cover all of the poor, however, but only persons who meet specified eligibility criteria. Subject to broad Federal guidelines, States determine the benefits covered, program eligibility, rates of payment for providers, and methods of administering the program.

Medical Audit
Detailed retrospective review and evaluation of selected medical records by qualified professional staff. Medical audits are used in some hospitals, group practices, and occasionally in private, independent practices for evaluating professional performance by comparing it with accepted criteria, standards, and current professional judgment. A medical audit is usually concerned with the care of a given illness and is undertaken to identify deficiencies in that care in anticipation of educational programs to improve it.

Medical Error
An error or omission in the medical care provided to a patient. Medical errors can occur in diagnosis, treatment, preventative monitoring, or in the failure of a piece of medical equipment or another component of the medical system. Often, but not always, medical errors result in adverse events such as injury or death.

The Medical Expenditure Panel Survey (MEPS)
MEPS is the third in a series of medical expenditure surveys conducted by the AHRQ. It is nationally representative survey that collects detailed information on the health status, access to care, health care use and expenses, and health insurance coverage of the civilian non-institutionalized population of the United States and nursing home residents.

MEPS is comprised of four component surveys:
1) the Household component;
2) the Medical Provider Component;
3) the Insurance Component; and
4) the Nursing Home Component.

The Household Component is the core survey and is conducted each year using an overlapping panel design to collect data for two calendar years from each sampled household.

The Medical Expenditure Panel Survey – Household Component (MEPS-HC)
The Medical Expenditure Panel Survey – Household Component (MEPS-HC) is a basic component of the Medical Expenditure Panel Survey, which documents the demographic, economic, and health status of households to determine medical costs incurred by households. MEPS-HC began in 1996 and is sponsored by the Department of Health and Human Services (HHS) through the Agency for Healthcare Research and Quality (AHRQ). It is a nationally representative survey of households across the United States that is conducted over a two-year span.

The Medical Expenditure Panel Survey – Insurance Component (MEPS-IC)
The Medical Expenditure Panel Survey – Insurance Component (MEPS-IC) produces estimates on employer-sponsored health insurance data in the United States. Conducted by NCHS, the NEHIS provides baseline data for state and national estimates on various aspects of employer-sponsored health insurance. This includes the number of employees eligible for coverage and take-up rate, type and number of plans offered, and detailed characteristics of plans. Between April and December of 1994, the NEHIS was conducted using the Computer-Assisted Telephone Interview (CATI) program. The NEHIS was the first federal survey to represent all employers in the United States by State and obtain information on all plans offered to employees by their employers. In 1996, the NEHIS was re-named the Medical Expenditure Panel Survey – Insurance Component (MEPS-IC) and is now conducted by the Agency for Healthcare Research and Quality (AHRQ).
Synonyms: National Employer Health Insurance Survey (NEHIS)

Medical Informatics
The systematic study, or science, of the identification, collection, storage, communication, retrieval, and analysis of data about medical care services that can be used to improve decisions made by physicians and managers of health care organizations.

Medical Management Information System (MMIS)
A data system that allows payers and purchasers to track health care expenditure and utilization patterns.

Medical Review Criteria
Systematically developed statements that can be used to assess the appropriateness of specific health care decisions, services, and outcomes.

Medically Necessary
A treatment or service that is appropriate and consistent with a patient's diagnosis and that, in accordance with locally accepted standards of practice, cannot be omitted without adversely affecting the patient's condition or the quality of care.

Medically Needy
Persons who are categorically eligible for Medicaid and whose income, less accumulated medical bills, is below state income limits for the Medicaid program.

Medicare (Title XVIII)
A U.S. health insurance program for people aged 65 and over, for persons eligible for Social Security disability payments for two years or longer, and for certain workers and their dependents who need kidney transplantation or dialysis. Monies from payroll taxes and premiums from beneficiaries are deposited in special trust funds for use in meeting the expenses incurred by the insured. It consists of two separate but coordinated programs: hospital insurance (Part A) and supplementary medical insurance (Part B).

Medicare Advantage
Medicare Advantate was created as part of the 2003 Medicare Modernization Act to update Medicare+ Choice (Medicare part C), which was created by the 1997 Balanced Budget Act. Medicare Advantage allows the Centers for Medicare and Medicaid Services (CMS) to contract with a variety of different managed care and fee-for-service entities offering greater flexibility to Medicare participants. Persons eligible for Medicare parts A and B are also eligible for Medicare Advantage.

The Medicare Current Beneficiary Survey (MCBS)
The Medicare Current Beneficiary Survey (MCBS) collects information of the health status of the Medicare population and is analyzed by the Office of Strategic Planning of the Centers for Medicare & Medicaid Services (CMS) to determine the expenditures and sources of payment for Medicare services. The MCBS is conducted through contract with Westat Incorporated. The survey was first conducted in 1991 through household interviews and is representative of the entire Medicare population.

Medicare Payment Advisory Commission (MedPAC)
MedPAC is an independent federal body that advises the U.S. Congress on issues affecting the Medicare program. It was established by the balanced Budget Act of 1997 (P L. 105-33), which merged the Prospective Payment Assessment Commission (ProPAC) and the Physician Payment view Commission (PPRC).

Medigap Policy
A private health insurance policy offered to Medicare beneficiaries to cover expenses not paid by Medicare. Medigap policies are strictly regulated by Federal rules.
Synonyms: Medicare Supplemental Insurance

MedWatch is the Food and Drug Administration’s voluntary reporting system for adverse events. The MedWatch system is intended to detect safety hazard signals for medical and pharmaceutical products. If a problem is detected, the FDA can issue medical product safety alerts or order product recalls, withdrawals, or labeling changes to protect the public health. The MedWatch also collects information on other FDA-regulated products (e.g., dietary supplements, cosmetics, medical foods, and infant formulas). Voluntary reporting by healthcare professionals, consumers, and patients is conducted using a single-page reporting form. Reporting can be conducted online, by phone, or by submitting the MedWatch 3500 form by mail or fax.

Meta Analysis
A quantitative method for combining the results of multiple studies (on the same topic) to obtain an overall estimate of a particular treatment or intervention.
Synonyms: Meta-Analysis

Minimal Risk
A guarantee that the level of anticipated harm or discomfort associated with participating in a study is not greater than those encountered in daily life or routine examinations or tests.

Missing at Random (MAR)
A situation in which missing values in a sample of data follow a pattern that can be predicted based on the values of the other variables in the data set and is not due to the missing values themselves. If, for example, the probability of missing data on income varies according to the age of a respondent, but for any given respondent age, the probability of missing data on income is unrelated to the level of income, then the data are MAR.

Missing Completely at Random (MCAR)
Cases in which missing values are truly missing at random; that is, they are randomly distributed across all observations in the sample. Although data that are MCAR may result in less precise estimators (due to an increase in their standard errors), they do not cause the estimators to be biased.

The most frequently occurring value in a set of discrete data. It is possible to have more than one mode in a dataset if one or more values are equally common. For example, a "bi-modal" distribution is one in which two values are equally or most common.

A simplified representation of an actual phenomenon or system that specifies the components of the phenomenon or system and the relationships among those components.

Moderator Variable
A variable that alters the effect of an explanatory variable on a dependent variable. For example, appropriate medical care can in some cases alter the effect of disease on mortality. Moderating effects often are tested by including interaction terms in estimated equations.

The extent of illness, injury, or disability in a defined population. It is usually expressed in general or specific rates of incidence or prevalence.

Death. Used to describe the relation of deaths to the population in which they occur. The mortality rate (death rate) expresses the number of deaths in a unit of population within a prescribed time and may be expressed as crude death rates (e.g., total deaths in relation to total population during a year) or as death rates specific for diseases and, sometimes, for age, sex, or other attributes (e.g., number of deaths from cancer in white males in relation to the white male population during a given year).

Mortality Rate
The number of deaths in a population within a prescribed time, expressed as either crude death rates or death rates specific to diseases and sometimes to age, sex, and other attributes (Turnock 2004).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Research that encompasses perspectives and/or research methods from two or more disciplines. Multi-disciplinary research may be conducted by a single researcher or a team of researchers; in either case, each individual uses the analytic tools specific to his or her own discipline to conduct the examination.

Multi Stage Sampling
Sample selection using a combination cluster and stratified of sampling methods applied in stages.

Multinomial Logistic Regression
A model used for a nominal outcome with three or more values where the determinants of the choice probabilities are based on characteristics of subjects.
Synonyms: Multinomial Logit Regression; Polytomous Logistic Regression

Mulitnomial Probit Regression
A model used for a nominal outcome with three or more values where the determinants of the choice probabilities are based on characteristics of subjects and a multivariate normal distribution.

Multiple Imputation [FIG]
A strategy for dealing with missing data that uses data on non-missing observations to predict new values for each missing observation in the sample. For example, in order to estimate missing values for Y, this procedure uses the existing values of X to create new values for Y, and then adds a random error component to each predicted value.

After several completed data sets are generated in this fashion (five is the norm), the estimates obtained from analyzing each data set separately are combined into an overall estimate.

Multiple Regression
A statistical model in which the values of the outcome variables are influenced by two or more (multiple) regressors.
Synonyms: Multivariate Regression

Multiple Time Series Design [FIG]
A form of Time Series Design that adds a control equivalent control group to the Time Series design.

Multitrait Multimethod Approach
A formal test of the construct validity for measures of complex concepts. A matrix is used to compare correlations between two or more traits (i.e. physical and mental functioning) as measured by two or more methods (i.e. a survey and exam). The hypothesis is that measures should be more highly correlated with measures of the same trait across methods than with different traits measured by the same method.

N of 1 trial
A RCT (randomized controlled trial) involving one single patient. This patient will receive multiple, different treatments in a random order (multiple crossover design). An N of 1 trial is used to determine and compare the effects of various treatments on a specific patient.

Narrative Analysis
A form of research that relies on participant's stories/narrative to obtain an understanding of the goals and intentions of a group.

The process an patterns used by a research subject to tell a story, and the way these are interpreted by the observer. Narrativity is distinguished from narrative, which limited to the content of a story.

The National Ambulatory Medical Care Survey (NAMCS)
The National Ambulatory Medical Care Survey (NAMCS) is a national survey designed to meet the need for objective, reliable information about the provision and use of ambulatory medical care services in the United States. It was conducted annually by the Center for Health Statistics (NCHS) from 1973 to 1981, again in 1985, and has been conducted annually since 1989. Physicians are trained by NCHS interviewers on data collection and are assigned a one-week reporting period. Physicians are responsible for collecting data about patients’ symptoms, physicians’ diagnosis, and medications prescribed. For survey years 1973-91, there are two data files--one for patient visit data and a second for drug mention data. The second file is limited to those visits with mention of medication therapy. For the 1991 data, it is possible to link information on the drug file with information on the patient visit file. Beginning with the 1992 survey year, only one data file is produced annually that contains both patient visit and drug information.

