Summary of findings for the main comparison. Supervision versus no supervision to improve the quality of primary health care.
| Supervision versus no supervision to improve the quality of primary health care | ||||||
| Patient or population: providers Settings: low‐ and middle‐income countries Intervention: supervision Comparison: no supervision | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect (95% CI) | No of Participants (studies) | Quality of the evidence (GRADE) | Comments | |
| Assumed risk | Corresponding risk | |||||
| No supervision | Supervision | |||||
| Providers practice | See comment | See comment | Not estimable | 134 (2 studies1) | ⊕⊕⊝⊝ low2,3,4 | 2 RCTs: both studies showed small benefits in provider practice (prescribing & family planning practices) with supervision5 |
| Providers knowledge | See comment | See comment | Not estimable | 134 (2 studies1) | ⊕⊝⊝⊝ very low2,4,6 | 2 RCTs: one showed small benefits in provider knowledge (family planning) with supervision whilst one study (prescribing knowledge) was inconclusive7 |
| Drug supply | Study population | Not estimable | 0 (1 study8,9) | ⊕⊝⊝⊝ very low10 | 1 CBA: study showed small benefit in drug stock management with supervision8 | |
| See comment | See comment | |||||
| Moderate | ||||||
| *The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; RR: Risk ratio; | ||||||
| GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. | ||||||
1Stanback 2007: A cluster randomised study of family planning services in Kenya: 6 of 13 training areas were randomly assigned to intervention (training plus supervision) or control (training alone). Within each training area family planning providers were randomly selected proportional to size of area. Data were obtained from 177 providers and 482 clients (before) and 176 providers and 451 clients (after 9 to 10 months). Kafle 1995: A cluster randomised study of primary healthcare workers in Nepal: 6 districts were randomly allocated to intervention (supervision visits by district health officers) or control (no intervention). Data was obtained from 21 intervention health facilities and 21 control facilities at baseline and after 7 months. A third study was not included in this profile because it was an observational study. 2 Serious study limitations: Both studies were cluster randomised (and took clustering into account when analysing their results). The baseline outcome measurements of intervention and control groups were similar although the studies did not report on the baseline characteristics for the comparison groups. Neither of the studies blinded the outcome assessor. 3 No serious inconsistency: Both studies found supervision produced small improvements in practice. 4 Serious imprecision: These studies contained a small number of clusters in the intervention and control groups and further used multi‐stage sampling from the cluster areas to select health facilities and posts. 5Stanback 2007 found the overall provider practice sore increased by 22.8% in the intervention group compared to 16.1% in the control group. This difference is reported as statistically significant (P = 0.004). Kafle 1995 found that 3 out of 13 indicators of good prescribing practice were higher in the intervention group following the intervention (P < 0.05) 6 Serious Inconsistency: Stanback 2007 showed a small benefit in knowledge scores in the intervention group compared with the control whereas Kafle 1995 did not demonstrate any significant differences between the intervention and control groups. 7Stanback 2007 found the overall knowledge and attitude sore increased by 19.9% in the intervention group compared with 12.6% in the control group. This difference is reported as statistically significant (P = 0.002). Kafle 1995 conducted a knowledge questionnaire following the intervention and the post intervention scores were inconclusive: knowledge scores in 3 of the 19 indicators higher in the intervention group, 13 were higher in the control group and 3 were the same. No significance values presented. 8Trap 2001 found significant difference in drug stock management score. Following the intervention, the score increased by 7% and decreased by 7% in the control group. The reasons for this decrease are not clear. 9Trap 2001 is a controlled before and after study of primary healthcare workers in Zimbabwe; 24 health facilities were included in each of the intervention and control groups respectively. The intervention group received two supervisory visits three months apart on drug stock management versus no intervention. Outcomes were assessed 6 to 8 months after the last supervisory visit. 10 No serious inconsistency: Only one study and therefore not applicable.