E2041 – Introduction to Epidemiology and Environmental Health

Seminar, 14.11.2024


Comparison of cohort studies and randomized trials


Randomized trials are very expensive to conduct so they are often preceded by years of
observational research. In some instances, the results of RCTs confirm the findings from
observational studies, in others they show something different. In one notable case, hormone
replacement therapy for women after menopause, while long-term follow-up of cohort studies
suggested protective benefits of supplemental estrogen on heart disease, RCTs seemed to indicate
harm. One trial or at least one of its arms to test the effects of estrogen plus progestin hormone
replacement on women’s health, the Women’s Health Initiative, was stopped because of indications of
excess harm to women. It is important to compare the observational and experimental evidence to
understand the potential grounds for these discrepancies.


Using the following publications, compare the evidence on hormone replacement and coronary heart
disease risk. Complete the table below and answer the questions that follow.


WHI Working Group. Risks and benefits of estrogen plus progestin in healthy postmenopausal women.
JAMA 2001; 288: 321-333.


Stampfer MJ et al. Postmenopausal estrogen therapy and cardiovascular disease, ten-year follow-up
from the Nurses’ Health Study. NEJM 1991; 325: 756-62.



Detailed study analysis – following the CONSORT Clinical Trial Reporting Checklist


Study design feature

                                       Stampfer et al, 1991

                                      WHI Working Group, 2001

What is the specific hypothesis for the study?



Setting & location where the data were collected



Eligibility criteria for study participation



Intervention – for the trial, describe the intervention in each study arm



Randomization – describe how women were randomized to study arm



Blinding – describe how blinding was done in the trial and whether blinding/masking was done in the
cohort study



Outcomes – name the outcomes and describe how they were measured as well as how often



Sample size – for the trial, how was this calculated? For the cohort study, what is the sample size
on which reporting is based (find in text or tables)



Statistical analyses: what kinds of models were conducted and what confounders/covariates were
taken into account when analyzing the study data?



Length of follow-up: how long were the women followed to observe the outcomes?



Loss to follow-up: how many women were lost to follow-up or left the trial or cohort study? For the
trial, did the loss to follow-up differ by study arm? What were characteristics of people lost to
follow-up, if discussed?



What were the characteristics of the women in the study? For the trial, did the study arms differ
or were they similar on basic characteristics?



For the study outcome, what were the results in relation to hormone replacement therapy
exposure/use? Provide RR/OR/HR as well as 95% confidence interval. Interpret the main findings.



Limitations: what were the stated limitations of the study?



Generalizability: to what population groups can the results be generalized?





Questions


1.      What do you see as the main differences between the study populations?

2.      Are there important differences in how the outcomes were assessed and the length of
follow-up? What are they?