7/2/2015Ch 7: Cohort Studies11 Chapter 7: Observational Cohort Studies

7/2/2015Ch 7: Cohort Studies2 Ch 7 Outline 7.1 Introduction 7.2 Historical perspective 7.3 Assembling and following a cohort 7.4 Prospective, retrospective, and ambidirectional 7.5 Addressing the potential for confounding 7.6 Data analysis 7.7 Wade Hampton Frost (optional)

7/2/2015Ch 7: Cohort Studies3 Basic Cohort Design Closed population Incidence 1 Incidence 0 RR or RD Recruit study subjects Classify individual as exposed or non-exposed Follow individuals over time, ascertain outcomes Compare incidences of study outcomes Exposed individuals non-exposed individuals RR or RD

7/2/2015Ch 7: Cohort Studies4 Simple Example Note: Incidence Proportions (“risks”) are 100% in both cohorts  use incidence rates (IR) instead

7/2/2015Ch 7: Cohort Studies5 Do not lose sight of individual experiences from the beginning to the end (William Farr) Historical Perspective From the Latin cohors –“an enclosure” –a unit of the Roman army Historical notes Ramazzini studied worker health (1713) Louis studied clinical outcomes in patients (18 th century) Pinel studied mental health outcomes associated with humane treatment (18 th century) Doll & Hill studied British doctors and smokers ( present) Americans studied the causes of heart disease in Framingham Massachusetts

7/2/2015Ch 7: Cohort Studies6 Goldberger on Pellagra Joseph Goldberger (1874–1929) demonstrated nutritional basis of pellagra Pellagra initially thought to be contagious Goldberger’s cohort study of cow ownership & pellagra: Cow Ownership Pellagra + Pellagra − Cow Cow –

7/2/2015Ch 7: Cohort Studies7 British Doctors Study Survival Curves Source: 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), Begun in 1951 by Doll and Hill with a mailing of a 6 question questionnaire sent to 59,600 individual addresses. 80% of nonsmoker survived to age 70 50% heavy smokers survived to 70

7/2/2015Ch 7: Cohort Studies8 British Doctors Cohort has confirmed excess mortality in smokers for dozens of neoplastic, vascular, and respiratory diseases

7/2/2015Ch 7: Cohort Studies9 WHI Observational Cohort WHI project included both experimental and observational cohorts Observational study: 93,676 women, 40 clinical centers Recruitment period Up to 15 years of follow-up

7/2/2015Ch 7: Cohort Studies10 WHI Observational Cohort Breast CA & Analgesic Use Group Duration Of use n BreastCA cases PYs Inc. Rate per 1000 P-value trend Referent< 1 yr54, , N/A NSAIDs1–4 yr ≥5 yr 9,000 10, ,127 36, Aspirin1–4 yr ≥5 yr 5,124 6, ,231 24, Acetamin.1–4 yr ≥5 yr 2,450 4, ,608 16,

7/2/2015Ch 7: Cohort Studies11 Prospective, Retrospective, Ambidirectional proximityBased on proximity of data collection to actual events ProspectiveProspective = data collected near time of event RetrospectiveRetrospective = data are from the past (“historical data”) AmbidirectionalAmbidirectional - combination of prospective and retrospective data

7/2/2015Ch 7: Cohort Studies12 WHI analgesic / breast cancer study Proximity of Data Collection Analgesic use historyBreast cancer occurrence Initial data collection The ascertainment of the exposure was retrospective. The ascertainment of the outcome (breast cancer) was prospective. Overall, the study was ambi-directional. 12

7/2/2015Ch 7: Cohort Studies13 Example 7.4 (Retrospective Cohort) Chemical workers and bladder cancer Exposure to benzidine, β-naphthylamine, alpha-naphthylamine Historical cohort: occupational records of 4622 men in the British chemical industry, 1900 to 1950 Bladder cancer info from death certificates 1921 – 1950 Data collection: 1952 Retrospective (“historical”) cohort study Initial data collection Chemical ExposuresBladder tumors

7/2/2015Ch 7: Cohort Studies14 Multiple levels of exposure Compare rates to least exposed referent group LungCA Rate (per 100,000 person-years) RR Non-smoker (0)101.0 (ref.) Light smoker (1)525.2 Mod. smoker (2) Heavy sm. (3)

7/2/2015Ch 7: Cohort Studies Multiple levels of exposure Analgesics and Breast cancer Group Duration Of use n BreastCA cases PYs Rate per 1000 P-value trend Referent< 1 yr54, , N/A NSAIDs1–4 yr ≥5 yr 9,000 10, ,127 36, Let R 0 ≡ rate per 1000 p-yrs in referent group Let R 1 ≡ rate per 1000 p-yrs in NSAID users of 1-4 years Let R 2 ≡ rate per 1000 p-yrs in NSAID users of 5+ years RR 1 = R 1 / R 0 = 46.4 / 49.0 = 0.95 [no units] RR 2 = R 2 / R 0 = 40.5 / 49.0 = 0.83 [no units] RD 1 = R 1 − R 0 = 46.4 − 49.0 = −2.6 per 1000 p-yrs RD 2 = R 2 − R 0 = 40.5 − 49.0 = −8.5 per 1000 p-yrs 15

7/2/2015Ch 7: Cohort Studies OpenEpi.com for data analysis “Counts” menu for incidence proportions, prevalences, and case- control data “Person Time” menu for rate data inferential confidence intervals P-valuesDescriptive and inferential (confidence intervals and P-values) statistics Can be used as a learning tool 16

7/2/2015Ch 7: Cohort Studies17 Comparison of RR and RD Rates (per ) of Lung CA & CHD assoc. w/smoking SmokerNonsmokeRRRD LungCA CHD Smoking  causes more CHD Smoking  stronger relation with LungCA RR  relative effect RD  absolute effect

7/2/2015Ch 7: Cohort Studies18 Wade Hampton Frost First Professor of epidemiology in U.S. First Dean of US School of Public Health Bridged gap between infectious disease epi and chronic disease epi with TB studies (infectious disease with long latency) 1880 – 1938

7/2/2015Ch 7: Cohort Studies19 Frost’s TB Studies TB morality per 100,000 p-yrs Columns  cross-sectional rates by age (NO follow-up of individuals )

7/2/2015Ch 7: Cohort Studies20 Rows  cross-sectional rates by year (NO follow-up of individuals) Frost’s TB Studies TB mortality per 100,000 p-yrs

7/2/2015Ch 7: Cohort Studies21 Frost’s TB Studies TB morality per 100,000 p-yrs Diagonals  mimic experience of birth cohort (1870 birth cohort shaded)

7/2/2015Ch 7: Cohort Studies22 TB Study Cross-Sectional Rates Note shifting peak occurrence (marked by *) with age Not a true cohort: False impression of cohort experience.

7/2/2015Ch 7: Cohort Studies23 Frost’s Birth Cohorts True cohort perspective. Note consistent peak at in late 20s (corresponds childbearing age) Comstock, G. W. (2001). Cohort analysis: W.H. Frost's contributions to the epidemiology of tuberculosis and chronic disease. Social and Preventive Medicine, 46(1), 7-12.

7/2/2015Ch 7: Cohort Studies24

7/2/2015Ch 7: Cohort Studies25