Intended for healthcare professionals

Endgames Statistical question

Cohort studies

BMJ 2010; 340 doi: https://doi.org/10.1136/bmj.c1002 (Published 25 February 2010) Cite this as: BMJ 2010;340:c1002
  1. Philip Sedgwick, senior lecturer in medical statistics
  1. 1Centre for Medical and Healthcare Education, St George’s, University of London, Tooting, London SW17 0RE
  1. p.sedgwick{at}sgul.ac.uk

    Researchers investigated whether increased body mass index was a risk factor for prostate cancer.1 Data from a prospective cohort study known as the Health Professionals Follow-Up Study was used.2 In this study, 51 529 American health professionals aged between 40 and 75 years responded to a postal questionnaire. The health professionals were recruited in 1986 and self reported that they were free of prostate cancer at the start of the study. Participants were then followed up at regular intervals until 1994.

    Which of the following statements, if any, are true?

    • a) The health professionals were studied retrospectively

    • b) The cohort was prone to the healthy entrant effect

    • c) The health professionals were selected on the basis of their BMI

    • d) The natural epidemiology of prostate cancer could be studied in this cohort

    Answers

    Answers b and d are true, whereas a and c are false.

    The group of health professionals was collectively known as a cohort. The cohort was recruited in 1986 and followed at regular intervals until 1994—the study was, therefore, prospective (answer a is false). Cohort studies that use historical information are retrospective. A group of people are selected and events that have occurred in the past are identified. However, unlike prospective studies, the quality of information collected in retrospective studies is typically subject to bias because it is dependent on the accuracy of medical records and memory recall.

    The cohort was prone to the healthy entrant effect (answer b)—they would have initially been healthier than the population of all health professionals aged between 40 and 75 years. When recruited, the health professionals were only included if they self reported that they were free of prostate cancer. This was an essential requirement in order to establish whether any increases in BMI occurred before prostate cancer (so BMI would be a potential risk factor). Therefore, the incidence of prostate cancer in the cohort in the first few years of the study would have been reduced compared with that in the full population of health professionals aged between 40 and 75 years.

    The cohort was not selected on the basis of the proposed risk factor BMI (answer c is false). The cohort consisted of health professionals aged between 40 and 75 years who responded to a postal questionnaire. Hopefully, the cohort was representative of the population from which they were sampled. The health professionals had their BMI calculated after recruitment.

    By studying the cohort prospectively, the natural epidemiology of prostate cancer in health professionals aged between 40 and 75 years could be investigated (answer d). If increased BMI was associated with prostate cancer, BMI would be considered a risk factor. It would also be possible to estimate the incidence of prostate cancer and investigate prognosis.

    Notes

    Cite this as: BMJ 2010;340:c1002

    Footnotes

    • Competing interests: None declared.

    References

    View Abstract

    Log in

    Log in through your institution

    Subscribe

    * For online subscription