Abstract:
In developing countries, quantifying disease burden is still a challenge. Disease syndrome studies could overcome this challenge since it is inexpensive, easy to collect self-reported data, and longitudinally track disease events. Moreover, such studies in communities could estimate the occurrence and burden of disease syndromes for targeted public health interventions. This study determined the occurrence of selected disease syndromes and utilization of community health care services among communities in Suna West Sub-County. The selected study disease syndromes include influenza-like illness (ILI), gastrointestinal illness (GI), and injuries. This study adopted a prospective cohort study design with households forming the cohort. We generated 92 random points using QGIS version 3.6.1 and used them to recruit 92 households best proximal to the random points in Wasweta II ward, Suna West sub-county. From these households, 390 study participants were cluster sampled, enrolled, and followed weekly for 12 weeks through a phone call for a report of illness. Upon a report of illness during the weekly call, a visit to the affected household was made within the week and a questionnaire seeking to characterize the reported illness was administered. The illnesses were then grouped into syndromes based on the study case definition. This study yielded highly structured data necessitating a Poisson and logistic multilevel data analysis depending on the distribution of the syndrome responses. The individual and household levels were included as the random effects, while independent variables were identified at the significant level of P ≤ 0.05. The study outcome variable was the count of attained GI and injuries, which followed a Poisson distribution, and the number of attained ILI syndrome that followed a binomial distribution. Adjusting for sex and age, making a visit outside the local sub-county of residence (Odds ratio (OR) =2.7, 95% CI 1.8, 4.1) and living in a cement floored house (OR=1.9, 95% CI 1.1, 3.3) independently predicted the attainment of ILI syndrome. On the other hand, making a visit outside the local sub-county (incidence rate ratio (IRR)=3.9, 95% CI 2.3, 6.4) and the presence of stagnant water due to rain (IRR=1.9 95% CI 1.1, 3.5) predicted the attaining of GI syndrome. Additionally, the independent risk factors for the occurrence of injuries included making visits outside the local sub-county (IRR=2.2, 95% CI 1.5, 3.1) and keeping domestic animal (IRR=0.13, 95% CI 0.02, 0.72). The burdens of attaining ILI, GI syndromes, and injuries across the 12 weeks were 1.6, 1.5, and 1.2 episodes per participant, and 3.7, 2.4, and 1.5 episodes per household respectively. Across time (Level 1), the intra-class correlation coefficient (ICC) of weekly repeated measures was highest in ILI syndrome (ICC=0.82) compared to GI (0.73), and injuries (0.64). At the participant level (Level 2), clustering was highest in injuries (ICC=0.31) compared to GI (ICC=0.18) and ILI (ICC=0.04). At the household level (Level 3) the contextual or household influences were highest in ILI syndrome (ICC=0.14) compared to GI (0.08), and injuries (0.05). Disease prevention measures targeting individuals and households should be instituted to reduce ILI, GI, and injury burden. Besides, a deeper understanding of gender and age roles in determining the occurrence of ILI, GI, and injuries is needed to reduce the burden. Studies are needed to establish granular exposures associated with the increasing risks of making a visit outside the local sub-county of residence, living in a cement-floored house, and owning domestic animals. As well, there is a need to increase community awareness of risks associated with stagnant pools of water within households and support interventions using community health volunteers (CHV) in implementing community disease prevention activities to reduce the burden caused by ILI, GI, and injuries.