Who Benefits from Removing User Fees for Health Care? Evidence from a Randomised Experiment in Ghana

Abstract

The extent to which removing user fees for health care in developing countries improves population health rests, in part, on how behavioural responses vary across individuals with different health needs. Using data from a randomised experiment of free care in Ghana and a measure of baseline health that is both objective and broad-based, we examined differential effects for initially ill and healthy children. We find that free care improved health seeking behaviour, lowered out-of-pocket spending and reduced anaemia amongst children who were ill at baseline but had no effect on initially healthy children. Whilst there was no health effect on the intervention population taken overall, the evidence suggests that removing user fees may have enabled individuals with the greatest health need to take up primary health care, thereby improving their health. There was no indication that free care encouraged frivolous use of services.

The Effect of Insurance Enrollment on Maternal and Child Health Care Utilization: The Case of Ghana

“…The results of this study are consistent with Asuming's (2013) findings, which suggested that poorer households have larger unmet needs for health services and are more responsive to interventions. Similarly Powell-Jackson et al (2014) found that the effect of free care on the use of primary care becomes larger as the sample is restricted to increasingly sicker (anemic) children.…”

“…Finally, 61 percent of insurance members seek formal medical treatment for their sick children, whereas only 33 percent of nonmembers do so. Yet the children of noninsured mothers do not seem to suffer from a fever or a cough at a different rate, which could be an indication of the absence of moral hazard, as also seen in Powell-Jackson et al (2014).…”

Can we trust measures of healthcare utilization from household surveys?

“…In total, we have endline data for 2,319 children. Further details of the study – its design, the main outcomes and the findings – are available elsewhere [ 10 , 11 ].…”