Background: Socioeconomic inequality including wealth distribution is a barrier to implementation of health policies. Wealth distribution can be measured effectively using household data on durable assets. Compared to other methods of analysing Socio-economic Status (SES) using durable assets, Multiple Correspondence Analysis (MCA) can create more reliable wealth quintiles. We therefore evaluated socioeconomic determinants of Schistosoma mansoni using MCA on household data among adult population in western Kenya. The hypothesis of this study was that MCA would be a useful predictor of S. mansoni prevalence and/or intensity.
Methodology: Twelve villages, 6 villages that had showed the greatest decrease in S. mansoni prevalence (Responder villages) and 6 villages that showed relatively lower decrease (Hotspot villages) between the year 2011 and 2015 were randomly selected for this study. This was according to a previous Schistosomiasis Consortium for Operational Research and Elimination (SCORE) report from western Kenya. From each village, convenience sampling was used to identify 50 adults from 50 households for inclusion in this study. An interview with a questionnaire based upon MCA indicators was conducted. One stool sample from each of the 600 adults was examined based on four slides for S. mansoni eggs using Kato Katz technique. Mean Eggs per gram(EPG) was calculated by taking the average of the readings from the four slides. A log binomial regression model was used to identify the influence of the various age-groups(<30 years, 30-60 years and >60 years), household size, wealth class, occupation, education status, main water supply, sex and sub-county of residence on S. mansoni infection. EPG was then compared across variables that were significant based on multivariate log binomial model analysis using a mixed model.
Principal findings: Overall prevalence of S. mansoni was 41.3%. Significantly higher prevalence of S. mansoni were associated with males, those aged below 30 years, those who use unsafe water sources (unprotected wells, lakes and rivers), residents of Rachuonyo North, Hotspot villages and those earning livelihood from fishing. Only sex and household size were significant predictors in the multivariate model. Males were associated with significantly higher prevalence compared to the females (aPR = 1.37; 95% CI = 1.14-1.66). In addition, households with at least four persons had higher prevalence compared to those with less than four (aPR = 1.29; 95% CI = 1.03-1.61). However, there was no difference in prevalence between the wealth classes(broadly divided into poor and less poor categories). Intensity of infection (Mean EPG)was also significantly higher among males, younger age group, Rachuonyo North residents and Hotspot Villages.
Conclusion: Socioeconomic status based on an MCA model was not a contributing factor to S. mansoni prevalence and/or intensity possibly because the study populations were not sufficiently dissimilar. The use of convenience sampling to identify participants could also have contributed to the lack of significant findings.