The National Asthma Survey
The National Asthma Survey is a one time survey that examines the health, socioeconomic, behavioral, and environmental predictors that relate to asthma. Findings of the National Asthma Survey help characterize the health care experiences of persons with asthma. In 2003, telephone interviews were conducted using the State and Local Integrated Telephone System (SLAITS). The survey is sponsored by the National Center for Environmental Health (NCEH) and Centers for Disease Control and Prevention.

The National Epidemiologic Survey on Alcohol and Related Conditions (NESARC)
The National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) measures the prevalence of alcohol, drug use, and associated psychiatric co-morbidities among the civilian, non-institutionalized adult population in the United States. NESARC is a longitudinal survey. The first wave of interviews was conducted from 2001-2002 and a second wave was conducted from 2004-2005. NESARC is sponsored by the National Institute on Alcohol Abuse and Alcoholism (NIAAA) and conducted through personal household interviews with computer assistance.

The National Fire Incident Reporting System (NFIRS)
The National Fire Incident Reporting System (NFIRS) collects and maintains information on the response to fires and incidents from fire departments across the United States. Authorized by the Federal Fire Prevention and Control Act of 1974, the National Fire Data Center in the United States Fire Administration (USFA) helps State and local governments develop fire reporting and analysis capability for their own use, and obtain data that can be used to assess and combat fires on a national scale. Incident, causality, and optional reports are completed by fire-departments for their fires and incidents, which are validated and consolidated into a single computerized database.

National Guideline Clearinghouse (NGC)
NGC (sponsored through a partnership between AHRQ, AAHP, and the AMA) is a publicly available electronic repository for clinical practice guidelines and related materials that provides online access to guidelines at www.guideline.gov.

The National Health and Nutritional Examination Survey (NHANES)
The National Health and Nutritional Examination Survey (NHANES) was established in 1959 to assess health and prevalence of disease in the United States’ population. NHANES is authorized by the National Health Survey Act of 1956 and is conducted by the National Center for Health Statistics Information is collected through health exams, which consist of a physical, urine, vaginal, blood, and nutritional analysis. The survey has been fielded seven times, and has been conducted annually since 1999 through home interview and a health examination of approximately 7,000 people.

The National Health Interview Survey (NHIS)
The National Health Interview Survey (NHIS) is used to evaluate progress towards national health goals and programs, as well as monitoring patterns and trends in health. Conducted by the National Center for Health Statistics (NCHS), NHIS was authorized in 1957 as part of the National Health Survey Act of 1956 and is administered by the US Census Bureau. This is an annual, nationwide, in-person survey of approximately 40,000 households.

The National Health Interview Survey on Disability (NHIS-D)
The National Health Interview Survey on Disability (NHIS-D) is an extension of the National Health Interview Survey that was conducted from 1994-1995. Authorized by the Americans With Disabilities Act (ADA) in 1990, NHIS-D is conducted by the National Center for Health Statistics (NCHS) in two phases. Phase I of the NHIS-D is a personal interview, designed to identify a wide range of children and adults with chronic conditions, impairments, disabilities and elevated service needs. Phase II of the NHIS-D was designed as a follow-up questionnaire with individuals who were identified in Phase I and focused utilization and need for services.

The National Home and Hospice Care Survey (NHHCS)
The National Home and Hospice Care Survey (NHHCS) was established in 1992 to be a continuing survey designed to track changes in home health and hospice care usage. Supported by the Department of Health and Human Services (HHS) and conducted five times between 1992 and 2000 by the National Center for Health Statistics (NCHS), the survey provides information to identify areas in need of support and promote policy-making. Through personal interviews with certified home hospice administrators and staff, information is collected about the health status of current patients, their length of stay, services provided, and reasons for discharge.

The National Hospital Ambulatory Medical Care Survey (NHAMCS)
The National Hospital Ambulatory Medical Care Survey (NHAMCS) provides a national estimate of the use of hospital ambulatory medical care services in the United States. Since 1992, NHAMCS has been conducted annually by the National Center for Health Statistics (NCHS). Information is collected by a national sample of hospitals’ emergency departments and outpatient departments. Over a 4-week period patients are randomly sampled and interviewed using the Patient Record Form.

The National Hospital Discharge Survey (NHDS)
The National Hospital Discharge Survey (NHDS) has been conducted annually since 1965 by the National Center for Health Statistics (NCHS) to provide information on characteristics of inpatients discharged from non-Federal, short-stay hospitals. The NHDS collects data from a sample of approximately 270,000 inpatient records acquired from a national sample of about 500 hospitals.

The National Immunization Survey (NIS)
The National Immunization Survey (NIS) is a joint effort between the Childhood Immunization Initiative and the National Center for Health Statistics (NCHS) to measure vaccination levels of young American children. The Survey is conducted by the National Opinion Research Center (NORC) at the University of Chicago for the Centers of Disease Control and Prevention (CDC). Since 1994, the survey has been conducted by a random digit-dialing (RDD) telephone system targeting households with children between the ages of 19 and 35 months. Vaccination information is validated with a follow up mail survey to the child’s provider.

The National Long-Term Care Survey (NLTCS)
The National Long-Term Care Survey (NLTCS) is conducted to evaluate the health and lifestyles of adults, ages 65 or older. The NLTCS is a longitudinal, written survey that is distributed approximately every five years and is nationally representative of the general and institutionalized population. Established in 1982, it has since completed six cycles, sharing and changing sponsorship between several organizations. The most recent survey was conducted in 2004 with sponsorship from the National Institute on Aging (NIA) and Duke University and was conducted by the United States Census Bureau.

The National Longitudinal Mortality Study (NLMS)
The National Longitudinal Mortality Study (NLMS) was conducted and sponsored by the United States Census Bureau to study the effects of differentials in demographic, socio-economic and occupational characteristics on mortality. It is co-sponsored by the National Cancer Institute, the National Heart, Lung, and Blood Institute, the National Institute on Aging, the National Center for Health Statistics. Records from the Current Population Surveys (CPS) of March 1979, April 1980, August 1980, December 1980, and March 1981 are matched to the National Death Index (NDI) to identify the occurrence and cause of death between 1979 and 1989.

The National Nursing Home Survey (NNHS)
The National Nursing Home Survey (NNHS) is used for trend analysis of utilization, services offered, charges for care, and the cost of providing nursing home care. The NNHS has been conducted seven times between 1969 and 2004 by the National Center for Health Statistics (NCHS). The 2004 survey was conducted using a computer-assisted personal interviewing (CAPI) system, along with a supplemental survey of nursing assistants. Requirements for participating nursing homes include certification by the Medicare, Medicaid or State licensure.

The National Survey of Ambulatory Surgery (NSAS)
The National Survey of Ambulatory Surgery (NSAS) is used to evaluate ambulatory services and plan for improvements in ambulatory care. Conducted by the National Center for Health Statistics (NCHS), written surveys are distributed to randomly selected hospitals and ambulatory surgery centers to collect information on patient demographics, source of payment, and the reasons and resolutions of the patient’s visit. The survey is authorized by Section 306 of the Public Health Services Act, sponsored by the Centers for Disease Control (CDC), and was first conducted from 1994 to 1996 and was last conducted in 2006.

The National Survey of America’s Families (NSAF)
The National Survey of America’s Families (NSAF) is used to track the local impact of federal policy changes and produce estimates for federal and state indicators of child health. Participants in this survey are asked about their participation in government and community programs, their demographics and attitude about society and health. First conducted in 1997, the survey is a part of the New Federalism Project, with sponsorship by the Urban Institute. The last survey was last conducted in 2002 by Westat Incorporated, which employed random-digit dialing (RDD) telephone household survey as well as in-person household surveys.

The National Survey of Children with Special Health Care Needs (CSHCN)
The primary goal of this survey is to assess the prevalence and impact of special health care needs among children in the United States. The survey is sponsored by the Maternal and Child Health Bureau of the Health Resources and Services Administration (HRSA), as well as the Office of the Assistant Secretary for Planning and Evaluation of the Department of Health and Human Services. CSHCN provides data to explore the extent to which children with special health care needs have medical care, adequate health insurance, and access to needed services. Other topics include care coordination and satisfaction with care. The CSHCN was conducted over a period between October of Part of state and local are telephone survey, conducted by NCHS, SLAITS – access through NCHS website 2000 and April of 2002, and again between 2005-2006. Using the State and Local Area Integrated Telephone Survey (SLAITS) interviews were conducted with children and their parents and a supplemental survey were conducted in low-income households with uninsured children.

The National Survey of Children's Health
The National Survey of Children’s Health examines the physical and emotional health of children ages 0-17 years of age. Sponsored by the Maternal and Child Health Bureau of the Health Resources and Services Administration (HRSA), data is collected from January of 2003 to July of 2004. Results are stored in an interactive public database housed by the Department of Health and Human Services, called the Data Resource Center for Child and Adolescent Health. The survey can be found through the website for the NCHS’ State and Local Area Integrated Telephone Survey (SLAITS). Special emphasis is placed on factors that may relate to well-being of children, including medical care, family interactions, parental health, school and after-school experiences, and safe neighborhoods. Telephone interviews are conducted in all 50 states and the District of Columbia, and are adjusted for households without telephones.

The National Survey of Early Childhood Health (NSECH)
The National Survey of Early Childhood Health (NSECH) investigates pediatric care from the parent’s perspective. Questions focus on the delivery of pediatric care and health education for families with children 4 – 35 months old. One-time only telephone interviews were conducted from February to July of 2000 using the State and Local Area Integrated Telephone Survey (SLAITS). The survey is conducted by the Centers for Disease Control and Prevention (CDC), supported by the Maternal and Child Health Bureau and the AAP Friends of Children fund, and funded by a grant to the American Academy of Pediatrics.

The National Survey of Family Growth (NSFG)
The National Survey of Family Growth (NSFG) collects statistics on family dynamics, including marriage and divorce, birth control methods, infertility, pregnancy, family health and life. Sponsored by the Department of Health and Human Services (HHS) and administered by the National Center for Health Statistics (NCHS), the NSFG was first conducted in 1973 by the University of Michigan. Authorized by the Public Health Service Act, it is currently in its seventh cycle and is conducted through household interviews.

The National Survey on Drug Use & Health (NSDUH)
The National Survey on Drug Use and Health (NSDUH) collects information on the non-institutionalized population’s use of alcohol, tobacco and illegal drugs. Since 1999, the survey has been able to produce both national and state level usage estimates. The NSDUH was established in 1971 and is authorized by Section 505 of the Public Health Service Act. Sponsored by the Substance Abuse and Mental Health Services Administration (SAMHSA), administration is currently under contract to the Research Triangle Institute (RTI), who has conducted the survey since 1988. Participants are selected through random household sampling of persons over 12 years of age, and interviews are conducted through a face-to-face interview.

