Table 2.
Baseline facility and area-based characteristics by study arms
| Characteristics | Description | Intervention (n = 75) | Comparison (n = 75) | Difference (P-value) |
|---|---|---|---|---|
| Panel A: facility-based characteristics | ||||
| Facility ownership | =1 for public owned (%) | 84.0 | 82.7 | 1.3 (0.828) |
| Facility level of care | =1 for dispensary (%) | 70.7 | 70.7 | 0.0 (1.000) |
| Availability of facility utilities | =1 for electricity and water supply (%) | 54.7 | 52.0 | 2.7 (0.745) |
| Availability of drugs—index | Mean index (0–1) of 37 drugs [SD] | 0.61 [0.16] | 0.66 [0.12] | –0.05 (0.031) |
| Availability of drugs—subgroup | =1 for availability below the median (%) | 57.3 | 42.7 | 14.6 (0.073) |
| Baseline coverage level (deliveries) | =1 for facility below the median (%) | 53.3 | 46.7 | 6.6 (0.418) |
| Baseline coverage level (IPT2) | =1 for facility below the median (%) | 54.6 | 45.3 | 9.3 (0.256) |
| Panel B: area-based characteristics | ||||
| Wealth status index | Mean wealth index [SD] | –0.43 [1.8] | 0.32 [2.4] | –0.75 (0.028) |
| Wealth status—tercile 1 | =1 for poorest population (%) | 40.0 | 26.7 | 13.3 (0.084) |
| Wealth status—tercile 2 | =1 for middle wealth population (%) | 34.7 | 32.0 | 2.7 (0.731) |
| Wealth status—tercile 3 | =1 for least poor population (%) | 25.3 | 41.3 | –16.0 (0.038) |
| Facility location | =1 for facility in rural district (%) | 78.7 | 84.0 | –5.3 (0.405) |
Three quantiles (terciles) were used for wealth status of the facility’s catchment population; Availability of drugs include 37 drugs and analysis used a dummy variable classified based on baseline availability distribution (=1 for availability below the median/bottom half and 0, otherwise); SD, standard deviation; reference category in brackets: public (vs non-public), dispensary (vs health centre and hospital), with electricity and water supply at baseline (vs none), baseline availability of drugs below the median/in bottom half (vs top half), baseline lower performer/below the median (vs higher performer), rural (vs urban district); for distributional analyses, wealth index and drugs availability index were re-classified on each arm separately and equally to avoid the imbalance across arms at baseline.