Table 3. Effects of the interventions on the primary outcome compared with control, by facility type.
| Outcome | Study arm | Clusters † | Individual-level prevalence | Cluster-level prevalence | Stratified risk difference ‡ | ||||
|---|---|---|---|---|---|---|---|---|---|
| Public | Private | Public | Private | Public | Private | Public | Private | ||
| n | n | n/N (%) | n/N (%) | Mean (SD) | Mean (SD) | RD(95% CI) | RD(95% CI) | ||
| Febrile patients tested for malaria | Control | 6 | 6 | 118/229 (51.5%) | 314/1307 (24.0%) | 50.0% (32) | 17.9% (15) | 0 | 0 |
| Provider | 10 | 7 | 230/325 (70.8%) | 186/1507 (12.3%) | 69.2% (21) | 16.9% (12) | 22.7%(-5.6, 51.0) | 0.5%(-15.2, 16.2) | |
| Provider-school | 7 | 7 | 122/213 (57.3%) | 109/1283 (8.5%) | 61.8% (28) | 11.1% (12) | 13.1%(-16.7, 42.9) | -6.1%(-21.5, 9.3) | |
| Test positive patients receiving ACT | Control | 6 | 6 | 45/66 (68%) | 193/254 (76%) | 61.9% (34) | 66.5% (23) | 0 | 0 |
| Provider | 10 | 7 | 40/73 (55%) | 55/69 (80%) | 57.3% (36) | 51.3% (35) | -8.3%(-46.7, 30.0) | 3.2%(-33.7, 40.2) | |
| Provider-school | 7 | 7 | 16/37 (43%) | 43/61 (70%) | 48.6% (32) | 64.3% (36) | -13.4%(-55.2, 28.5) | -2.2%(-34.6, 30.3) | |
| Test negative patients receiving an antimalarial | Control | 6 | 6 | 18/52 (35%) | 33/60 (55%) | 41.1% (41) | 73.0% (29) | 0 | 0 |
| Provider | 10 | 7 | 27/157 (17%) | 67/117 (57%) | 24.7% (32) | 33.0% (33) | -14.0%(-54.7, 26.7) | -37.6%(-83.3, 8.3) | |
| Provider-school | 7 | 7 | 21/85 (25%) | 40/48 (83%) | 26.5% (38) | 63.7% (41) | -13.8%(-56.7, 29.1) | -9.6%(-56.0, 36.7) | |
| Treatment according to malaria guidelines | Control | 6 | 6 | 79/229 (34%) | 220/1307 (17%) | 33.3% (27) | 11.8% (11) | 0 | 0 |
| Provider | 10 | 7 | 170/325 (52%) | 105/1507 (7%) | 52.3% (28) | 10.0% (8) | 20.1%(-10.5, 50.8) | -0.7%(-12.3, 11.0) | |
| Provider-school | 7 | 7 | 80/213 (38%) | 51/1283 (4%) | 40.7% (26) | 6.4% (9) | 7.8%(-24.6, 40.1) | -4.9%(-16.3, 6.5) | |
TABLE NOTES
†A cluster contributes to the analysis of public facilities if it has at least one such facility, and similarly for private. Clusters with at least one of each type contribute to both analyses.
‡The stratified risk differences were calculated as before, but separately for public and private facilities.