Despite nearly universal access to CBHI in Rwanda, up to a quarter of patients in our study still experienced CHE when paying for the direct and indirect costs of this essential surgical care. Eliminating the medical cost of care is not sufficient to alleviate this burden to patients as it is the addition of transportation, food, and other indirect costs that cause financial catastrophe for many patients. Even modest nonmedical direct and indirect costs such as the median of ~$30 for a c-section are nontrivial for poor patients. Additional informal payments are noted in some African contexts21, and these would certainly increase this OOP expenditure and rates of CHE. However, given strongly enforced anti-corruption policies throughout Rwanda, it was felt by hospital leadership that such payments are rare, and hence this was not assessed in our study. There are, however, other nonmonetary costs even when patients are able to find the finances for surgical care: nearly half of patients had to borrow money from family or friends to pay their bill, thereby incurring an informal debt they would have to repay35, and 12% had to sell possessions.
Though we found disappointingly high rates of CHE, these rates would be exponentially higher if there was no insurance at all. Our model showed that those with subsidized care had lower rates of CHE. Even if medical costs and transportation are fully subsidized there will still be a low rate of CHE (3-4%) for impoverished patients. Not surprisingly, though, we found that when a higher percentage of costs are covered by insurance or by the government, far fewer patients are in danger of CHE. Consequentially, in the current system, the poorest patients who are fully subsidized with free care are better protected, while those in the next tier are more vulnerable to financial catastrophe. Notably, for all groups, c-section is not categorized as impoverishing. This is largely because across the Ubudehe categories, these women and their families are already classified as poor and most are extremely poor, hence are not considered impoverished by the expense according to the definitions used. Despite the introduction of Mutuelles which reduced rates of OOPs and CHE, inequalities in the population have not been reduced36. Thus, further strategies are needed to target the challenges of accessing care for those living in poverty. Informal insurance structures have been shown to be of particular utility in reducing CHE in sub-Saharan Africa37. In Ghana and Ethiopia, rates of CHE are also much higher in uninsured patients, and CBHI in particular was found to decrease rates of CHE by up to 23.2%38,39. However, studies have also shown inequalities in who chooses to access CBHI with the rich utilizing it more and dropout rates being highest among those required to pay a premium for services40. This dynamic has also been anecdotally reported in Rwanda where patients may have felt they were being taxed by the compulsory enrollment and voluntary enrollment with more community level governance could be benficial11.
Our findings suggest that there is need for greater financial protection for impoverished households in order to achieve the best medical and social outcomes for patients requiring hospital services and surgical care. In the Democratic Republic of Congo, 16% of women experienced CHE related to obstetric and neonatal care, particularly if there were any complications, while in Ghana, substantial rates of CHE were found due to transportation and indirect costs, particularly among patients with complications, despite free maternal health care41,42. These issues extend beyond maternal health and c-sections, with studies from Malawi and Uganda demonstrating high rates of CHE for other surgical conditions despite free surgeries or no user fees43–46. A study by Mercy Ships (which provides free surgery to patients) demonstrated that paying for transportation decreased the no-show rate for surgery by 45%47. Policy makers could consider such initiatives as well as others that either offset the cost of transportation or decrease the need for transportation for these patients. To avoid compounding the financial risk in the post-operative period and obtain the best outcomes, these strategies might include vouchers to offset the cost of post-operative follow-up, home-based follow-up utilizing community health workers (CHWs), empowering CHWs with mobile applications on smartphones to facilitate SSI detection and other concerns to increase the specificity of need for in-person post-operative follow up22.
Importantly, we note that true rates of CHE may be underestimated if patients choose not to have surgery due to associated costs. This is less likely for c-section patients in Rwanda, for whom surgery is usually urgent or emergent. However, when extrapolating to other types of operations including those considered elective initially or definitively, some patients may lack access to care based on financial capacity to pay. Poverty in Rwanda is correlated with lower health care services coverage48, which suggests that despite financial protection for essential services, overall access to medical care for uncovered services is still threatened. The rate reported here is likely an underestimate as we did not include post-discharge costs or the cost of care for complications. Our group, in a previous study, found that the cost of travel from home-to-health center was a significant predictor of surgical site infection49, potentially suggesting that these expenses are prohibiting the necessary follow-up care.
This study had several key limitations. First, it was only conducted at a single site, which may have local geographic features, though culturally and socioeconomically the population does resemble other parts of rural Rwanda in terms of care seeking behavior and resources. Rwanda also has a unique political context, particularly the CBHI program, and so the results and ensuing policy recommendations may not fully generalize to other LMICs. However, the general lessons learned can still help inform policy development in other countries. Second, this study may not capture any patients who did not seek hospital care due to inability to pay for services. That number is likely very small based on the high penetration of CBHI membership in the community around Kirehe. Furthermore, this only looks at the rates of CHE for one essential procedure which is of value to an entire household, therefore the rates may not represent the financial risk of other surgical procedures which may be considered elective by either the family or government.
Another known limitation of this type of data collection is that the calculated household expenditure depends on patient memory of their regular expenses and therefore suffers from recall bias. Interestingly, when patients were asked their total monthly expenditure, the amount was generally higher than the sum of its components. Furthermore, in a rural population, households may not have regular expenses but rather have occasional larger purchases related to agriculture or home maintenance. Therefore, estimating daily expenditure may not be a consistent measure of a patient’s true resources. Finally, indirect costs are likely underestimated since lost wages also include recovery time at home post-operatively.