Natural Log [EQ]
A logarithm to the base e. The natural log is commonly abbreviated as "ln".

The NCHS National Death Index (NDI)
The National Death Index (NDI) contains records on virtually all deaths in the United States since 1979. Established in 1982 by the National Center for Health Statistics, the NDI is a national, computerized index of death record information compiled from computer files submitted to the National Center for Health Statistics (NCHS) by each state's vital statistics office. About 2.4 million death records are added to the NDI each year.

National Committee for Quality Assurance (NCQA)
A national organization founded in 1979 composed of 14 directors representing consumers, purchasers, and providers of managed health care. It accredits quality assurance programs in prepaid managed health care organizations and develops and coordinates programs for assessing the quality of care and service in the managed care industry.

Near Miss
A medical error that does not result in harm.
Synonyms: Close Call

Nearest Neighbor Imputation
A method of generating 'donor' estimates for hotdeck imputation. In nearest-neighbor imputation, observations are sorted in order to determine which observations are most like the response units with missing values. The value for the observation that is most similar to the observation with missing values is then applied to the missing data.

Failure to act in a manner that a "reasonably prudent person" would have in a similar situation. Conduct or treatment falling below the accepted legal standard.

Negligence Per Se
Negligence that has been established as a matter of law. For example, violation of a statue is typically considered negligence per se.

Nested Case Control Study [FIG]
A case-control study design in which cases and controls are from the subjects in an ongoing cohort study. These studies are hybrids of retrospective and prospective analyses because they start by identifying a population, then collect baseline data and follow the population for the next several years.

The objective is to compare the histories (and sometimes contemporaneous clinical measurements) of individuals who develop the disease with those who do not develop the disease during the study period. Nested case-control studies are an efficient method for studying case-control relationships among a cohort once a number of cases have emerged.

Network Meta-Analysis [EX]
A method used to assess the comparative effectiveness of experimental treatment among similar patient populations that have not been compared directly in a randomized clinical trial. Unlike traditional meta-analysis, which summarize the results of trials that have evaluated the same treatment/placebo combination, Network Meta-Analyses compare the results from two or more studies that have one treatment in common.

Nominal Group Technique
A qualitative study design that uses a small group discussion format to elicit opinions from the group while preventing the discussion from being dominated by one or a few members. Participation from more passive group members is sought by generating a large number of ideas from individual group members. The goal is equal contributions from each participant in an environment without fear of criticism.

At the conclusion, a set of prioritized solutions or recommendations is achieved.

Non-randomized Trial with Contemporaneous Controls [FIG]
A type of trial in which treatment and control group participants are matched at the group level based on demographic and clinical characteristics, and receive different treatments (e.g. A and B) at the same time.

In these studies, outcomes are compared between the group receiving treatment A and the group receiving treatment B contemporaneously. Trials with contemporaneous controls are generally considered superior to a trial with historical controls.
Synonyms: Simultaneous Nonrandomized controls

Non-randomized Trial with Historical Controls [FIG]
A study in which investigators compare outcomes among a group of patients who are receiving a new treatment with outcomes among patients who received standard treatment in a previous period. In these studies, patients receiving new treatment compose the experimental group and patients who received standard treatment are historical controls.

Bias may be a concern in this approach when subjects receiving new treatment are not comparable to subjects who received traditional treatment.

Nonequivalent Comparison Group Pretest-Posttest Design
A study in which two nonrandomized groups are compared in a pretest-posttest design. Pretest measurements are made of both a study group and a comparison group. The study group is then exposed to an intervention or experiment, after which both groups are measured in a posttest.

Nonequivalent Groups Posttest-Only Design
A study in which two nonrandomized groups are compared in a posttest design. In this design, the study group is exposed to an intervention or experiment, after which both groups are measured in a posttest.

Nonexperimental Data
Information not gathered as part of a controlled experiment or from random assignment of study subjects. Nonexperimental data are commonly used in social science research, particularly when gathering experimental data would be too costly or unethical. Because the researcher cannot control assignment of subjects to the treatment and control groups, nonexperimental data are more difficult than experimental data to analyze and interpret. Examples of nonexperimental data include survey data, administrative records, and standardized test scores. They also are known as observational data.
Synonyms: Observational Data

A defining principle of naturalistic observation that researchers who conduct naturalistic observation should not disrupt the events they are observing.

Non-Interrupted Time Series Design
A type of longitudinal study in which measures are repeated on the same study participants over time without, in contrast to an Interrupted Time Series Design, the occurrence of manipulation or a natural event.

Nonparametric Methods
Methods of estimation and hypothesis testing that are not based upon researcher's assumptions about the underlying sampling distribution of the data. These methods allow considerable flexibility in how a model is specified, which helps prevent researchers from reaching improper conclusions by erroneously specifying the models functional form (for example, by assuming the relationship between the dependent and independent variables is linear, when in fact it is not).

Nonparametric Statistics
Inferential statistical methods that do not assume samples are drawn from normally distributed populations. The major distinction from parametric statistics is that non-parametric statistics make no assumption about the frequency distributions of the variables being studied.
Synonyms: Distribution-Free Statistics

Nonparametric tests
A nonparametric test does not make assumptions about the underlying distribution of the data. Nonparametric tests tend to be less powerful than parametric tests, meaning that more observations are required to see meaningful differences between groups.

Nonprobability Sampling [EX]
A sampling technique in which the sample selection is not random. In these studies, the probability of obtaining a particular sample is not known and cannot be calculated. As a result, non-probability samples cannot be generalized to the broader population from which they were sampled.

Normal Distribution [FIG]
A theoretical, continuous probability distribution that represents the range of possible values of a variable. Under the Central Limit Theorem, the sample mean will be approximately normally distributed.

Novelty Effect
A type of threat to external validity that is similar to the Hawthorne effect. A novelty effect refers to the fact that an outcome or result may be due in part to the uniqueness of an experimental situation or testing rather than the intervention itself. In this instance, the novelty of the experiment or situation is a confounder.

Null Set
A grouping, or set, of objects that does not contain any observations or elements.
Synonyms: Empty Set

Observational Data
Information not gathered as part of a controlled experiment or from random assignment of study subjects. Observational data are commonly used in social science research, particularly when gathering experimental data would be too costly or unethical. Because the researcher usually has no control over how the data are collected, observational data are more difficult than experimental data to analyze and interpret. Examples of observational data include survey data, administrative records, and standardized test scores. They are also known as nonexperimental data.
Synonyms: Nonexperimental data

Observational Study
Studies in which researchers observe the subjects in their natural settings and do not attempt to control or manipulate allocation to treatment.

Observed Confounder
A factor or condition, measured by the scientist, that is associated with both an explanatory variable and the dependent variable but is itself not caused by the explanatory variable, e.g. not a consequence of the treatment.
Synonyms: Observed Confounding Variable; Observed Confounding Factor

Observed Score [EQ]
A score directly observed or obtained in a measurement or assessment study. Classical test theory postulates that the observed score is composed of the true score and error, and therefore provides imperfect information.

A type of caused that triggers a specific result.

Odds [EQ]
The ratio of the probability of an occurrence of an event (such as a disease) to that of non-occurrence. Also, the number of individuals in a population that experience an event divided by those in the population who did not experience the event.

Often, odds are calculated based on the probability of an occurrence divided by 1- probability of the occurrence.

Odds Ratio [EQ]
The odds ratio is the ratio of the odds of having the disease if exposed to a risk factor divided by the odds of having the disease if not exposed to a risk factor. This is particularly valuable in cases where the population incidence (NEW cases) of a condition or disease is not known -- as in cohort or case control studies. The odds ratio is approximately equal to the relative risk (RR) for small rates (low prevalence)

Office of Public Health Preparedness (OPHP)
The Office of Public Health Preparedness (OPHP) directs the Department of Health and Human Services' (HHS) efforts to prepare for, protect against, respond to, and recover from all acts of bioterrorism and other public health emergencies that affect the civilian population, and serves as the focal point within HHS for these activities. OPHP is headed by a director, who reports directly to the Secretary, and serves as the Secretary's principal advisor on HHS activities relating to protecting the civilian population from acts of bioterrorism and other public health emergencies. The office was created in January 2002.

Olmstead Decision
A 1999 Supreme Court decision in the case of Olmstead vs. L.C. whereby the court found that necessary institutionalization of individuals with disabilities is discrimination under the Americans with Disabilities Act (ADA). The decision has relevance for State Medicaid programs that provide both institutional and home- and community-based long-term care services. The Court explained that a State may meet its obligation under the ADA by having comprehensive, effectively working plans ensuring that individuals with disabilities receive services in the most integrated setting appropriate to their needs.

Omitted Variable
A variable that is unobserved by the analyst.

Omitted Variable Bias [FIG]
Bias that results from omitted variables.

Not all omitted variables result in biased estimates of causal effects. In order to result in bias, the omitted variable must exert a causal influence on both the explanatory variable(s) of interest and the dependent variable.

One-group Posttest-Only Design
A type of experimental study in which only one group receives a treatment and is then measured in a post test -- after treatment. In this design, there is no control group or baseline condition to compare with.
Synonyms: One-Shot Case Study.

One Group Pretest Posttest [FIG]
A study design in which a sample is observed twice, one prior to (pre), and once after (post) an intervention or experiment.
Synonyms: Before and After Design; One Group Before and After Design

One-Sample T-Test [EQ]
A statistical test to determine whether the observed mean value of a sample differs from a hypothesized population mean.

A courts written statement explaining its decision. Opinions include the statement of facts, points of law, and a rationale for the decision.

Opportunity Cost
The cost of foregone outcomes that could have been achieved through alternative investments.

Ordered Logitistic Regression
A model used for an ordinal outcome with three or more discrete, ordered groups (e.g. a rating scale with response categories such as poor, fair, good, excellent). The rank order of the values must be taken into account.
Synonyms: Ordered Logistic Regression; Ordinal Logistic Regression; Ordinal Log-Linear Regression

Ordinal Scale
Measurement scale that indicates the magnitude of responses relative to one another but does not give any information about differences between responses.

A change in individuals, populations or organizations attributed to (or associated with) independent variables not predetermined by the investigator.
Synonyms: Dependent variable; Response; Endogenous Variable; Regressand; Left Hand Variable; Finding

Outcomes Research
Outcomes research is designed to evaluate the impact of health care on health or economic outcomes. Large population-level datasets are often used to conduct outcomes research, though primary data collection is sometimes conducted. Where large datasets are used, these are often gleaned from administrative or financial data which may not be ideal for research purposes.

Generally, a data point outside the normal or expected range. May be defined specifically by the number of standard deviations above or below the mean value.

