Studies of prevalence: how a basic epidemiology concept has gained recognition in the COVID-19 pandemic [communication].

Abstract

BACKGROUND

Prevalence measures the occurrence of any health condition, exposure or other factors related to health. The experience of COVID-19, a new disease caused by SARS-CoV-2, has highlighted the importance of prevalence studies, for which issues of reporting and methodology have traditionally been neglected.

OBJECTIVE

This communication highlights key issues about risks of bias in the design and conduct of prevalence studies and in reporting them, using examples about SARS-CoV-2 and COVID-19.

SUMMARY

The two main domains of bias in prevalence studies are those related to the study population (selection bias) and the condition or risk factor being assessed (information bias). Sources of selection bias should be considered both at the time of the invitation to take part in a study and when assessing who participates and provides valid data (respondents and non-respondents). Information bias appears when there are systematic errors affecting the accuracy and reproducibility of the measurement of the condition or risk factor. Types of information bias include misclassification, observer and recall bias. When reporting prevalence studies, clear descriptions of the target population, study population, study setting and context, and clear definitions of the condition or risk factor and its measurement are essential. Without clear reporting, the risks of bias cannot be assessed properly. Bias in the findings of prevalence studies can, however, impact decision-making and the spread of disease. The concepts discussed here can be applied to the assessment of prevalence for many other conditions.

CONCLUSIONS

Efforts to strengthen methodological research and improve assessment of the risk of bias and the quality of reporting of studies of prevalence in all fields of research should continue beyond this pandemic.