Estimates of Prevalence of Cognitive Impairment From Research Studies Can Be Affected by Selection Bias
The authors of the commentary discuss cognitive impairment related to HIV infection. They state that research samples can be biased in estimating the prevalence of HIV-associated cognitive impairment because patients must consent to participate in studies. If researchers don’t carefully document the source population, selection bias may occur. The authors conducted a study to examine the impact of selection bias in their Positive Brain Health Now Cohort, which recruited participants from five Canadians. They found that many eligible patients refused to participate (64%), with the common reason being lack of time. Refusers tended to be working, younger, female, and less likely to report cognitive difficulties. By accounting for these differences, they estimated the adjusted prevalence of cognitive impairment in the full sample, which was much lower than the prevalence when these differences were not accounted for. The study highlights the importance of documenting key variables for those who refuse to participate in research to ensure accurate estimates of prevalence.