A measure of "statistical significance". Most often, if the P value is less than a specified level chosen prior to the study (typically set at 0.01 or 0.05), then the null hypothesis is rejected. If the null-hypothesis is rejected, the P value represents the likelihood that the observed difference between the intervention and control groups was obtained by chance alone.

Pair-wise Comparisons [EX]
After tests for main effects have been performed pair-wise comparisons and other post hoc tests may be conducted to identify where differences occur (e.g. which groups are different from the others.)

A post hoc pair-wise comparison involves calculating an analysis of variance (e.g., One-way between-subjects ANOVA), selecting two means and applying a method such as Tukey to adjust for multiple comparisons.
Synonyms: Post-Hoc Test

Panel Data Study
Study in which the same sample of people are studied at regular intervals over time.
Synonyms: Nested Case-Control Study

Parallel Boxplots
A visual display of quantitative information that compares data from two distributions in parallel box plots on the same scale.

A measurable, numeric characteristic about the population of interest, such has the mean value of some variable, or a parameter in a regression model relating a dependent and an independent variable.

To specify a mathematical model representing unknown population characteristics for theoretical or estimation purposes.

Parametric Test
A parametric test makes assumptions about the underlying distribution of the data.

Participant Observation
A form of gathering data in which the research is an active participant in the study group, in order to gain a first person perspective within the observation. A researcher may either take the role of a covert participant and not disclose their role, or maybe an overt participant and publicly acknowledge the study.

Patient Origin Study
A study, generally undertaken by an individual health program or health planning agency, to determine the geographic distribution of the residences of the patients served by one or more health programs. Such studies help define catchment and medical trade areas and are useful locating and planning the development of new services.

Patient Set
All information related to a patient is contained in the patient set. The patient set may contain multiple layers that represent registry or facility views.

Peer Review
Generally, the evaluation by practicing physicians or other professionals of the effectiveness and efficiency of services ordered or performed by other members of the profession (peers). Frequently, peer review refers to review of research by other researchers.

Per Curiam
Latin for "by the court as a whole."

Per Curiam Opinion
An opinion of the appellate court that does not identify the judge who wrote it.

Perfect Collinearity [EX]
An exact linear relationship between two independent variables included in a regression model.

Performance Measures
Methods or instruments to estimate or monitor the extent to which the actions of a health care practitioner or provider conform to practice guidelines, medical review criteria, or standards of quality.

Period Prevalence Rate
The number of people who have a disease at some point during an interval (period) divided by the number of people at risk.
Synonyms: Prevalence, Limited Duration

Permutations [EQ]
The number of ways "r" objects can be selected from a group of "n" objects.

Person Years [EX]
The total number of years experienced by individual at risk for a particular condition of interest such as a certain disease. Often used in cohort studies as the denominator to standardize calculations of rates.

Personal Liability
Liability in which one individual is personally accountable for an action.

The party presenting a request or action to a court or other official body.

Pharmaceutical Care System
A strategy that attempts to utilize drug therapy more efficiently to achieve definite outcomes that improve a patient's quality of life. A pharmaceutical care system requires a reorientation of physicians, pharmacist, and nurses toward effective drug therapy outcomes. It is a set of relationships and decisions through which pharmacist, physicians, nurses, and patients work together to design, implement, and monitor a therapeutic plan that will produce specific therapeutic outcomes.

The study of the costs and benefits associated with various pharmaceutical treatments.

The study of drug actions/effects in human body, focusing on the amount of time required for the absorption, distribution, metabolism, and elimination of a drug and its metabolites.

The process of detecting, assessing, understanding, and preventing adverse effects of authorized drugs often based on data from post marketing reports, including adverse drug event (ADE) reports. The aim is to promote the safe and effective use of medicines.

Pharmacy Benefit Manager (PBM)
Many insurance companies, HMOs, and self-insured employers contract with PBMs to manage drug benefit coverage for employees and health plan members. Common tools employed by PBMs to manage drug benefits include management of pharmacy networks, implementation of generic substitution and mail-order programs, negotiation of rebates with drug manufacturers, formulary management, and clinical programs such as disease management.

Phase I Trial
A clinical pharmacologic study to determine whether there are any unanticipated safety issues not discovered during preclinical testing and initial evidence of effectiveness. Generally includes 20-80 participants. If the drug passes this phase of testing it then undergoes Phase II Trials.

Phase II Trial
Phase II Trials are designed to accomplish two objectives. IIA studies are designed to assess the appropriate doses of a medication that are well-tolerated and exhibit biological effects. IIB studies attempt to assess the efficacy and relative safety of the drug among a larger population sample, generally 100 to 200 participants.

Phase III Trial
Large scale trial to assess the effectiveness and relative safety of a drug. Once the effectiveness of the drug has been ascertained in Phase II studies, the Phase III trial may enroll up to several thousand patients across multiple centers or locations. Phase III trials tend to compare the new drug against a commonly used drug in a randomized controlled trial, attempting to evaluate the overall benefit-risk relationship of the drug and assist physicians in appropriate use and labeling.

Phase IV Trial
Post-marketing studies to evaluate clinical indications and assess the risk of specific adverse effects. Studies may also be conducted with different dosing variations, treatment intervals, and combinations or comparisons with other drugs used to treat the same indicated condition.

Physical Environmental Determinant
A proposed or established causal factor in the natural and built environment that affects health outcomes (e.g., air and water quality, lead exposure, the design of neighborhoods).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Physician Hospital Organization (PHO)
A legal entity formed by a hospital and a group of physicians to further mutual interests and to achieve market objectives. A PHO generally combines physicians and a hospital into a single organization for the purpose of obtaining payer contracts. Doctors maintain ownership of their practices and agree to accept managed care patients according to the terms of a professional services agreement with the PHO. The PHO serves as a collective negotiating and contracting unit. It is typically owned and governed jointly by a hospital and shareholder physicians.

Physician Payment Review Commission (PPRC)
Congress created the Physician Payment Review Commission in 1986 to advise it on reforms of the methods used to pay physicians under the Medicare program. The Commission has conducted analyses of physician payment issues and worked closely with the Congress to bring about comprehensive reforms in Medicare physician payment policy. Its recommendations formed the basis of 1989 legislation that created the RBRVS, a resource-based fee schedule limiting the amount physicians may charge patients.

Placebo Effect
A real or apparent improvement in the patient's condition attributable to the patient (or provider) expectation that the treatment will have an effect. The placebo effect is independent of the real effect of the intervention or treatment.

The party bringing suit in a civil court.

Plurality Opinion
An opinion lacking a majority of the judges but having more votes than any other.

Point Prevalence
Disease prevalence (total new and old cases of a disease) at a point in time. May have a different value from a period prevalence estimate.

A group of individuals for whom the results of an analysis are relevant. A population may also be thought of as the group(s) of interest which the researcher wishes to describe or draw conclusions about.

Population Attributable Risk (PAR)
The proportion of disease in the whole population that is attributable to a risk factor of concern (exposure). In other rods, PAR measures how much of the disease occurrences can be attributed to a particular risk. PAR is the incidence rate of disease in the total population minus the incidence rate of unexposed.

PAR is a measure of the potential for prevention of disease if the risk could be eliminated. As a result, PAR is often used to assess the potential impacts of public health interventions.

Population Health
1) Health outcomes and their distribution in a population. These outcomes are achieved by patterns of health determinants (such as medical care, public health, socioeconomic status, physical environment, individual behavior, and genetics) over the life course produced by policies and interventions at the individual and population levels.

2) A conceptual framework for thinking about why some populations are healthier than others, as well as the policy development, research agenda, and resource allocation that flow from it (Young 1998).

3) The health outcomes of a group of individuals,
including the distribution of such outcomes within the group (Kindig and Stoddart 2003).

4) The health of a population as measured by health status indicators and as influenced by social, economic, and physical environments; personal health practices; individual capacity and coping skills; human biology; early childhood development; and health services (Dunn and Hayes 1999).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Positive Predictive Value (PPV)
In screening and diagnostic testing, the probability that an individual has the condition of interest conditional on having a positive test.
Synonyms: Predictive Value Positive (PVP)

The philosophy that an objective reality exists apart from human perception of reality and that reality can be understood though observation and research. Positivist methods are experimental and observation, in order to maximize objectivity and a separation between observer and subject.

Posterior Distribution
An estimated probability distribution for an estimated quantity based on information (observations) from collected data combined with a prior distribution based on related data and subjective beliefs. In epidemiology, it is the probability distribution of having a disease given the patients symptoms.

Posttest Only Control Group Design
A type of experimental design in which the experimental and control groups are measured and compared after implementation of an intervention. Comparisons are made only after the intervention, since this design assumes that the two groups are equivalent other than the randomly assigned intervention. Between-group differences are used to determine treatment effects.

Potential Outcomes Framework
An approach to measuring the causal impact of an intervention by conceptualizing the outcome that would have been observed had a group of study participants already given an experimental treatment not received that treatment.

The causal impact would be estimated by comparing the results for the group who received treatment to the (hypothetical) results for the same group in the absence of treatment. The set of potential outcomes that might have resulted had this exact same treatment group not received the treatment (but instead received something else, be it no treatment, alternative treatment, etc) is called the counterfactual.

Power [FIG]
The ability of a statistical test to reject the null hypothesis when it is truly false, In other words, power is the ability of the test to detect the true relationship as a function of the parameter value under the alternative hypothesis. A statistical test is considered to have "high" power if the probability of making a type II error is low.

Practice Guidelines, Parameters
Standards used to guide healthcare providers based on accepted clinical treatment protocols for typical cases.

The making of a law by a court through judicial decisions or opinions. A case ruling that serves as a foundation or a basis for future rulings in similar matters.

The quality of being sharply defined or stated. One measure of precision is the number of distinguishable alternatives from which a measurement was selected, sometimes indicated by the number of significant digits in the measurement.
Synonyms: Accuracy

Preclinical Disease
The early stage in the natural history of a disease when apparent signs or symptoms have not yet developed.

Preclinical Phase
The testing of experimental drugs or treatments in animals before clinical trials in humans that are considered safe to carry out.

Predetermined Variable
A predetermined variable is one whose values are considered fixed in repeated samples. In causal diagrams, the values of predetermined variables are not caused by other variables in the model. Some texts distinguish between predetermined variables that are truly exogenous such as age, and predetermined variables that are determined in a previous time period, but might be treated as endogenous in a more comprehensive model.

Prediction [EX]
Using the relationships derived from a regression model to estimate the value of the dependent variable based upon the values of the independent variables.

Predictive Validity
A form of criterion validity analysis that measures the extent to which a measure/item is predictive of the criterion measure ("gold standard") in the future.

Predictive Value [EX]
The statistic generated by dividing the number of true positives by the sum of true positives and false positives.

