Impact of an innovative financing and payment model on tuberculosis patients’ financial burden: is tuberculosis care more affordable for the poor?

Please see Additional file 1 for translations of the abstract into the five official working languages of the United Nations

China is among the countries with the highest burden of tuberculosis (TB). According to the WHO Global Tuberculosis Report 2017, China ranked 3rd in terms of incident TB cases and 2nd on multi-drug resistant TB (MDR-TB) [1]. In order to achieve the health-related Sustainable Development Goal (SDG) proposed by the United Nations and end the TB epidemic by 2030, standard medical treatment needs to be made accessible and affordable for all TB patients [2].

Currently the TB service delivery system in China has been undergoing a transformation whereby responsibility for provision of clinical services is being gradually shifted from TB dispensaries to designated hospitals, typically general hospitals or infectious disease hospitals, in most of the provinces [3]. This transformation has had considerable cost implications, as studies have shown that, while quality of care has often been improved, income-pursuing behaviors, such as over-prescription and unnecessary hospitalization have driven up the treatment cost of TB [4–7]. Previous studies also show that although over 95% of the Chinese population is covered by the three public health insurance systems – Urban Employee Basic Medical Insurance (UEBMI), Urban Resident Basic Medical Insurance (URBMI) and New Cooperative Medical Schemes (NCMS) [8] – TB patients bear a high economic burden for medical treatment, especially low-income and rural patients [6, 9–12]. Financial factors are also frequently mentioned as a major reason for non-adherence to treatment, suspended treatment and eventual non-cure [13–17]. As a majority of TB patients live in poor households [18, 19], it seems likely that the financial protection provided by the current health insurance schemes is far from sufficient.

In response to the financial difficulties faced by patients, the China-Gates TB project, Phase II has implemented a new financing and payment model as one component of a set of interventions that aims at improving TB control in China. Before project implementation, studies from the baseline survey have found that the medical cost was relatively high in the project area [4]. The effect of NCMS on reducing catastrophic health expenditure (CHE) for TB patients was very limited [9], and the percentage of poor patients who incurred CHE was much higher than the rich [20]. The new model focuses on improving the reimbursement rate for TB patients and introducing hospital payment reforms, moving from a fee-for-service to a case-based payment approach. A treatment guideline was also launched with a standard service package, the cost of this being covered under the new model. Transportation and subsistence allowances are also provided to patients who adhere to treatment. The other components such as establishing a comprehensive TB control model, are documented elsewhere [12].

While increasing the reimbursement rate appears to have obvious potential benefits for patients, the implications of the case-based payment approach are not clear. By setting a fixed payment rate for each case, case-based payment aims to reduce non-necessary medical services by altering provider incentives. Diagnostic-related group (DRG) case-based payment mechanisms, have been implemented in numerous developed and developing countries, and in some cases have demonstrated effectiveness in cost containment [21–23]. Previous studies in China have found mixed results regarding the impact of case-based payments in terms of reducing medical costs for different types of disease, but little evidence of reduced overall inpatient expenditure [24–29]. In addition, it has been found that the impact may be compromised by problems that have arisen in project implementation, for example with hospitals excluding patients whose treatment cost is higher than the payment limit [25, 29]. For TB specifically, at least one study suggests that a case-based payment approach with no co-payment for patients may be feasible and beneficial under the NCMS [30]. However, there is a lack of evidence as to the effects in terms of cost control and patient financial burden of adopting a model combining increased reimbursement rates and case-based payment. In particular, it remains unknown whether this approach could protect the most economically vulnerable groups from financial difficulty.

At the end of the second phase of the China-Gates TB project, a team led by Duke Global Health Institute carried out an evaluation of the effects of the interventions on equity in access to/use of TB services, and on changes in financial protection for TB patients. This article aims to investigate changes in the out-of-pocket (OOP) health expenditure and financial burden of TB patients before and after the implementation of the new model, while a companion paper by Dong et al. which is published in the following issue considers changes in service utilization. It also examines the equity implications of the new model for patients from different socio-economic groups, especially those living in poverty.

