Table 2 Sensitivity analyses of the simulated client (SC) and conspicuous observation (CO) surveys in Benin

1

An analysis was performed of results for the subset of health workers included in both the SC and CO surveys, to account for differences among workers in the two surveys.

2

SC data were weighted from a given health facility with caseload data from the CO survey at the same facility, to adjust for different sampling schemes (the SC survey included one observation per facility, and the CO survey included multiple observations per facility).

3

Results from both surveys were adjusted for IMCI training status of the health worker who performed the consultation, to adjust for different proportions of children seen by IMCI-trained workers in the two surveys.

4

An analysis was performed that weighted SC data with CO survey caseloads and adjusted for IMCI training status.

5

An analysis was performed that excluded the diarrhea indicators, because diarrhea was ‘hidden’ in the SC case history (i.e., SCs never spontaneously complained of diarrhea, and requiring health workers to ask made SC cases more difficult than diarrhea cases in the CO survey) and if health workers failed to identify the diarrhea, they scored poorly on seven of the 25 performance indicators (i.e., performance for diarrhea had a moderately strong influence on overall estimates of CO bias).