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 December 2009.
Download a PDF of the glossary (179 KB). (Adobe Reader is required to open or print PDF files.)





Absolute difference
The absolute difference in rates of good or harmful outcomes between experimental groups (experimental event rate, or EER) and control groups (control event rate, or CER), calculated as the event rate in the experimental group minus the event rate in the control group (EER – CER). For instance, if the rate of adverse events is 20% in the control group and 10% in the treatment group, the absolute difference is 20% – 10% = 10%.

Absolute risk
The risk of an event (eg, if 10 of 100 patients have an event, the absolute risk is 10% expressed as a percentage, or 0.10 expressed as a proportion).

Absolute risk increase
The absolute arithmetic difference in rates of harmful outcomes between experimental groups (experimental event rate, or EER) and control groups (control event rate, or CER), calculated as rate of harmful outcome in experimental group minus rate of harmful outcome in control group (EER – CER). Typically used to describe a harmful exposure or intervention (eg, if the rate of adverse outcomes is 20% in treatment and 10% in control, the absolute risk increase would be 10% expressed as a percentage and 0.10 expressed as a proportion). See also Absolute risk reduction; Number needed to harm.

Absolute risk reduction
The absolute difference (risk difference) in rates of harmful outcomes between experimental groups (experimental event rate, or EER) and control groups (control event rate, or CER), calculated as the rate of harmful outcome in the control group minus the rate of harmful outcome in the experimental group (CER – EER). Typically used to describe a beneficial exposure or intervention (eg, if 20% of patients in the control group have an adverse event, as do 10% among treated patients, the ARR or risk difference would be 10% expressed as a percentage or 0.10 expressed as a proportion).

Academic detailing
A strategy for changing clinician behavior. Use of a trained person who meets with professionals in their practice settings to provide information with the intent of changing their practice. The pharmaceutical industry frequently uses this strategy, to which the term detailing is applied. Academic detailing is such an interaction initiated by an academic group or institution rather than the pharmaceutical industry.

Active alternatives
The set of diagnoses that can plausibly explain a patient’s presentation. See also Differential diagnosis.

Additive
Describes any trait that increases proportionately in expression when comparing those with no copy, 1 copy, or 2 copies of that allele, ie, those with 1 copy of the allele show more of the trait than those without, and in turn, those with 2 copies show more of the trait than those with 1 copy.

Adherence
Extent to which patients carry out health care recommendations, or the extent to which health care providers carry out the diagnostic tests, monitoring equipment, interventional requirements, and other technical specifications that define optimal patient management.

Adjusted analysis
An adjusted analysis takes into account differences in prognostic factors (or baseline characteristics) between groups that may influence the outcome. For instance, when comparing an experimental and control intervention, if the experimental group is on average older, and thus at higher risk of an adverse outcome than the control group, the analysis adjusted for age will show a larger treatment effect than the unadjusted analysis. See also Cox regression model.

Alerting systems
A strategy for changing clinician behavior. A type of computer decision support system that alerts the clinician to a circumstance that might require clinical action (eg, a system that highlights out-of-range laboratory values). See also Reminder systems.

Algorithm
An explicit description of an ordered sequence of steps with branching logic that can be applied under specific clinical circumstances. The logic of an algorithm is as follows: if a, then do x; if b, then do y; etc.

Allele
One of several variants of a gene, usually referring to a specific site within the gene.

Alpha level
The probability of erroneously concluding there is a difference between comparison groups when there is in fact no difference (type I error). Typically, investigators decide on the chance of a false-positive result they are willing to accept when they plan the sample size for a study (eg, investigators often set alpha level at 0.05).

Alvarado model
The Alvarado model is a clinical decision rule for diagnosing appendicitis. This model uses the MANTRELS mnemonic to examine 8 findings from the medical history or the examination, and the resulting score provides guidance on whether to operate in the setting of suspected appendicitis. Of 10 potential points, patients with a score of 7 or higher are considered “positive” and are more likely to have appendicitis. The various components are Migration (1 point), Anorexia-acetone (1), Nausea-vomiting (1), Tenderness in RLQ (2), Rebound pain (1), Elevation of temperature (1), Leukocytosis (2), and Shift to the left of normal WBC count (1). See Table 5-5 in The Rational Clinical Examination.

Amoss sign
In patients with severe meningeal irritation, the patient may spontaneously assume the tripod position (also called Amoss sign or Hoyne sign), sitting on the edge of the bed with the knees and hips flexed, the back arched lordotically, the neck extended, and the arms brought back to support the thorax.

Anchor
An independent standard that is itself interpretable and at least moderately correlated with the instrument being explored. The anchor is usually designed to establish a minimum important difference (MID) in change.

Arthralgia
Joint pain.

Atelectasis
Loss of air in all or part of the lung (collapse).

Autocorrelation
Occurs when the likelihood of an observation is not independent of its relationship with other observations. For example, autocorrelation occurs when a good day for a patient with chronic disease is more likely to follow a "good day" than a "bad day."