The authors included in the review all the studies assessing the uptake of health-insurance programmes and their impact on financial coverage, health-services utilisation and health outcomes, using credible methods of causal identification (RCTs, Quasi-experiments, controlled-before versus after panel data and instrumental variables-based studies). It focused on government-run (or community-run but government-sponsored) health-insurance systems that seek to offer financial protection for people facing health shocks. The study attempted to identify all relevant studies, regardless of language or publication status.
The authors conducted a systematic search within published and unpublished literature in a wide range of electronic databases, including ISI Web of Knowledge, ECONLIT and IDEAS. The authors also checked citations found in articles, searched the websites of relevant organisations, including the World Bank and the World Health Organization, and contacted subject experts for additional studies. Finally, the authors synthesised the results in a narrative format and discussed explanatory factors and the potential biases arising from different identification strategies.
Theauthors systematically review the evidence from 34 studies in East Asia and Pacific, South Asia, CIS, Sub-Saharan Africa and Latin America and the Caribbean. Most insurance schemes covered in the review required no premium payment from beneficiaries but charged some co-payment at the point of use. Out of the 34 studies, 19 were considered sufficiently methodologically strong to be of relevance to policy decision makers.
The authors find that enrolment rates in health insurance are frequently lower than expected and seem to be related to the perception of the programme, and to the education levels and cultural factors of the participants, rather than to structural factors such as distance to health-care services or initial state of health. The results also suggest that modifications in the health-insurance payment scheme affect the effectiveness of the insurance. Some evidence points to a positive impact of health insurance on health-service utilisation in terms of outpatient visits and hospitalisation. However, because of methodological inconsistencies between studies and an ambiguous pattern of results in terms of the effectiveness of health insurance for health-service utilisation and health outcomes, no clear policy conclusion can be drawn. Results suggest that health insurance can contribute to improved financial protection, in terms of reduced catastrophic spending, although many studies do not find any effects in terms of reduction in out-of-pocket payments among insured people. From those studies which report effects on health care spending by poorer groups, the authors found that the impact appeared to be smaller for the poorer population than other groups.
The systematic review is based on comprehensive literature searches and appropriate methods to reduce risk of bias in terms of study selection and synthesis methods. However, it has some limitations. It is not clear whether some of the included studies have a unit-of-analysis error problem, and, if this is the case, the review does not report whether the authors effectively address this problem. The authors also present study findings by 'vote counting' as opposed to more preferable methods of synthesising effect sizes which take into account both magnitude of findings (policy significance) as well as sample size (statistical credibility).