Using Markov Chain Monte Carlo Models to Estimate the Severity, Duration and Cost of a Salmonellosis Outbreak of Known Size
Water utilities need to include the costs of waterborne disease and how much consumers would be willing to pay to prevent them when making cost-benefit decisions about proposed municipal water system improvements. Two Markov Chain Monte Carlo models were developed to estimate the severity and duration of illness and resultant costs of Salmonella outbreaks. Transition probabilities were estimated by conducting a meta-analysis of disease severity. 26.4% of salmonellosis patients see a physician. 12.5% of patients who see a physician are hospitalized, and 3.2% of hospitalized patients die. For patients in nursing homes infected with Salmonella, 10.6% of all patients were hospitalized and 31.1% of hospitalized patients died. Estimates of medical costs were obtained from reported physician charges, medication costs and hospitalization charges. Productivity losses due to Salmonella illnesses were calculated from median household income and the duration of illness determined by the model. Surveys of consumer willingness to pay for quality-adjusted life years and mortality reductions were used to estimate the cost of premature death. These models were then tested against the study of the 1993 Salmonella typhimurium outbreak in Gideon, Missouri by Angulo et al. (1997). For 131 patients in the study, the models predicted 94-97 would recover on their own, 30-33 would recover after seeing a physician, 4-5 would see a physician, and 0.14-0.18 patients would die. These results match those found by the Angulo et al. The models were then used to calculate a cumulative probability distribution of possible costs of the Gideon outbreak. For the reported 7 deaths among 650 cases, the model predicts an economic cost of $31.9 to $32.3 million.
School:University of Cincinnati
School Location:USA - Ohio
Source Type:Master's Thesis
Keywords:salmonella markov chain model monte carlo waterborne disease cost estimate
Date of Publication:01/01/2008