Difference-in-Differences Estimates of Program Effect on Phase II PNC Client-Reported Outcomes.

*** p** p* pNotes: p-values generated using unpaired t-tests assuming unequal variance. “2010” p-values indicate probability of differences between the intervention vs. comparison mean scores at the 2010 data collection time point; “post-rollout” p-values compare the intervention vs. comparison mean sores at the 2012 data collection time point. Difference-in-differences estimates are reported as incidence rate ratios. Robust standard errors are clustered at the health facility level. Columns (1)–(3) report difference-in-difference estimates of program effect under 3 model specifications. Covariates in models (2) and (3) include categorical variables for facility type, facility sector, and client socioeconomic status quintile.Difference-in-Differences Estimates of Program Effect on Phase II PNC Client-Reported Outcomes.