Parameter estimates for each process in the cascade and decision-tree models.

The values are stratified by whether the data were collected before or after the introduction of WHO diagnostic policy recommending universal diagnostic-led treatment for malaria. Values specific to a Tanzanian case study are also shown. The median and interquartile range from the published studies is presented. For the probability of seeking treatment at the public sector clinic, diagnostic sensitivity and diagnostic specificity, separate values for Tanzania were not available and so the general parameters were used. The probability of fever being due to malaria was assumed the same in the aggregated analysis but set to reflect the reduction in malaria incidence seen in Tanzania. In the Tanzanian case study, the probability of at least one dose of ACT being in stock was used rather than the probability of all doses of ACT being in stock due to limited data on the latter.