Predictive Value Negative (PVN)
The proportion of persons who receive a "negative" test result who truly do not have the disease. PVN varies with the prevalence of the disease in the population.
Synonyms: Negative Predictive Value (NPV)

Predictive Value Positive (PVP)
The proportion of persons who receive a "positive" test result who truly have the disease. PVP varies with the prevalence of the disease in the population.
Synonyms: Positive Predictive Value (PPV)

Pretest-Posttest Control Group Design [FIG]
A study design in which subjects are assigned to either a study or a control group. The study group experiences an intervention or experiment while the control group does not. Both groups are observed at two points in time, before and after an intervention or experiment.
Synonyms: Randomized Two-Group Pretest-Posttest Only Design

Prevalence [EQ]; [FIG]
At a specific point in time, the proportion of a population that has the given disease. This is equal to the number of diseased individuals in the population divided by the number of individuals in the population at that time.
Synonyms: Prevalence, Complete

Primary Authority
Authority that comes directly from a law-making body, such as a constitution or statute, case law, administrative regulations, executive orders, and treaties. This is the opposite of secondary authority, which is material that explains or opines about primary authority (e.g. a law review article or Restatement).

Primary Case
A person who is infected with the disease from an exposure and introduces the disease into a group.

Prior Distribution
In Bayesian statistics, the distribution that reflects the researcher’s beliefs about the parameters of a model prior to the analysis. These beliefs may be shaped by results of earlier studies or by subjective judgment and may differ from one researcher to the next.

The likelihood that an event will occur.

Probit [EQ]; [FIG]
A regression model in which the dependent variable is a binary variable limited to a value of zero or one. Probit models the cumulative distribution function of the normal distribution to model the relationship between the explanatory variables and the probability that the dependent variable equals the independent variable. This function squeezes the estimated probabilities into the interval between zero and one.

Productivity Costs
The monetary value of productivity losses attributable to an illness or intervention. Includes the value of goods and services not produced due to illness (morbidity costs). In cost-of-illness analyses, but not in cost-effectiveness analyses, may also include the value of the goods and services the individual could have produced were it not for premature death (mortality costs). In cost-benefit analysis, morbidity-related productivity costs for morbidity are always included, but mortality-related productivity costs are excluded if willingness-to-pay is used to value mortality reductions.

Prediction of ways a patient's disease will progress based on an individuals' symptoms. Often used to assess the chance of recovery.

Prognostic Selection
The selection of study subjects based on prognosis, i.e. based on forecasting possible medical outcomes in the subjects.

Propensity Score
The propensity score expresses the probability of being in one group as opposed to another, based on observable characteristics prior to an intervention. Propensity scores can be obtained from a variety of discrete dependent variables such as logit or probit. Propensity scores can be included as explanatory variables in regression equations or used to match subjects with similar propensity scores.

Join a discussion on the definition and use of Propensity Scores

Proportional Hazards Modeling [EQ]; [EX]
A type of statistical model often used in survival analysis to assess the effect of explanatory variables (e.g. treatment) on survival.

A key assumption of this model is that the effect of the studied variable (e.g. treatment) has a constant proportional effect on the hazard rate at all times.

Proportionate Mortality [EX]
The proportion of deaths in a defined population attributable to a specific condition over a period of time.

Expressed as the number of deaths from a specific condition divided by total number of deaths from all causes.

Proportions [EQ]
Based on count data, a ratio in which the numerator is expressed as a subset of the denominator. The range of a proportion must be between 0 and 1.0.

In medicine, proportions are often used to describe the fraction of a population that is affected by a condition.

Prospective Analysis [FIG]
A study design in which one or more groups of individuals who are free of a particular disease of interest are monitored over time, during which some of them develop the disease. Certain characteristics of all individuals, known as exposure variables, are monitored at the beginning of the study and again at each follow-up, in order to determine whether particular exposures are risk-factors for developing the disease.

Prospective Clinical Research Study
A prospective study design in which a specific group of people are followed over time to collect clinical data throughout the group's life span.

Prospective Observational Study
A study that begins with the selection of exposed and unexposed subjects and then follows the subjects over a sufficient number of years to observe health events of interest (e.g. incidence rate or mortality rate). Considered "observational" because researchers do not randomize participants or manipulate factors in the study.

Prospective Payment
Any method of paying hospitals or other health programs in which amounts or rates of payment are established in advance for a defined period (usually a year). Institutions are paid these amounts regardless of the costs they actually incur. These systems of payment are designed to introduce a degree of constraint on charge or costs increases by setting limits on amounts paid during a future period. In some cases, such systems provide incentives for improved efficiency by sharing savings with institutions that perform at lower than anticipated costs. Prospective payment contrasts with the method of payment originally used under Medicare and Medicaid (as well as other insurance programs) where institutions were reimbursed for actual expenses incurred.

Protected Health Information (PHI)
Under the Health Insurance Portability and Accountability Act (HIPAA), PHI is individually identifiable health information transmitted by electronic media, maintained in electronic media, or transmitted or maintained in any other form or medium. PHI excludes education records covered by the Family Educational Rights and Privacy Act, as amended, 20 U.S.C. 1232g, records described at 20 U.S.C. 1232g(a)(4)(B)(iv), and employment records held by a covered entity in its role as employer.

Pseudo Randomization
Used in econometrics to refer to a mechanism, often an instrumental variable, that assigns subjects to different values of endogenous explanatory variable, but that otherwise has no direct effect on the dependent variable after controlling for other variables in the equation of interest.

Public Health
The science dealing with the protection and improvement of community health by organized community effort. Public health activities are generally those that are less amenable to being undertaken by individuals or are less effective when undertaken on an individual basis and do not typically include direct personal health services. Public health activities include: immunizations; sanitation; preventive medicine, quarantine and other disease control activities; occupational health and safety programs; assurance of the healthfulness of air, water, and food; health education; epidemiology, and others.

Publication Bias
Lack of representativeness in peer-reviewed research publications as a result of factors unrelated to the quality of the research. One example is a higher rate of publications which illustrate positive research findings (i.e. show a new treatment is beneficial compared to standard therapy).

Punitive Damages
Damages awarded in addition to Actual Damages when the defendant acted with recklessness, malice or deceit. These damages are intended to punish the defendant and thus deter future actions of the sort in question.

Purposive Sampling
Generates a sample based on expert judgment of major dimensions of a population. Purposive sampling is used if a simple random sample will not generate sufficient numbers of subjects representing the most important components of a population.

Qualitative Research
Qualitative research attempts to describe or understand complex human phenomena. Qualitative researchers seek illumination and understanding, and work to extrapolate to similar situations rather than use statistical methods to quantify phenomena. Ultimately, qualitative researchers hope to develop new concepts and theories to explain and describe the phenomena they are studying.
Synonyms: Categorical Data Analysis

Qualitative Variable
Variables that have non-numeric values such as names or labels.

Quality-Adjusted Life Year
A measure of health outcome that assigns toeach period of time a weight, ranging from 0 to 1, corresponding to the health-related quality of life during that period, in which a weight of 1 corresponds to optimal health and a weight of 0 corresponds to a health state judged equivalent to death; these are then aggregated across time periods (Gold et al. 1996).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Quality of Care
The degree to which delivered health services meet established professional standards and judgments of value to the consumer. Quality may also be seen as the degree to which actions taken or not taken maximize the probability of beneficial health outcomes and minimize risk and other outcomes, given the existing state of medical science and art. Quality is frequently described as having three dimensions: quality of input resources certification and/or training of providers), quality of the process of services delivery (the use of) appropriate procedures for a given condition), and quality of outcome of service use (actual improvement in condition or reduction of harmful effects).

Quality of Life
A broad construct reflecting a subjective or objective judgment concerning all aspects of an individual’s existence, including health, economic, political, cultural, environmental, aesthetic, and spiritual aspects (Gold, Stevenson, and Fryback 2002).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Quantiles are points that divide a statistical sample or probability distribution into ordered subgroups containing an equal number of observations. Examples of quantiles include quartiles (4 groups), quintiles (5 groups), and deciles (10 groups).

Quantitative Research
A systematic process in which numerical data are analyzed to quantify relationships between variables. The researcher aims to provide independent, unbiased analysis of the relationship between one thing (independent variable) and another (dependent variable) with the intent of generalizing the results of the analysis to the population of interest.

Quantitative Variable
Variables that have numeric values representing measurable quantities.

Quantiles that divide the distribution into four equal, consecutive groups (quarters).

Quasi-Experimental Study [EX]
A study design in which researchers manipulate an active independent variable but do not have full control over the allocation or timing of the intervention. Quasi-Experimental designs are often used when it is not possible to conduct a true experiment with complete random assignment, as is often the case in policy or real-life settings.
Synonyms: Quasi-Experimental Approach; Quasi-Experimental Design

Quantiles that divide the distribution into five equal, consecutive groups.

Quota Sampling
A type of non-probability sampling in which the population is categorized into mutually exclusive subgroups similar to strata and samples are taken from each subgroup. During sample selection, individual assignment to the sample is non-random, but is based on quotas or proportions of each stratum that must be sampled.
Synonyms: Consecutive Sampling

Random Effects Models
Models used in the analysis of longitudinal data, particularly when measures are taken repeatedly from the same subjects over time. Random effects models treat differences across individuals as random, rather than fixed. These models allow for estimation of between- and within-individual variation, by allowing both the intercept and slope coefficients to vary across individuals. Random effects models are useful when observations can be thought of as random samples from a larger population about which the researcher wishes to make inferences. The primary disadvantage of these models is that they assume that there is no correlation between the explanatory variables and the error term, which often times is unlikely to be true.

Random Error
Data fluctuations that cause measured values to differ from their true values in unsystematic and unpredictable ways. Although random error increases sampling variability (making it harder to detect a statistically significant result when one actually exists), it does not produce bias. Random error is typically caused by the inability to take precise measurements and can often be measured or reduced.

Random Sampling
Selection of a group of subjects (sample) from a larger group (population) using a process that selects each individual randomly (by chance) so that each member of the population has a known chance of being included in the sample. A strategy used to reduce the likelihood of bias.

The process of assigning individuals in an experimental study to the experimental or control group by chance. Randomization is required in experimental studies in order to ensure that the groups being studied are similar at the onset of the investigation, thereby minimizing selection bias.

Randomization is often achieved using a random number table to assign participants to the treatment or control group.

Randomized Clinical Trial [FIG]
The "gold" standard for evaluating efficacy of therapeutic, preventive and other intermediate interventions. Randomization is considered crucial to assess treatment effects because random assignment of participants to treatment or control groups allows the researcher to assume all treatment arms are comprised of subjects who are roughly equivalent. If randomization has been properly conducted, it is then possible to assume that differences in outcome are attributable to the intervention (treatment) and not participant characteristics.

A measure of the dispersion of data. The range is the difference between the smallest and largest value of a specific variable.