This study shows that the average OOP health expenditure and financial burden on the patients increased after implementing the new financing and payment mode. National statistics show that the average outpatient cost per visit increased 13.3% and average inpatient cost per admission increased 11.1% over the intervention period. Clearly, the percentage increase following the intervention was much higher and cannot be fully explained by the general rising trend in healthcare costs [34]. As evidenced in the companion paper by Dong et al., the new model contributed to a substantially increased use of outpatient and inpatient services, especially for the lowest income group. Their hospitalization rate increased from 48.5 to 70.7%, and the number of outpatient visits from 4.6 to 5.7. From an equity prospective, the reduced differences in service use across different income groups can be seen as a positive outcome of the intervention. No one would wish to see poorer patients avoiding CHE by choosing not to access the services they need. Nevertheless, the intervention also resulted in an increased inequality in OOP payments for TB care. Clearly the model did not provide sufficient financial protection to offset the substantially increased costs resulting partly from the increase access promoted by the intervention, and both the project designers and implementers at the national and local levels may not have fully anticipated this situation.

Qualitative results from the study offer several possible reasons why poorer patients paid more out-of-pocket during the intervention period. While the new model encouraged poor patients to start and adhere to regular treatment, many of the medical services they received were not included in the standard treatment package, and were thus not reimbursed at a higher rate. The costs of these medical services were also not restricted by the case-based payment limit as defined by the project policies. Studies have shown that patients in lower socio-economic status tend to delay treatment [35], and thus they are more likely to result in more treatments for complications or comorbidities eventually, thus increasing out-of-pocket payments for services not covered by the model. Moreover, if their condition was severe, they risked being excluded from the project because physicians in these designated hospitals were concerned that their department/unit would have to incur a deficit if the treatments they provide to the patients with serious co-morbidities are paid for using the case-based payment. Alternatively, they might be transferred to higher level hospitals where they would not enjoy the benefits of the new policies. From the physician side, as their incomes are closely related to the net income of the department, they had an incentive to exclude patients with severe comorbidities from the program, or prescribe medical services outside the treatment package to increase revenue.

In light of the financial difficulty faced by the poor patients, future policies concerning the financing of TB treatment might need to consider a strategy that is more supportive. A systematic review on interventions to reduce illness and injury related financial burden shows that eliminating or largely reducing copayments in current insurance schemes is effective in reducing OOP medical payments [36]. Evidence from cash transfer programs to support TB households in low and middle income countries also shows that this strategy could increase the TB cure rate [37], and one study in Peru also finds a reduction in CHE [38]. Based on this international experience and findings from our study, improving the reimbursement rate for all TB related medical services is necessary to reduce poor TB patients’ financial burden. However, it is also critical to ensure that the costs of all these medical services are restricted by the case-based payment; otherwise there remains an incentive for physicians to pursue income by prescribing unnecessary drugs and treatments. In addition, more financial resources, for example earmarked funds from government, should be mobilized to subsidize poor patients, especially those with complications and comorbidities.

This study has several limitations. One major limitation is that we did not have a formal list of patients who were included in the program throughout their treatment and those who were excluded. Nevertheless, as the new model was required by the government to be implemented in all designated hospitals, it would seem likely that the vast majority of patients were included. It can also be argued that assessment of the model under the current health system implies that its effectiveness may have been influenced by implementation barriers known to be imbedded in that system, for example the tendency for hospitals to avoid innovations that they perceive as potential threats to income generation. We did not ask about complications and comorbidities in the survey, and we cannot exclude the possibility that the patients enrolled in treatment during the project had more severe conditions and incurred higher costs for this reason. In addition, recall bias is likely as we asked patients about the whole treatment period since the onset of symptoms.

The mean out-of-pocket health expenditure and financial burden of TB patients increased after the implementation of the new model and we would argue that the limitations mentioned above did not substantially influence this conclusion, given the degree of the observed increase. Apparently, the new financing and payment model was not successful in protecting poor patients from financial difficulty during their treatment. Future research may be required to look into strategies to provide more financial protection for poorer patients and impose effective cost control.