Our study showed that providing financial incentives to migrant TB patients was effective in improving treatment completion and reducing default rates. Our early qualitative study found that financial constraints were substantial barriers to the provision of TB services among migrant patients in Shanghai. This study also suggests that providing financial incentives reduces the access barrier for migrants and contributes to treatment completion. In the intervention district, the financial assistance provided to migrants accounted for a significant proportion of their medical costs, transportation costs, nutrition and food, and accommodation costs associated with treatment, and was especially beneficial for the poor patients. Tulsky et al. suggests that cash incentives, such as the ones used in this study, may be more effective than non-cash incentives in improving treatment adherence. However, cash incentives also have limitations. Patients may use this to purchase items such as alcohol and cigarettes, which may in turn inadvertently affect treatment results[29, 30]. However, we did not find evidence of this in our study.
Our study also found that providing the financial subsidy in four installments during the course of the treatment was effective as it appeared to motivate patients to complete their treatment. Another study found that monetary incentives were effective in improving adherence to TB treatment, however, this was limited to the first follow-up appointment in homeless individuals only. Davidson et al. found that the improved adherence to TB treatment was attributed to the financial incentives which were distributed on a progressive and cyclical basis with a bonus at the end of each cycle period. Although Davidson’s model may be a better way to motivate the completion of treatment, it is too complicated to apply in routine busy practice. In our study, the monetary incentives, working as performance rewards, were given to migrants in stages when they had to renew their drugs. This was proved to be just as effective as Davidson’s model.
Our study found that, in both districts, the majority of the migrant TB patients were living in poverty. In the intervention district, the majority of the poor patients received a financial subsidy of RMB 1,000. The study proved that our simple process of poverty evaluation is feasible and patient-friendly, which is an important step in ensuring the appropriate distribution of financial incentives[27, 28]. Our project suggests that it is feasible to engage community-based health workers, especially as they are closer to the migrant residents and have the responsibility of following up with TB patients.
In this study, 11% of the patients in the intervention district defaulted from treatment, while a large proportion of patients defaulted within one month of registration despite the incentives. This may be due partially to a delay in conducting the poverty assessment and distributing the financial incentives. This again highlights the importance of timely and appropriate administration of financial incentives for the scheme to be successful[27, 31]. However, given the higher living costs in Shanghai and the lack of family support for migrants, the financial subsidy may not be sufficient to attract patients to stay in the city to complete their treatment. Our earlier study showed that patients who did not own any property in Shanghai, and/or were unemployed because of TB, were likely to return to their hometown upon diagnosis for better family support and care. The post-reimbursement procedures of TB treatment in Shanghai and the profit-seeking behaviors of public hospitals may also add greater financial burdens for TB patients during their treatment.
Our study found that an additional cost of RMB 6,550 (US$1301) is needed in the intervention district to achieve a 1% increase of treatment completion, and an additional cost of RMB 5,240 (US$825) is needed for a 1% decrease of default rates as compared to the control district. If these financial incentives are to continue, the government will be required to provide this additional funding. However, given the financial capacity of Shanghai and other similar cities, as these costs are relatively small, they should not pose an obstacle to reducing TB transmission in migrants and local residents.
This project was designed to be embedded in existing TB services, that is, no extra staff time was required in TB clinics and CDCs. In the meantime, a similar project funded by the Global Fund to Fight Against AIDS, Tuberculosis and Malaria (GFATM) was implemented in Shanghai. This project has increased clinical hours and appointed new staff, in addition to providing financial incentives to migrant TB patients. Further economic evaluation is being conducted to compare the effects of two different ways of providing financial incentives to migrant TB patients elsewhere.
China is building a harmonious equitable society. Our study may contribute to a policy change for providing the general social protection system for migrants in Shanghai and other large cities in China. Various government sectors, such as health insurance, civil affairs, and TB control can be mobilized to replicate this program.
This study had a few limitations. Firstly, it had limited generalizability as it was conducted in one big city, resulting in a test pool of a relatively limited number of migrant TB patients. Nevertheless, the study has detected a significantly statistical difference on treatment completion rates and default rates, and therefore this may provide an insight for migrant TB control in similar settings. Secondly, nearly a quarter of the poor patients did not receive the financial subsidy in the intervention district. This may be due partially to program administration problems and to the fact that poor patients were reluctant to receive public support. Thirdly, our study was conducted in the context of the increasingly improved TB control efforts and outcomes. However, the improvement effect can be largely balanced by comparing the control and intervention districts during the same period. To our knowledge, no other policies or economic changes could be taken into account that may influence the treatment of the migrant TB patients at that time. Finally, due to the mobile nature of migrants, only around 70% of the new migrant TB patients completing treatment were able to complete the questionnaire about their financial burdens. Our study may be subject to recall bias as patient direct non-medical costs were collected from patient surveys as no other methods of data collection were available. However, patient direct medical costs in TB clinics were collected from patient charts review. Furthermore, the study was conducted shortly after patients' treatment completion to minimize recall bias.