Transcript John Snow
Epidemiology Kept Simple Section 11.4 Cohort Studies Gerstman Cohort Studies 1 Cohort Studies Cohort studies follow two or more cohorts over time. Incidences are calculated and compared. Select n1 exposed individuals n0 nonexposed individuals Incidence, exposed Incidence, nonexposed Compare incidences Cohort studies are nonexperimental studies with follow-up of individuals over time Gerstman Cohort Studies 2 Cohorts (Closed Populations) Recall from Chapter 6 that a cohort is a closed population. Individuals are followed over time. The experience of cohort members is monitored to determine onsets of new events. Incidences of events are compared in two or more groups. Gerstman Cohort Studies 3 Illustrative example: British Doctor’s Study This famous cohort study started in the 1950 by Doll and Hill has been followed for half a century This survival curves show that 80% of the nonsmoker cohort survived to age 70. In contrast, 50% of the > 25 cigarette / day cohort survived to age 70 Doll, R., Peto, R., Wheatley, K., Gray, R., & Sutherland, I. (1994). Mortality in relation to smoking: 40 years' observations on male British doctors. British Medical Journal, 309(6959), 901-911. Gerstman Cohort Studies 4 British Doctor’s Study The British Doctors study has discovered or confirmed more than 40 serious adverse health outcomes associated with smoking. Here is a table from one of its earliest publications: Gerstman Cohort Studies 5 Historical Illustration Wade Hampton Frost’s Birth Cohorts Wade Hampton Frost was the first Professor of epidemiology and public health in U.S. His birth cohort studies have stood the test of time in illustrating the distinction between cross-sectional and longitudinal studies These studies also provided the bridge between acute and chronic disease epidemiology We will slowly go through some of Frost’s analyses Wade Hampton Frost (1880 – 1938) Gerstman Cohort Studies 6 Illustrative Example: Cross-Sectional Rates vs. Cohort Rates TB Mortality per 100,000 by Age and Year 1880 Cross-Sectional Rates Shaded Reading down columns shows cross-sectional rates (NO follow-up of individuals, NOT cohorts) Gerstman Cohort Studies 7 Historical Illustration: Cross-Sectional vs. Cohort TB Mortality per 100,000 by Age and Year Serial Cross-Sectional Rates Shaded Reading across rows just provides cross-sectional rates in serial fashion (NO follow-up of individuals, NOT cohorts) Gerstman Cohort Studies 8 Historical Illustration: Cross-Sectional vs. Cohort TB Mortality per 100,000 by Age and Year 1870s Birth Cohort Rates Shaded Reading along the diagonals mimics the experience of a birth cohort over time. The experience of the 1870 birth cohort is shaded. Gerstman Cohort Studies 9 Historical Illustration: Cross-Sectional vs. Cohort Analysis of Cross-Sectional Rates Cross-sectional rates decreases over time Peak in young children and adults – note that the adult peak is shifting to older ages (marked by *) over time Gerstman Cohort Studies 10 Historical Illustration: Cross-Sectional vs. Cohort Analysis of Birth Cohort Rates Birth cohorts rates show decreases over time with a consistent agerelation (peak in late 20s associated with child birth) Practical implication: squelched idea of the benefit of early exposure Bridge between infectious disease and chronic disease epidemiology Gerstman Cohort Studies 11 The Framingham Study Gerstman Cohort Studies 12 Retrospective Cohorts Cohort data can be collected prospectively (as events occur over time) or retrospectively (historical records) Dyestuff workers and bladder cancer illustrative example (p. 208) Gerstman Employment records and vital statistic used to capture events from the past On-the-job exposure to aniline dyes was associated with a large increase in bladder cancer Cohort Studies 13