Rate [EX]
A measure of the intensity of the occurrence of an event. Rates are usually expressed using a standard denominator such as 1,000 or 100,000 persons. Rates may also be expressed as percentages.

Rating Scale (RS)
1) Generally, a measurement tool that individuals use to assess their preferences. Typically the tool is a line or depiction of a thermometer with one end representing best outcomes and the other end representing the worst outcomes. Individuals are asked to report their behavior, attitudes, etc. on the scale, indicating their observed preferences or feelings regarding a variety of conditions. 

2) In Cost-Effectiveness Analysis: a method of measuring an individual's preference for a health state.  A respondent is asked to assign a number between 0 and 100 (or 0 and 10, or 0 and 1) to a health stage, where 1 represents perfect health and 0 represents the state of being dead.  Rating scale preferences are technically values, not utilities, and therefore cannot be interpreted as utility measures unless the data are transformed using a conversion curve. 

Resource-based Relative Value Scale (RBRVS)
Established as part of the Omnibus Reconciliation Act of 1989, Medicare payment rules for physician services were altered by establishing an RBRVS fee schedule. This payment methodology has three components: a relative value for each procedure, a geographic adjustment factor, and a dollar conversion factor.

Reason [EX]
A type of cause that refers to a traceable or explainable process relating the cause and result or outcome.

Recursive Partitioning Analysis
A nonparametric technique used to divide a sample of study participants into smaller and smaller sub-samples based on selected predictor variables, or risk factors. With each successive split, the sub-samples become more homogeneous. Recursive partitioning analysis provides a visual summary of the data, which may be useful for identifying interactions between sub-samples that multivariate regression would fail to detect.

Reference-based Price
A method of setting prices (that a third party is willing to reimburse) based on a set of similar products. Adopted in cases where an agency wants to negotiate price, perhaps by limiting the amount a manufacturer can charge for a drug or device. One approach is to limit the price of a new drug or device to the same price as the least expensive drug in a therapeutic category.

Regression Analysis
A statistical procedure used to establish and describe the nature and magnitude of the relationship between a dependent variable (Y) and two or more independent variables (X). Often used to predict the value of a dependent variable given the specific values of independent variable or variables. The most common form of model used in regression analyses is linear. However, in epidemiology and other fields where dichotomous dependent variables are used, logistic regression is used.

Regression Coefficient
The measure of association between an independent variable and the outcome.
Synonyms: Beta Coefficient; Parameter Estimate

Regression Imputation
An imputation approach that estimates missing values based on predicted values generated by a regression model. The regression model is based on known variables for the missing unit of observation and requires that all values used as predictors are complete cases for the variables used in the regression.

Regression to the Mean
A type of threat to internal validity that refers to the tendency for observations to gravitate toward the mean over time.
Synonyms: Regression Toward the Mean

Relative Risk (RR) [EQ]
A measure of the association between exposure to a risk factor and development of a disease. A measure of relative risk indicates the number of times more or less likely an individual is to contract a disease when exposed to the risk factor versus being unexposed.
Synonyms: Risk Ratio

Relative Risk Reduction (RRR) [EQ]; [FIG]
The relative, proportional reduction in a certain risk (ratio of bad outcome) between experimental and control groups.

RRR is the difference in risks (bad outcomes) between the two groups, expressed as a proportion by dividing the absolute risk reduction (ARR) among the treated group by risks in the untreated population.  RRR expresses the extent to which a treatment reduces the risk of interest, in comparison to patients not receiving the treatment.

Relative Survival Rate [EQ]
The ratio of the observed survival rate in a patient group with a specific disease compared with the expected survival rate in a similar group (i.e. same age group) without disease. The latter is often taken as the survival rate among a general population that is equivalent to that patient group in all aspects except the disease condition.

The survival rate is determined at specific time intervals, such as survival 2 years and 5 years after diagnosis.

The extent to which an instrument, scale or other measure generates consistent results. Reliability is dependent upon the context in which an instrument is administered, rather than representing a specific property of an instrument under all conditions.

Reliability Coefficient
Statistic used to measure the consistency of a test or observations. The reliability coefficient is calculated by measuring the correlation between repeated observations on the same subjects.

Reliability coefficients range from -1 to 0 to +1.
Synonyms: Reliability

Report Card
A report presented on quality of health services designed to inform patients and health care purchasers of practitioner and organizational performance.

Residual [EQ]
In a regression analysis, the difference between the observed value of the dependent variable (y) and the predicted value (ŷ).

Residual Plot
A graphical display of information that shows the difference between the observed value of the dependent variable (y) and the predicted value (ŷ). If the model is measured correctly, residual points should be scattered randomly around the line (error=0) with a mean equal to zero.

The party against whom a motion, petition, or appeal is filed.

Response Rate
A measure of the extent to which respondents refuse to be studied, including refusing to participate in a study or refusing to respond to specific items, which is called "item nonresponse". In sample surveys, the response rate refers the proportion of individuals who completed a survey divided by the number of individuals who were selected for the sample (complete reporting units divided by the number of eligible reporting units), adjusting for individuals who were unable to be contacted.

Restatement Law
A series of volumes authored by the American Law Institute that explains the general direction of law and provides opinions about the direction in which the authors believe the law should change.

Retrospective Clinical Research Study
A research study that examines historical clinical data or analyzes the clinical data currently available in existing databases.

Retrospective Cohort Study
Most cohort studies are prospective; however, cohort studies that have reconstructed exposure data from historical records are referred to as retrospective cohort studies. In these studies, exposure and outcome data are followed up without actually following cases, which can result in considerable savings of time and money.
Synonyms: Historical Cohort

Retrospective Study
A study design that selects participants based on known outcomes of interest and then analyzes data based on participants past experiences. Past experiences are compared between those who do and do not have the outcome of interest. Retrospective studies are planned to study and collect data on events that have already occurred. In medicine, if the disease being studied takes a long time to appear, researchers tend to use the retrospective approach.
Synonyms: Retrospective Analysis

Reversal Time Series
A study design in which an intervention is introduced and withdrawn one to several times while the outcome of interest is measured repeatedly over time. A series of pretests is generally conducted to generate a baseline, after which the intervention is introduced and the outcome of interest is measured several times. The intervention is then withdrawn, and the outcome is measured several times. This process may be repeated as many times as desired.

Revised Statute
Laws that have been collected, arranged and reenacted by the legislative body.

Rich Descriptions
A framework of a study in which the descriptions of observations include a broad spectrum of categories to ensure quality in the study.

Risk Adjustment
A set of techniques used to correct for underlying differences between two or more groups that predict the differences in the observed outcome.

Risk-based Capital Formula
A method of establishing the minimum amount of capital appropriate for an insurance company to support its overall business operations in consideration of its size, structure, and risk profile. It is used to assess a managed care organization's financial viability and help prevent insolvency.

Risk-bearing Entity
An organization that assumes financial responsibility for the provision of a defined set of benefits by accepting prepayment for some or all of the cost of care. A risk-bearing entity may be an insurer, a health plan or self-funded employer, or a Physician Hospital Organization (PHO) or other form of a Patient Safety Net (PSN).

Risk Factor
Something which, if present, either increases or decreases an individual's risk of developing a disease or other outcome.

Risk Selection
Occurs when a disproportionate share of high or low users of care join a health plan.

Risk Sharing
The distribution of financial risk among parties furnishing a service. For example, if a hospital and a group of physicians from a corporation provide health care at a fixed price, a risk-sharing arrangement would entail both the hospital and the group being held liable if expenses exceed revenues.

Receiver Operating Characteristic (ROC) Curve [FIG]
An ROC curve illustrates the degree to which a test can discriminate between healthy and diseased persons.

ROC curves plot the true positives (sensitivity) compared to false positives (1-specificity) as the threshold value for the test varies. In general, the results of a diagnostic test are considered more accurate to the extent the area under the ROC curve is closer to 1.

Root Cause Analysis
A process for identifying the basic or causal factor(s) that underlie variations in performance, including the occurrence or possible occurrence of an error.

Run-in Period
A period prior to random assignment in which all qualified potential study participants collectively receive the experimental treatment or a placebo. The goal is to increase the efficacy of the study by screening out certain participants, such as those who cannot tolerate the treatment, non-compliers, or placebo responders.

Rural Health Clinics Act
Establishes a reimbursement mechanism to support the provision of primary care services in rural areas. Public Law 95-210 was enacted in 1977 and authorizes the expanded use of physician assistants, nurse practitioners, and certified nurse practitioners; extends Medicare and Medicaid reimbursement to designated clinics; and raises Medicaid reimbursement levels to those set by Medicare.

Substance Abuse and Mental Health Services Administration (SAMHSA)
The mission of SAMHSA is to provide, through the U.S. Public Health Service, a national focus for the Federal effort to promote effective strategies for the prevention and treatment of addictive and mental disorders. SAMHSA is primarily a grant making organization, promoting knowledge and scientific state-of-the-art practice. SAMHSA strives to reduce barriers to high-quality, effective programs and services for individuals who suffer from, or are at risk for, these disorders, as well as for their families and communities.

A group of individuals (or units) selected from a larger group (population) to be representative of the population. A sample is generally selected for a study because the population as a whole is too large to study.

Sample Selection Bias
Sample estimates that do not generalize to estimates for the population of interest because the study sample was not obtained from the population of interest by simple random sampling.
Synonyms: Incidental Truncation

Sampling [EX]; [FIG]
The process of selecting and observing individual units in order to make statistical inferences or make conclusions about the population.

The proportion of subjects in a population selected for analysis is called the sampling fraction.

Sampling Design
A designated plan for selecting sample participants in order to generate estimates of an underlying population parameters.

The extrapolation of a particular study's findings to a program, organization, or population that differs in size than the one being evaluated.

Scatterplot [FIG]
A visual display of information on the relationship between two quantitative variables X and Y. In a scatterplot, data is arrayed in an X-Y plane where each point represents one observation from a data set.

State Children's Health Insurance Program (SCHIP)
This program was enacted as part of the Balanced Budget Act of 1997, which established Title XXI of the Social Security Act to provide States with $24 billion in Federal funds for 1998-2002 targeting children in families with incomes up to zoo percent of the Federal poverty level.

Testing a large population in order to identify unrecognized or undiagnosed cases of a particular disease, or any risk factor of interest. A screening test is only an initial examination and is not intended to diagnose diseases or conditions.

The Second Longitudinal Study of Aging (LSOA II)
The Second Longitudinal Study of Aging (LSOA II) is used in conjunction with data from the original LSOA to determine whether the prevalence and incidence of functioning, pathology, and impairments in the elderly population have changed over 10 years, and whether the change is due to differences in cohort characteristics or to technological and medical advancements. The LSOA II is a collaborative effort between the National Center for Health Statistics (NCHS) and the National Institute on Aging (NIA), which followed a cohort of older persons through two follow-up interviews conducted in 1997-1998 and 1999-2000. The Second Supplement on Aging (SOA II), conducted in conjunction with the 1994 National Health Interview Survey (NHIS), served as the baseline for the study. Baseline in-person interviews are conducted in the respondents’ homes and are followed-up with the Computer Assisted Telephone Interview program (CATI).

