2. Summary of contracting out programmes reported since 2009.

Publication	Setting	Contracting model	Key messages	Study design
Alonge 2014 Ameli 2008 Arur 2010	Afghanistan, 2003‐2006/7 (post‐Taliban conflict)	Three models: 1. Province‐wide lump sum contracts; performance bonuses; an independent group monitored performance; a high degree of NGP autonomy; limited capacity building of NGP; government managed contracts 2. Monthly reimbursements made; monitoring through an international non‐profit organisation; no performance bonuses 3. 80% of Year 1 budget paid in advance; donor‐monitored NGP performance; no performance bonuses	1. Contracting out has been associated with substantial increases in use of curative care, in particular that of poor and female patients. 2. No conclusive evidence shows that any 1 model is more effective than another. 3. Linking equity goals to performance bonuses may reduce the inequity of service utilisation between the poor and the non‐poor. 4. Using service characteristics and geographical distances as planning parameters does not guarantee better resource allocation. 5. The impact of contracting out on the quality of services needs to be researched.	Contracting out was implemented as routine care.
De Costa 2014 Mohanan 2014	India, 2000‐2010	1. The government contracted private obstetricians who own hospitals to enable poor women in rural areas to deliver at these facilities. 2. Hospitals had to meet criteria related to size and emergency services. 3. Obstetricians received a fixed reimbursement per 100 deliveries. 4. The reimbursement amount had a build‐in disincentive for caesarean deliveries.	1. Institutional deliveries increased by 50%. 2. Quality of care and provider attrition need to be researched. Mohanan 2014 3. Investigators contested the success of the programme: Studies claiming programme success did not (i) address the impact of self‐selection of institutional delivery, or (ii) address inaccurate reporting from hospitals. 4. Investigators found no important changes in the probability of institutional delivery.	Contracting out was implemented as routine care.
Heard 2013	Bangladesh, 1999‐2004	1. The government contracted with an NGP or with local government to deliver basic PHC. 2. Competitive bidding for NGP contracts 3. NGPs, but not the local government, were allowed to recruit staff and set salaries and working conditions. 4. NGPs, but not the local government, procured products directly from suppliers. 5. Both NGPs and the local government were reimbursed for documented expenditures.	1. Improvement in PHC was seen in both models, but the overall quality of care was better in the NGP facilities. 2. NGP facilities provided more PHC services per capita spending. 3. Investing in PHC facilities and contracting with NGPs may improve urban health services.	Contracting out was implemented as routine care.
Kane 2010	India, 1‐year project, 2007‐2008	1. The government partnered with NGPs to improve TB case finding through including it in routine HIV prevention services. 2. 48% of NGPs had formal contracts. 3. The model was translated into national policy through a public sector‐funded TB‐HIV partnership scheme with NGPs. 4. No other details were reported.	1. TB services can be effectively integrated into HIV prevention services and can be delivered through public‐private partnerships (PPPs).	The PPP was implemented as routine care.
Mairembam 2012	India, 2008‐2012	1. PPP to attract and retain skilled health workers 2. Management functions in facilities were contracted to NGPs through a memorandum of understanding. 3. No other details were reported.	1. Improved service delivery, building maintenance, and staff availability 2. NGPs’ flexible approach in staff recruitment and creating a supportive working environment reduced staff attrition. 3. Being isolated from government‐supported functions limited access to training programmes. 4. Contracting out must happen in the context of broader government support to address isolation from government support.	The PPP was implemented as routine care.
Shet 2011	India, 2004‐2007	1. At the public‐private facility, the government provided free treatment and the private hospital provided the premises, infrastructure, and human resources. 2. No other details were reported.	1. The fully public and PPP facilities had notably better health outcomes compared with the fully private facility. 2. The fully public facility reported fewer treatment failures compared with PPP and private facilities. 3. Larger studies are required.	The PPP was implemented as routine care.
Tanzil 2014	Pakistan, 2005‐2011	1. The government outsourced administration of PHC to a semi‐autonomous government entity. 2. No other details were reported.	1. Healthcare services were better managed in contracted out facilities than in fully governmental facilities. 2. Contracting may be effective in rebuilding PHC in low‐ and middle‐income countries.	Contracting out was implemented as routine care.
Vieira 2014	Guinea Bissau, 2012‐2013	1. The government entered a PPP with an NGP to manage a national TB reference centre. 2. Government provided the drugs and electricity, and paid staff. 3. The NGP topped up salaries and provided services.	1. Since the contracting period, mortality and treatment failure were notably lower compared with during the pre‐contracting period. 2. Direct costs to patients were reduced. 3. PPP may, in the short term, increase adherence to the hospitalisation phase of intensive treatment.	The PPP was implemented as routine care.
Zaidi 2012	Pakistan, 2003‐2008	1. HIV prevention services were contracted out to NGPs through competitive bidding. 2. These were performance‐based contracts according to predefined targets. 3. Contracts were managed by the government.	1. Contracting out is inherently a political process affected by the wider policy context. 2. Rapid roll‐out in unprepared contexts can be confounded by governments’ capacity to manage it. 3. Governments should be careful that contracting out does not distance NGPs from their historical attributes. 4. Governments’ political willingness and technical capacity are key components of successful programmes.	Contracting out was implemented as routine care.

HIV: human immunodeficiency virus.

NGP: non‐governmental provider.

PHC: primary health care.

PPP: public‐private partnership

TB: tuberculosis