Step 1 | Health facility admissions in 3 regions of Tanzania over the period of 12Â months were monitored prospectively from Jan 2006 to December 2006 using intensified surveillance to generate population-based estimates for the incidence of APP, stratified by circumstances of poisoning (occupational, accidental, suicide, homicide and unknown circumstances). These generated baseline APP rates for further adjustment. |
Step 2 | Data from a household survey of farmers was compared to records in health facilities to generate a ratio of occupational APP cases reported at health facilities, reflecting the extent of under-reporting of APP cases due to occupational circumstances. The 95% confidence interval for this ratio was used to generate a high and low boundary for the estimate. |
Step 3 | Cases of APP in the prospective study for which circumstances were unknown were allocated in a sensitivity analysis to each of the other four known circumstancesa, and the contribution of newly allocated occupational APP adjusted for under-reporting. |
Step 4 | The rate of occupational APP recorded at health facilities was adjusted to account for the under-reporting of occupational APP identified in Step 2, including high and low boundary estimates, and to account for reallocation of unknown to known circumstances as identified in in Step 3. |
Step 5 | Total APP incidence rates, summing across all circumstances, were generated under the different contingencies identified. The contribution of occupational poisoning as a proportion to overall APP incidence was described. |