The Second Supplement on Aging (SOA II)
The Second Supplement on Aging (SOA II) is an on-going, collaborative effort of the National Center for Health Statistics (NCHS) and the National Institute on Aging (NIA) to provide information about the causes of changes in health and functioning in older Americans and risk factors associated with health events. Administered by the U.S. Census Bureau, the SOA II was conducted as part of the Phase II National Health Interview Survey on Disability (NHIS-D), a follow-up supplement to the 1994 National Health Interview Survey (NHIS). The sample is nationally representative of the elderly population and comprised of non-institutionalized persons 70 years or older at the time the survey was conducted.

Secondary Attack Rate (SAR)
The attack rate among the population that come into contact with a primary case. Recording the number of new cases of a disease within an accepted incubation period following contact with the primary case, the SAR is often used to assess the transmission risk of infectious diseases. Not technically a rate but a proportion.

Secondary Authority
Authority that explains the law but does not establish it (i.e. an annotation).

Secondary Case
Individuals who acquire a disease due to exposure to a primary case.

Secondary Data
Data that were collected prior to the current analysis and perhaps for a purpose different from the current analysis.

Secondary Prevention
Activities aimed at reducing the disease prevalence and shortening duration of disease through early detection of disease and prompt, effective treatment.

Section 1115 Medicaid Waiver
"Section 1115 of the Social Security Act grants the Secretary of Health and Human Services broad authority to waive certain laws relating to Medicaid the purpose of conducting pilot, experimental or demonstration projects that are ""likely to promote the objectives"" of the program. Section 1115 demonstration waivers allow states to change provisions of their Medicaid programs, including: eligibility requirements, the scope of services available, the freedom to choose a provider, a provider's choice to participate in a plan, the method of reimbursing providers, and the statewide application of the program."

Section 1915 (b) Medicaid Waiver
Section 1915(b) waivers allow states to require Medicaid recipients to enroll in HMOs or other managed care plans in an effort to control costs. The waivers allow states to: implement a primary care case-management system; require Medicaid recipients to choose from a number of competing health plans; provide additional benefits in exchange for savings resulting from recipients' use of cost-effective providers; and limit the providers from which beneficiaries can receive non-emergency treatment. The waivers are granted for two years, with two-year renewals. Often referred to as a "freedom-of-choice waiver."

Selection Bias
The estimated treatment effect estimate does not represent just the unique effect of treatment on an outcome, because a subset of unobserved confounders non-random assignment of the subjects to the treatment and control groups resulted in the treatment effect estimate.
Synonyms: Omitted Variable Bias; Self-Selection Bias; Unmeasured Confounding; Unobserved Confounding; Confounding by (Contra-)indication; Residual Confounding; Hidden Selection Bias

Semi-Structured Interview
A data collection method in which the researcher interviews a participant based on a set of topics rather than a specific set of questions. The types of questions asked are usually similar to those in an unstructured interview consisting of open-ended questions.

The study and theory of signs, especially as a component of communication. Often in socialization participants interpret interactions as symbols. In qualitative research, semiotics is the study of these symbols and their significance.

Sensitivity [EQ]
The proportion of diseased individuals who are correctly identified by a test as having the disease.

Sensitivity Analysis
The process of changing one or more key assumptions underlying a particular analysis and determining what effect, if any, the change has on the findings. Add or dropping explanatory variables from the statistical model, changing the models functional form, and using alternative samples or statistical procedures to estimate results are all examples of sensitivity analyses.

Sentinel Event
An unexpected occurrence or variation involving death or serious physical or psychological injury, or the risk thereof. Serious injury specifically includes loss of limb or function. The event is called "sentinel" because it sends a signal or sounds a warning that requires immediate attention.

Sequential Probability Ratio Test (SPRT) [FIG]
An ongoing statistical analysis repeatedly conducted as data is collected. The data are repeatedly reassessed and a decision is made either to:
(1) Reject the null hypothesis and stop collecting data
(2) Fail to reject the null hypothesis and stop collecting data
(3) Continue collecting data until a decision regarding the null hypothesis can be reached.

The SPRT sets threshold boundaries, which take the form of parallel lines, one of which represents the expected outcome (representing the null hypothesis) and the other a significantly different outcome (representing the alternative hypothesis). When the value of the calculated test statistic falls outside of these threshold boundaries, a conclusion can be drawn and data collection stops.

Severity of Illness
A risk prediction system to correlate the seriousness of a disease in a particular patient with the statistically expected outcome (e.g., mortality, morbidity, efficiency of care). Most effectively, severity is measured at or soon after admission, before therapy is initiated, giving a measure of pretreatment risk.

A shorter version of the SF-36 (one-page, two minute) survey form that has been shown to yield summary physical and mental health outcome scores that are interchangeable with those from the SF-36 in both general and specific populations. This shorter version of the SF-36 was published in early 1995.

Shadow Pricing
Within a given employer group, pricing of premiums by HMO(s) based upon the cost of indemnity insurance coverage, rather than strict adherence to community rating or experience rating criteria.

The significance level of a statistical hypothesis. The probability of committing a type I error, or rejecting the null hypothesis if in fact it is true. Ideally, a researcher wants the significance level to be as small as possible, in order to prevent them from inadvertently rejecting the null hypothesis in favor of the alternative.
Synonyms: Alpha

Simple Random Sampling (SRS)
A specific sampling technique in which each individual in the sample is selected entirely by chance and each member of the population has an equal and independent chance of being included in the sample.

Simultaneous Equations
A set or series of equations that interact in a way that requires estimation of all related equations to produce estimates with desirable properties (e.g., unbiaseness, consistency, and efficiency). The equations may contain some of the same variables and/or the error terms in the equations may be correlated across equations.

Simultaneous Nonrandomized Design
Simultaneous comparison of two populations that are subject to different treatment conditions. Can be considered a type of cohort study since each of the treatment groups can be classified as the group "exposed" to a factor of interest (the treatment) in a cohort study.

Simultaneous Testing
In clinical practice, the process of using multiple screening tests for different diseases or conditions simultaneously, such as in a blood panel. The risk of one or more false positives increases with the number of tests administered.

Single Factor Design
A study design with only one independent factor (e.g. category) of treatment in which the factor is manipulated at multiple levels. Often used in experimental design to determine the effect of a certain treatment or intervention.

May be contrasted with factorial design, which evaluates the effects of two or more factors simultaneously.

Single-Subject Multiple-Baseline Design
A study design in which multiple behaviors are measured for each subject. The objective of the study is to identify how one behavior changes after being exposed to the intervention, where all other behaviors are held constant. Subjects are followed over time to establish a baseline for all behaviors of interest, after which time one of the behaviors is exposed to the intervention. As soon as the first behavior changes or shifts, the intervention is applied to the next behavior, and so on.

Snowball Sampling [FIG]
A type of non-probability sampling technique in which studying participants are recruited by first identifying a few subjects suitable for the research, and then asking the initial subjects to refer other qualified subjects.

Solomon Four Group Design [FIG]
An experimental design in which subjects are randomly assigned to two study groups and two control groups. Pretest measures are used for one of the study groups, and one of the control groups. Following exposure of the study groups to the intervention or experiment, posttest measures are collected on all four groups.

Spearman-Rank Correlation Coefficient
A nonparametric correlation estimate of continuous data that has been ranked. May also be a better indicator of a relationship between two variables when the relationship is non-linear.

Special Population
A class of study participants recognized as needing special protections for consent, disclosure, and confidentiality. These groups include minors, prisoners, pregnant women, human fetuses, neonates, and those unable to consent to research participation on their own.

The proportion of non-diseased individuals who are correctly identified by a test as not having the disease.

Split-Half Reliability
A form of internal consistency reliability analysis based on the correlation between randomly selected halves of a set of question on the same topic. The Spearman-Brown prophecy formula is used to compute the degree of similarity between the two subsets of responses.

A false or invalid association. In statistics, a spurious relationship is observed when an association between two variables that is assumed to be causal is shown not to be causal but instead is simply correlated due to an association between the two variables and a third unobserved factor.
Synonyms: False

Spurious Correlation [FIG]
An apparent correlation that does not reflect a true association in the variables.

Standard Deviation
A measure of the variability or dispersion of data. The standard deviation is calculated by taking the square root of the variance.

Standard Gamble (SG)
A method of measuring an individual's preference for a health state. The respondent is asked to compare a specific health state to a gamble between two outcomes, usually the chance of a specified span of life in perfect health versus the chance of immediate, painless death. The probabilities of the two outcomes are generally varied until the respondent is indifferent between the certain health state and the gamble involving perfect health and death. The probability of perfect health at the point of indifference is the standard gamble utility for the health state in question.

Stare Decisis
Latin for "to stand by things." This is the doctrine of precedent, used when it is necessary for a court to hold to a ruling in a previous case, and invoked when similar points arise.

Statistical Inference
The process of estimating the value of population parameters from observed data.

Statistical Significance
The likelihood that a test result or the association observed between the dependent and independent variable could occur by chance. Statistical significance is measured by comparing the P-value and a designated significance level.

A law passed by a legislative body (i.e., Legislature, Congress).

Statutory Law
The body of law derived from statutes rather than from constitutions, regulations, or judicial decisions.

Statutory Liability
Liability created by a statute as opposed to common or case law.

A random or probabilistic event or process.

Stochastic Regression Imputation
Estimation of missing values based on predicted values generated by a regression model plus a residual.  The regression model is based on known variables for the missing unit of observation, and includes a residual term to reflect uncertainty in the predicted value.

Mutually exclusive subgroups of the target population based on important characteristics or variables.

Stratified Randomization
In this sampling strategy the population of interest is stratified by key variables (e.g. sex, race/ethnicity) to ensure that the sample represents the correct proportion of individuals in the in the population with these characteristics. Subsequently, the sample is drawn randomly from each stratum in order to preserve proportionate representation of each stratum in the population.

Stratified Sampling [EX]
The process of dividing a population into mutually exclusive subgroups (strata) and taking random samples from each subgroup rather from the whole population. Can be used when the population being studied is heterogeneous but may be categorized into relatively homogenous subgroups by certain characteristics if samples were drawn from the population as a whole.

Structured Interview
A survey instrument in which the wording and sequence of questions asked in an interview are predetermined and identical for all subject to ensure consistency.

Substitution Imputation
A method of addressing non-response in the fieldwork stage of a survey by replacing non-responding units with alternate units that were not originally selected to be part of the sample.  Alternatively, substitution imputation may substitute plausible values for missing or incorrect data using comparable data from a previous a cycle of a survey or other external source of information.

