JAMAevidence Glossary

Terms are derived from Users' Guides to the Medical Literature: A Manual for Evidence-Based Practice, 2nd Edition and The Rational Clinical Examination: Evidence-Based Clinical Diagnosis. Updated March 2010.
Download a PDF of the glossary (185 KB). (Adobe Reader is required to open or print PDF files.)





I² statistic
The I² statistic is a test of heterogeneity. I² can be calculated from Cochrane Q (the most commonly used heterogeneity statistic) according to the formula: I² = 100% × (Cochrane Q – degrees of freedom) / Cochran Q. Any negative values of I² are considered equal to 0, so that the range of I² values is between 0% and 100%.

Inception cohort
A designated group of persons assembled at a common time early in the development of a specific clinical disorder (for example, at the time of first exposure to the putative cause or the time of initial diagnosis) and who are followed thereafter. See also Cohort study.

Incidence
Number of new cases of disease occurring during a specified period of time; expressed as a percentage of the number of people at risk during that time.

Inclusion criteria
The characteristics that define the population eligible for a study or that define the studies that will be eligible for inclusion in a systematic review.

Incorporation bias
Occurs when investigators study a diagnostic test that incorporates features of the target outcome. The result is a bias toward making the test appear more powerful in differentiating target positive from target negative than it actually is. See also Bias.

Independent association
When a variable is associated with an outcome after adjusting for multiple other potential prognostic factors (often after regression analysis), the association is an independent association.

Independent variable
The variable that is believed to cause, influence, or at least be associated with the dependent variable; in experimental research, the manipulated (intervention) variable.

Index date
The date of an important event that marks the beginning of monitoring patients for the occurrence of the outcome of interest.

Indirect costs and benefits
The impact of alternative patient management strategies on the productivity of the patient and others involved in the patient’s care.

Individual patient data meta-analysis
A meta-analysis in which individual patient data from each primary study are used to create pooled estimates. Such an approach can facilitate more accurate intention-to-treat analyses and informed subgroup analyses.

Informational redundancy
In qualitative research, the point in the analysis at which new data fail to generate new themes and new information. This is considered an appropriate stopping point for data collection in most methods and an appropriate stopping point for analysis in some methods. See also Theoretical saturation.

Informed consent
A participant’s expression (verbal or written) of willingness, after full disclosure of the risks, benefits, and other implications, to participate in a study.

Intention to treat analysis
Analyzing participant outcomes according to the group to which they were randomized, even if participants in that group did not receive the planned intervention. This principle preserves the power of randomization, thus ensuring that important known and unknown factors that influence outcomes are likely to be equally distributed across comparison groups. We do not use the term intention-to-treat analysis because of ambiguity created by patients lost to follow-up, which can cause exactly the same sort of bias as failure to adhere to the intention-to-treat principle.

Internal validity
Whether a study provides valid results depends on whether it was designed and conducted well enough that the study findings accurately represent the direction and magnitude of the underlying true effect (ie, studies that have higher internal validity have a lower likelihood of bias/systematic error).

Interobserver agreement
The degree to which a single evaluator agrees with her/his assessment at different points in time.

Interrater reliability
The extent to which a rater is able to consistently differentiate participants with higher and lower values of an underlying trait on repeated ratings over time (typically measured with an intraclass correlation).

Interview
In qualitative research, this is one of 3 basic data collection methods. It involves an interviewer asking questions to engage participants in dialogue to allow interpretation of experiences and events in the participants’ own terms. The 2 most common interviews are semistructured, detailed interviews of individuals or discussion-based interviews of several people, called focus groups. In quantitative research, a method of collecting data in which an interviewer obtains information from a participant through conversation.

Interviewer bias
Greater probing by an interviewer in one of the groups being compared, contingent on particular features of the participants. See also Bias.

Intraclass correlation coefficient
This is a measure of reproducibility that compares variance between patients to the total variance, including both between- and within-patient variance.

Intraobserver agreement
Agreement among 2 or more evaluators (observers). See also Interrater reliability.

Intussusception
The enfolding of one segment of the intestine within another.

Inverse rule of 3s
A rough rule of thumb, called the inverse rule of 3s, tells us the following: If an event occurs, on average, once every x days, we need to observe 3x days to be 95% confident of observing at least one event.

Investigator triangulation
Investigator triangulation requires more than one investigator to collect and analyze the raw data, such that the findings emerge through consensus among a team of investigators. See also Triangulation.

Isoform
Variant in the amino acid sequence of a protein.