Walson, Judd LBrander, Rebecca2019-10-152019-10-152019-10-152019Brander_washington_0250E_20652.pdfhttp://hdl.handle.net/1773/44812Thesis (Ph.D.)--University of Washington, 2019Children under age 5 in sub-Saharan Africa suffer a disproportionately high burden of infections disease, and the consequences of these conditions extend beyond the period during which the child is acutely ill. Children remain at high risk of mortality in the time period following a severe infectious disease, and, in the case of diarrhea, linear growth faltering. Interventions are needed to address these adverse outcomes following acute illness, but little is known about which children are at high risk, whether antibiotics may be effective, or the relative cost-effectiveness of various antibiotic administration strategies. We identified risk and predictive factors of linear growth faltering following moderate-to-severe diarrheal disease, and evaluated whether children who were exposed to antibiotics at diarrhea presentation had lower risks of linear growth faltering than children who were unexposed. Further, we compared two methods for collected patient-level hospitalization cost data and evaluated the comparative cost-effectiveness of mass distribution of azithromycin vs targeted azithromycin strategies. Using data from the Global Enteric Multicenter Study of children 0-59 moths old in 7 low- and middle-income countries in Africa and Asia presenting with moderate-to-severe diarrhea, we used linear regression to identify clinical and sociodemographic factors associated with loss in length-for-age z-score (LAZ) in the 50-90 days following presentation with moderate-to-severe diarrhea, and poisson regression with robust standard errors to identify factors associated with severe linear growth faltering (loss of ≥ 0.5 LAZ in the study period). Young age, nutritional status (low weight-for-length z-score, or high length-for-age z-score at presentation), high socioeconomic status, and severity of disease (hospitalization, presentation with fever, comorbidities, or general danger signs) identified children at high risk of linear growth faltering. These populations may benefit from diarrhea management interventions address post-diarrhea linear growth faltering. To evaluate the effects of antibiotics during moderate-to-severe diarrhea on linear growth, we used linear regression to estimate associations between antibiotic exposure (any antibiotic given or prescribed at diarrhea presentation) and linear growth faltering, using propensity score adjustment for factors associated with likelihood of receiving antibiotics. After propensity score adjustment, children who received antibiotics lost 0.04 less LAZ than those who did not (95% confidence interval: 0.01, 0.07) and were 20% less likely to experience severe linear growth faltering (adjusted odds ratio: 0.80 [0.69, 0.94]). Antibiotic management may offer modest protection against linear growth faltering in a sub-set of high risk children, but clinical trial evidence will be needed and the benefits should be weighed against the consequences. To evaluate the completeness of medical record documentation for the purposes of costing, we collected resource utilization data on children 1-59 months old hospitalized in a public hospital in western Kenya two different ways: by direct observation and medical record abstraction. Only 38% of children had medical records that completely documented all resources that were received. Micro-costing by medical record abstraction may slightly underestimate costs, but researchers should select the data collection method that best fits the goals and budget of the project Finally, we constructed a decision tree model to estimate the cost-effectiveness of several azithromycin strategies for preventing mortality: mass drug administration (MDA) of azithromycin to children 1-59 months old, MDA to children 1-5 months old, and azithromycin administered at hospital discharge to children recently hospitalized for any infectious condition. MDA to children 1-59 months old would cost approximately $14/disability-adjusted-life-year (DALY) averted, MDA to children 1-5 months old would cost approximately $5/DALY averted, and post-discharge azithromycin would cost approximately $3/DALY averted. All azithromycin strategies would be highly cost-effective for preventing mortality, but targeting azithromycin to a high mortality population would be even more cost-effective.application/pdfen-USnoneantibioticchild mortalityEconomic evaluationGlobal healthMalnutritionPediatricsEpidemiologyEpidemiologyThesis