A Brief Introduction to Epidemiology X Epidemiologic Research

A Brief Introduction to Epidemiology - X (Epidemiologic Research Designs: Cohort Studies) Betty C. Jung, RN, MPH, CHES BC Jung

Learning/Performance Objectives u. To develop an understanding of: – What cohort studies are – The value of such studies – The basic methodology – Pros and Cons of such studies BC Jung

Introduction Epidemiology studies the distribution of disease in a number of ways. The two major categories of epidemiological studies are: Observational and experimental studies. Most epidemiological studies are observational. BC Jung

Epidemiological Study Designs u. Observational Studies - examine associations between risk factors and outcomes (Analytical - determinants and risk of disease, and descriptive - patterns and frequency of disease) u. Intervention Studies - explore the association between interventions and outcomes. (Experimental studies or clinical trials) BC Jung

Research Designs in Analytic Epidemiology u. Ecologic Designs Cross. Sectional Study u. Case-Control Study u. Cohort Study BC Jung

Cohort Studies u“Motion Picture Studies” (Paffenbarger, 1988) u. Forward looking. The most powerful of observational studies u. Follow groups of individuals free from disease through a period of time u. Quantified with relative risk/incidence rates/attributable risk BC Jung

Examples of Cohort Studies u Framingham Heart Study u Body fat distribution and 5 -year risk of death in older women (1993) - 15 -unit increase in waist/hip circumference was associated with 60% greater relative risk of death. Waist/hip circumference ratio as a better marker than body mass index of risk of death in older women. u Vasectomy and Prostate Cancer - those who had a vasectomy and those who did not. Increase relative risk of those with vasectomy - increase risk of prostate BC Jungcancer.

Historical Cohort Studies u. Cohort formed in the past with period of follow-up ending also in the past u. Used in occupational settings were population registers (payroll records) are available u. Example: Atomic bomb blast survivors BC Jung

Value u. Gold standard for studying the association between a risk factor and outcome u. Useful for studying incidence, risk factors, natural history or prognosis u. Useful for studying multiple outcomes u. Useful for looking at multiple exposures and their interactions BC Jung

Cohort Study Design Time Direction of Inquiry Population People Without Disease Exposed Disease Not Exposed Disease No Disease BC Jung

Cohort Study Design Concurrent 1995 Define Population Retrospect 1975 Non-randomizing 2005 Exposed Non-Exposed 1985 2015 Disease No Disease 1995 BC Jung

Methodology u. Start with persons having the presumed cause (antecedent or exposure). BUT free from the effect (disease), and then wait for them to develop the effect u. Comparison group - also free from disease, but who, also DO NOT have the presumed cause BC Jung

Methodology u. Cohort - group or aggregate of persons who have presume antecedent characteristics in common and observe the development or non-development of a given health outcome u. Compare to those free of the disease or health outcome under study. Issue being at risk of repeated episode (i. e. , Stroke, antecedents may different between prestroke 1 and prestroke 2 BC Jung

Cohort Study Measures u. Cumulative Incidence - # new cases/at risk population u. Incidence Density - # new cases/at risk person-time u. Measures of association – Relative Risk – Odds Ratio BC Jung

Strength of Association Relative Risk; (Prevalence); Odds Ratio 0. 83 -1. 00 0. 67 -0. 83 0. 33 -0. 67 0. 10 -0. 33 <0. 01 None Weak Moderate Strong Approaching Infinity Source: Handler, A, Rosenberg, D. , Monahan, C. , Kennelly, J. (1998) Analytic Methods in Maternal and Child Health. p. 69. BC Jung 1. 0 -1. 2 -1. 5 -3. 0 -10. 00 >10. 0 Strength of Association

Pros u. Can study situations where randomization is not possible u. Time sequence strengthens the inference about cause (temporal relationship between exposure and outcome) u. Only way to establish population-based incidence BC Jung

Pros u. Direct measure of incidence (risk) and prognosis (natural history) u. Incidence rate is not influenced by the presence of the effect (outcome/disease) at the beginning of the study u. Magnitude of a risk factor’s effect can be quantified u. Can estimate the relative contribution of different (multiple) causes to the occurrence of the effect (disease or outcome) BC Jung

Pros u. Can count the number of prevalent cases, and new cases, as well as the number and proportion of cases that can be prevented u. Information bias is decreased (i. e. , selective recall/memory) u. Can better measure the impact of confounding BC Jung

Pros (Historical Cohort Studies) u. Easier to create the cohort u. Baseline measurements available u. Follow-up has already occurred u. Less costly and time consuming BC Jung

Cons u. Expensive u. Not good for low-incidence (rare) diseases u. Not good for chronic diseases with long latency u. Time needed to conduct these studies u. Unexpected changes to the environment can influence the association of disease and possible cause over time BC Jung

Cons u. Non-response/Migration bias – “loss to follow-up” u. Selection bias – zero time not defined (lead-time bias) u. Sampling bias u. Ascertainment/Assessment bias of outcome (can be reduced by blinding/masking) BC Jung

Cons u. Information bias – data are different (i. e. , different hospitals) – must to be comparable for exposed and unexposed u. Confounding bias u. Measurement bias - misclassification u. Analytic/Observer bias – how data are analyzed and interpreted BC Jung

Cons (Historical Cohort Studies) u. Incomplete data sets u. No control over the quality of the measurements that are available u. Incomplete control of confounding BC Jung

References u. For Internet Resources on the topics covered in this lecture, check out my Web site: uhttp: //www. bettycjung. net/ BC Jung