for the main comparison.

Contracting out compared with not contracting out for providing clinical healthcare services
Population: people who use governmental clinical health services that are contracted out to non‐governmental providers Intervention: provision of any clinical health service on behalf of the government by for‐profit and/or not‐for‐profit, non‐governmental providers Comparison: contracting out vs no contracting out
Outcomes	Net effecta	No. of studies	Certainty of the evidence (GRADE)b	Results in words	Comments
Utilisation of health services
Immunisation of children 12 to 24 months old (over a 12 month period)	Fully immunised Net effect = ‐39.4%, intervention effect P = 0.46, clustered SE = 9.0%; see Table 2 for the CI Measles Net effect = 46.5%, SE = 28.5%, 95% CI ‐9.4% to 102.4% DPT Net effect = ‐1.4%, SE = 22.9%, 95% CI ‐46.3% to 43.5% Polio Net effect = ‐7.6%, SE = 24.1%, 95% CI ‐54.8% to 39.6%	2c,d	⊕⊕⊕⊖ Moderatee	Contracting out probably makes little or no difference in immunisation uptake of children 12 to 24 months old over the previous 12 months.
Antenatal visits (over the previous 12 months)	> 2 antenatal care visits Net effect = ‐12.2 %, intervention effect P = 0.35, clustered SE = 10.0%; see Table 2 for the CI ≥ 3 antenatal care visits Net effect = 27.4%, SE = 22.2%, 95% CI ‐16.1% to 70.9%	2c,d	⊕⊕⊕⊖ Moderatee	Contracting out probably makes little or no difference in the number of women who had > 2 antenatal care visits over the previous 12 months.
Female use of contraceptives (over a 12 month period)	Net effect = ‐11.5%, intervention effect P = 0.78, clustered SE = 3.0%; see Table 2 for the CI Net effect = 1.9%, SE = 6.9%, 95% CI ‐11.6% to 15.4%	2c,d	⊕⊕⊕⊖ Moderatee	Contracting out probably makes little or no difference in female use of contraceptives over the previous 12 months.
Health outcomes
Mortality in the past year of children younger than 1 year (over a 12 month period)	Net effect = ‐4.3%, intervention effect P = 0.36, clustered SE = 3.0%; see Table 2 for the CI	1c	⊕⊕⊖⊖ Lowe,f	Contracting out may make little or no difference in the mortality of children younger than 1 year over a 12 month period.	Trial authors conclude that the sample size was too small to detect typical mortality.
Incidence of diarrhoea in children younger than 5 years (over a 12 month period)	Net effect = ‐16.2%, intervention effect P = 0.07, clustered SE = 19.0%; see Table 2 for the CI	1c	⊕⊕⊖⊖ Lowe,f	Contracting out may make little or no difference in the incidence of childhood diarrhoea over a 12 month period.
Equity in utilisation of clinical health services
Not reported in the included studies
Economic outcomes
Individual healthcare expenditures (over a 12 month period)	Net effect = $ ‐19.25 (2003 USD), intervention effect P = 0.01, clustered SE = $ 5.21; see Table 2 for the CI	1c	⊕⊕⊕⊖ Moderatee	Contracting out probably reduces individual out‐of‐pocket spending on curative care over a 12 month period.	The reduction in individuals’ healthcare expenditure is in line with the reported decrease in people visiting private healthcare providers.
Adverse effects
Not reported in the included studies.
a Calculated as the difference between the change in the intervention group and the change in the control group: Net effect = (INTpost – INTpre) – (CONTpost – CONTpre). bGRADE Working Group grades of evidence: ⊕⊕⊕⊕ High certainty: This research provides a very good indication of the likely effect. The likelihood that the effect will be substantially different* is low. ⊕⊕⊕⊖ Moderate certainty: This research provides a good indication of the likely effect. The likelihood that the effect will be substantially different* is moderate. ⊕⊕⊖⊖ Low certainty: This research provides some indication of the likely effect. However, the likelihood that it will be substantially different* is high. ⊕⊖⊖⊖ Very low certainty: This research does not provide a reliable indication of the likely effect. The likelihood that the effect will be substantially different* is very high. * Substantially different = a large enough difference that it might affect a decision. cBloom 2006 (cluster‐randomised trial). dCristia 2015 (CBA). e Downgraded by 1 for serious risk of bias. Study 1 (Bloom 2006) is at high risk of bias as baseline participant characteristics are not reported, and Study 2 (Cristia 2015) is at high risk of other bias because estimates of effects correspond with a strengthened model of the intervention compared with the initial model. f Downgraded by one for serious imprecision. The study reported treatment of the treated (ToT) estimates. Actual numbers for numerator and denominator were not provided.

DPT: diphtheria‐pertussis‐tetanus