Glossary terms for 'H'
|Hazard rate||An epidemiologic term that measures the instantaneous rate at which an outcome occurs in a population. For practical purposes, it is almost always estimated as the rate (see below) of an outcome. For example, the hazard rate for developing coronary artery disease among women ages 50 to 59 years old was estimated as 0.008 per year. |
|Hazard ratio||The ratio of the hazard rate in those exposed to a risk factor divided by the hazard rate in those who are not unexposed; almost always estimated from proportional hazards models (Cox models). For example, the hazard ratio for developing coronary artery disease was 2.0 comparing men ages 50 to 59 years with women of the same ages.|
|Heterogeneity||A situation in which the association between a predictor and an outcome is not uniform, either among different studies or among different subgroups of subjects. For example, there is substantial heterogeneity among studies that have looked at the effects of post-menopausal estrogen on mood and cognition, with some studies showing positive effects, some adverse effects, and some no effect.|
|Homogeneity||A situation in which the association between a predictor and outcome is uniform in different studies. For example, there is homogeneity among reasonably sized studies that have looked at the effects of smoking on lung cancer: All have found a substantially increased risk among smokers.|
|Hospital-based controls||In the context of a case-control study, the selection of control patients from the same hospital(s) from which the cases were chosen. For example, in her study of whether eating processed meats was associated with upper gastrointestinal cancer, the investigator used hospital-based controls selected from patients who had non-malignant gastrointestinal diseases treated at the same hospital as the case had been treated.|
|Hypothesis||A general term for a statement of belief about what the study will find. For example the study hypothesis was that chronic use of anti-epileptic medication was associated with an increased risk of oral cancer. See also null hypothesis and research hypothesis.|
Glossary material from Hulley SB et al. Designing Clinical Research, 4th ed. Philadelphia, Lippincott Williams & Wilkins, 2013.