Glossary terms for 'D'
|Data||A plural noun used to describe measurements, usually in numeric format. (The singular of data is datum.) For example, data concerning the prevalence of various diseases are useful when making decisions about allocation of health care resources.|
|Dependent variable||See outcome variable.|
|Descriptive study||A study that does not look for associations, test hypotheses, or make comparisons. For example, the investigator performed a descriptive study of the prevalence of obesity among pre-school children. See also analytic study.|
|Diagnostic test study||A study that looks at whether the results of a medical procedure are useful in assessing the likelihood of a particular diagnosis in a patient. For example, a diagnostic test study was designed to determine whether serum bicarbonate levels were useful in the diagnosis of sepsis among patients with fever.|
|Dichotomous variable||A variable that can have only one of two values, such as yes/no or male/female. For example, the examiner dichotomized systolic blood pressure into hypertensive (= 140 mm Hg) or not. See also continuous variable, nominal variable, and ordinal variable. |
|Differential bias||A general term for the situation in which a measurement varies systematically by the status of the subject, usually by whether or not the subject is case or a control; most commonly occurs with recalled exposures. For example, the investigators showed that cases of adult celiac disease were more likely to recall childhood exposures to wheat-containing products as children than their non-identical twins who had grown up in the same household, the investigators suspected that there was differential recall bias. See also non-differential bias.|
|Discrete variable||A type of variable that takes on only integer values. For practical purposes, continuous variables are sometimes treated as discrete variables.. For example, age is usually expressed as age in years at last birthday, and current smoking as average number of cigarettes smoked per day. See also continuous variable.|
|Dose-response||The phenomenon by which the greater the exposure (dose), the greater the magnitude or likelihood of the outcome (response). (If an exposure is protective, then the greater the exposure, the lower the likelihood of the outcome.) For example, one study reported a dose-response relation between sun exposure and numbers of melanocytic nevi; another reported a dose-response relation between numbers of nevi and risk of melanoma. |
|Double-cohort study||A study design in which subjects are enrolled into one of two distinct cohorts, often by occupation. For example, a double-cohort study was used to compare the risks of contact dermatitis of the hands, as well as fungal infections of the feet, among potters versus dancers.|
|Dropout||A study subject in whom outcome status cannot be ascertained, often because she refused follow-up. Sometimes includes subjects who drop-out because they died during the study. For example, there were 17 drop-outs: 8 due to refusal, 6 due to death, and 3 because of the development of dementia.|
Glossary material from Hulley SB et al. Designing Clinical Research, 4th ed. Philadelphia, Lippincott Williams & Wilkins, 2013.