Attrition in MRC Cohort Studies
This study has three main objectives:
- to examine the methods used to maintain participation in cohort studies
- to identify best practice for reducing differential attrition
- to investigate and disseminate methodologies suitable for the analysis of cohort studies suffering from attrition under different circumstances
Work has just started on the evaluation of retention strategies in MRC funded cohort studies. This project is funded by the MRC's Population Health Sciences Research Network (PHSRN) and jointly run from the MRC Social and Public Health Sciences Unit (SPHSU) in Glasgow and the MRC Biostatistics Unit (BSU) in Cambridge.
There was a workshop in June 2006, funded by the PHSRN, entitled "Maximising Return from Cohort Studies: Prevention of Attrition and Efficient Analysis". A report on the workshop, along with several presentations, can be found here.
Background to the study
Longitudinal cohort studies follow a group of individuals over a period of time, often many years. They can be used to study the incidence of disease and health behaviours, and the aetiology of many diseases. Studies of this type have played a vital role in increasing our understanding of the aetiological mechanisms underlying population and individual differences in the incidence of disease and health behaviours, as well as health inequalities.
It is important that the individuals in the cohort sample are representative of the population that the results are to be generalised to. How well the cohort sample represents this target population plays an important part in the analyses and interpretation of results.
Study participant recruitment and retention strategies are a vital part of the cohort study process. Poor strategies may lead to both selective recruitment and attrition. For example, participants who 'volunteer', or who are more easily recruited into studies, tend to be healthier, at lower risk of adverse exposures or outcomes, and less likely to come from an under-represented population (i.e. minority or immigrant population, lower socio-economic status, etc.). Attrition is also usually selective; those at greater risk of ill-health, or who are at high risk of adverse behaviours, are more likely to drop out of a study.
These recruitment and retention issues can result in fairly homogenous study populations (largely White, English-speaking, middle class adults). This limits how generalisable the findings are, and can result in biased estimates of association, which can lead to inaccurate conclusions and over/underestimations of study effects.
The researchers involved in this study are: Cara Booker, Seeromanie Harding and Alastair Leyland who are based at the SPHSU in Glasgow, and Ian White who is based at the BSU in Cambridge.
Brueton V, Tierney J, Stenning S, Nazareth I, Meredith S, Darbyshire J, Harding S, Rait G. Strategies to reduce attrition from randomised trials [protocol]. Cochrane Database of Systematic Reviews 2011;Issue 2:Art. No.:MR000032open access