Latin for "above" (i.e. in the above text, or a previously cited authority).

An investigation in which information is systematically collected. A population survey may be conducted by face-to-face inquiry, by self-completed questionnaires, by telephone, by postal service, or some other way. The generalizability of results depends upon the extent to which those surveyed are representative of the entire population.

Survival Analysis [FIG]
An approach to modeling the time to the occurrence of discrete outcomes that accounts for censoring. A variety of methods for conducting survival analysis exist, some of which are parametric, semi-parametric, or non-parametric.

Survival Regression
A model for an outcome measured as the duration (or length of time) that elapses before some event occurs (e.g., death). The observation of duration may be censored.
Synonyms: Survival Model; Cox Model; Cox Nonproportional Hazard Model; Event History; Hazard Model; Duration Model;

Systematic Random Sampling [EX]
A sampling technique in which the first sample is selected randomly and then subsequent samples are selected according to a simple, systematic rule -- often at a fixed interval or time period. For example, sampling every tenth individual in population.

Systematic Review
A type of review conducted in a way as to reduce bias that includes both a section on materials and analytic methods.

In the context of a literature review, a systematic review includes a formal, objective literature search based on specific inclusion and exclusion criteria.  Once literature has been selected, it is critically evaluated and synthesized.

In the context of comparative effectiveness research, a systematic review may be used to assemble information on a body of data to compare single clinical trials and where possible, arrive at a common set of conclusions about the data.

In each of these cases, a systematic review may or may not formally include a meta-analysis.

Target Population
The group of individuals (subjects) the researcher is interested in and about whom the researcher wishes to draw conclusions.

Technology Assessment
A systematic evaluation of features and or effects of technology. In the context of comparative effectiveness research, this includes identification and interpretation of data from clinical trials and other studies.

Temporal Validity
A type of external validity that refers to the ability to generalize results of a study across time.

Tertiary Prevention
Prevention activities that focus on the individual after a disease or illness has manifested itself. The goal is to reduce long-term effects and help individuals better cope with symptoms.

Test-Retest Reliability
Measures the correlation between responses given to the same questions over time.

Testing Bias
A type of threat to internal validity referring to the effect that taking a test may have on subsequent administrations of the test. Individual performance on a specific item may be impacted by repeated testing due to practice, or researcher expectations, ultimately influencing the study outcome in unintended ways that cannot be controlled by the researcher.

Time Series Design [EX]; [FIG]
A quasi-experimental research design in which periodic measurements are made on a defined group of individuals both before and after implementation of an intervention.

Time series studies are often conducted for the purpose of determining the intervention or treatment effect.
Synonyms: Longitudinal Study

Time Trade Off (TTO)
A method of measuring an individual's preference for a health state. The respondent is asked to compare a specified length of life in the health state with a shorter length of time in perfect health. The length of time in perfect health is generally varied until the respondent is indifferent between the life span in the health state and the reduced life span in perfect health. The ratio of these two life spans is the time tradeoff utility for the health state in question.

Title XIX (Medicaid)
The title of the Social Security Act that contains the principal legislative authority for the Medicaid program and therefore a common name for the program.

Title XVIII (Medicare)
The title of the Social Security Act that contains the principal legislative authority for the Medicare program and therefore a common name for the program.

TOPOFF and TOPOFF II (Top Officials)
A large-scale simulation of a terrorist event that involved high-level officials as well as law enforcement, emergency management first responders, and other non-governmental officials. The exercises were led by the Department of Justice, the Department of State, and the Federal Emergency Management Agency (FEMA) in May, 2000 and May, 2003, respectively.

A civil or legal wrong (other than a breach of contract) for which a remedy or damages may be obtained.

Research that integrates the theoretical and methodological perspectives of two or more disciplines to produce a uniquely collaborative approach to problem solving. This method is particularly relevant when a concept or method cannot be fully understood within the context of a single discipline, but requires the input of multiple disciplines in order to conceptualize and analyze a research topic.

Transitional Models
A statistical model for tracking observations over time that is capable of estimating the likelihood of changing from one outcome to another in-between follow-up periods. Transitional models assume that the probability of an observed outcome in the present time period is conditional on past responses, as well as on the other explanatory variables. Since past responses influence present outcomes, these responses function as an additional explanatory variable within the model.

A research strategy in which the researcher observes the same variable or phenomenon with multiple sources, measures, and methods. This ensures that multiple theoretical perspectives are considered, and improves the reliability and validity of the information. 

Two Stage Least Squares Regression (2SLS) [EQ]
A two-step application of the instrumental variables (IV) technique to correct for the correlation of a suspected endogenous explanatory variable with the error term in the equation of interest. In the first step, the suspected endogenous variable is regressed on all the exogenous (predetermined) variables in the model. The values of the suspected explanatory variable predicted from the first step then are used as instruments for the suspected endogenous explanatory variable in the equation of interest. Inclusion of a predicted value of an explanatory variable in the equation of interest requires adjustment to the standard errors of the regression coefficients. Most statistical software packages do this adjustment as part of the 2SLS package, but if the steps are run one at a time, the standard errors in the second stage must be corrected manually.

Two-way Analysis of Variance (ANOVA)
A way of studying the effects of two factors separately. It allows the total variance to be separated and attributed to specific sources, such as the variability between groups and the variability within groups.

Type I Error
The act of rejecting the null hypothesis when it is actually true.
Synonyms: False Positive; Alpha Error

Type II Error
The act of not rejecting the null hypothesis when it is in fact false.
Synonyms: False Negative; Beta Error

Uniform Statute
A law drafted with the intention that it will be adopted by all or at least most of the states.

The unit to which a performance measure is applied (e.g. patients, clinician, group of clinicians, institution).

Unobserved Confounder
A condition or state, not measured by the scientist, that is associated with both an explanatory variable and the dependent variable but is itself not caused by the explanatory variable, e.g. not a consequence of the treatment.
Synonyms: Unobserved Confounding Variable; Unobserved Confounding Factor; Spurious Correlation

Unstructured Interview
An interview in which the researcher asks the participant open-ended questions without explicit response choices and there is no predetermined order or plan. Unstructured interviews are used when the topic involves complex concepts so the types of questions act to lessen the ambiguity in the meaning of the responses.

The Uppsala Monitoring Center
The Uppsala Monitoring Center is located in Uppsala, Sweden, and is responsible for maintaining The World Health Organization’s (WHO) International Drug Monitoring Programme. Its primary function is collecting data about adverse drug reactions from countries around the world. Their aim is to maintain and develop useful products, services and tools to ensure quick detection and an international collaboration to protect the public from hazardous medication. In keeping with this goal, the center also aspires to develop a single, global database for drug safety information.

Stands for United States Code. The U.S. Code is the codification of the general and permanent statutory laws of the United States.

A quantitative measure of an individual's preference for a particular health state under uncertainty. Typically measured by the Standard Gamble (SG), Time Trade Off (TTO), or Rating Scale method. Utility scores are used to reflect preferences when computing a quality-adjusted life year (QALY).

Utility Scores
The conventional scale of utility measures ranges from 0.0 for "death" to 1.0 for "perfect health". A negative utility score (less than 0.0) is sometimes assigned to health states regarded as "worse than dead".

Utilization Review (UR)
Evaluation of the necessity, appropriateness, and efficiency of the use of health care services, procedures, and facilities. In a hospital, this includes review of the appropriateness of admissions, services ordered and provided, length of a stay, and discharge practices, both on a concurrent and retrospective basis. Utilization review can be done by a peer review group, or a public agency.

The extent to which an instrument accurately reflects the concepts it is intended to measure.

Variable Costs
Costs that vary depending on the amount of output produced.

Vertical Integration
Organization of production whereby one business entity controls or owns all stages of the production and distribution of goods or services.

Vicarious Liability
Liability held by an individual for the actions of another. Indirect responsibility for the actions of another (e.g., an employer's liability based upon the actions of an employee).

Vital Statistics
Statistics relating to births (natality), deaths (mortality), marriages, health, and disease (morbidity). Vital statistics for the United States re published by the National Center for Health Statistics.

The Vital Statistics Cooperative Program
The Vital Statistics Cooperative Program (VSCP) combines the registries of the federal government and individual state registries of births, deaths, marriages, divorces and fetal deaths. Since the late 1970s, the VSCP has been supported by the National Center for Health Statistics (NCHS) and is administered by the Division of Vital Statistics (DVS).

Voluntary Reporting
An error reporting system where the reporter chooses to report an error in order to prevent similar errors from occurring in the future. One theory of voluntary reporting systems is that they allow reporters to focus on a set of errors broader than just those that cause serious harm and that they help to detect system weaknesses before the occurrence of serious harm.

Happiness and meaning and self-realization (Ryan and Deci 2001).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

Life satisfaction or gratification in living (Cowen 1991).

(Taken directly from Kindig, D. "Understanding Population Health Terminology," Milbank Quarterly, Vol. 85, 2007, pp. 139-161.)

A subscription based legal research service providing access to statutes, case law materials, public records, secondary authorities, and other legal resources.

Wilcoxon Mann-Whitney Test
A nonparametric test used to estimate the nature and magnitude of differences between paired samples. Used when the distributional assumption of normality required by the t-test cannot be applied or when the data is ordinal (ranked.)
Synonyms: Mann-Whitney U test; Wilcoxon rank-sum test

Wilcoxon Signed Ranks Test
A nonparametric alternative to the t-test for estimating the nature and the magnitude of the differences between paired samples. Used to test the null hypothesis that the median difference between paired samples is zero.

The test statistic W+ is computed by calculating the differences between paired observations, ranking these differences from smallest to largest by absolute value, and summing all the signed ranks. The Z- is calculated using W+ value to determine whether the result is statistically significant.

Often used when the distributional assumption of normality cannot be applied in a study.

Willingness to Pay (WTP)
The maximum amount an individual or population would pay in order to receive a benefit, such as a reduced mortality risk.

Within Subject Design
A study design in which each participant is studied at all levels of the independent variable that researchers can manipulate. Different conditions and outcomes are then compared within each subject to determine the effect of the manipulated variable.

The within-subjects design is also called a repeated-measure design since it involves taking repeated measurements on each subject. This design is common in psychological and biomedical research.
Synonyms: Repeated-Measure Design

The written order of a court in their name of a state or other legal authority commanding either a specific act or inaction.

Writ of Certiorari
A writ issued by an appellate court direction the lower court to deliver the record of the case in review, indicating the appellate court will hear the case.

Years of Potential Life Lost (YPLL)
A measure of the relative social impact due to premature deaths due to specific diseases. YPLL is computed by estimating the additional number of years up to 65 or 75 that people would have lived had they not died from the disease before that age.