Estimating Hidden Population Size of COVID-19 using Respondent-Driven Sampling Method - A Systematic Review

Currently, the ongoing COVID-19 pandemic is posing a challenge to health systems worldwide. Unfortunately, the true number of infections is underestimated due to the existence of a vast number of asymptomatic infected individual's proportion. Detecting the actual number of COVID-19-affected patients is critical in order to treat and prevent it. Sampling of such populations, so-called hidden or hard-to-reach populations, is not possible using conventional sampling methods. The objective of this research is to estimate the hidden population size of COVID-19 by using respondent-driven sampling (RDS) methods.

This study is a systematic review. We have searched online databases of PubMed, Web of Science, Scopus, Embase, and Cochrane to identify English articles published from the beginning of December 2019 to December 2022 using purpose-related keywords. The complete texts of the final chosen articles were thoroughly reviewed, and the significant findings are condensed and presented in the table.

Of the 7 included articles, all were conducted to estimate the actual extent of COVID-19 prevalence in their region and provide a mathematical model to estimate the asymptomatic and undetected cases of COVID-19 amid the pandemic. Two studies stated that the prevalence of COVID-19 in their sample population was 2.6% and 2.4% in Sierra Leone and Austria, respectively. In addition, four studies stated that the actual numbers of infected cases in their sample population were significantly higher, ranging from two to 50 times higher than the recorded reports.

In general, our study illustrates the efficacy of RDS in the estimation of undetected asymptomatic cases with high cost-effectiveness due to its relatively trouble-free and low-cost methods of sampling the population. This method would be valuable in probable future epidemics.