| Description: |
Performance-based financing (PBF) is a funding strategy that pays for outcomes rather than the cost of inputs. Verification through facility records (quantity verification) and patient interviews in communities (community verification) is a known cornerstone of PBF to ensure reported results are accurate. However, the literature suggests it’s common to tie payment to quantity verification results, which measure internal record alignment but do not assess the validity of records (e.g., whether records represent delivered services). We sought to understand the extent to which reported voluntary medical male circumcisions (VMMCs) in a PBF program could be verified in facility records and with patients, and if the two sources aligned at the facility-level. We performed a mixed method verification including quantity verification and community verification to verify reported results for Population Services International’s VMMC program in Zimbabwe from 2016 – 2018. We also interviewed verifiers to help understand the findings and we assessed the correlation between quantity and community verification performance scores at the facility-level to see whether facilities that have strong record keeping tended to also have strong validation from patients and vice versa. Among the 36,877 VMMCs selected from DHIS2 for quantity verification, 94% of records were sufficiently complete. Among records selected for community verification, only 55% (2,010/3,676) of patients were interviewed. Among those interviewed, 17% (342/2,010) provided answers that did not plausibly match the record. Verifiers reported that some patients admitted providing incorrect contact information to avoid follow-up and most verifiers suspected staff had fabricated data. We found no correlation between performance scores at the facility-level. Overall, results from the quantity verification were not a good proxy for the community verification. Programs that pay based on facility records alone risk overpaying for services and misreporting performance. To ... |