Compared with the original benefit package in H8 towns, the new benefit package of H8SP aimed at further improving the target population's accessibility to health services and overcoming the barriers existed in the pilot project areas through extending coverage of target population, covering out-patient services and reducing the co-payment rate, etc.
In H8SP towns, the coverage of MFA was extended from 5% to 8%~11.3%, which allowed H8SP towns enroll more poor families with relatively better off economic status than that of H8. In fact, the data from ChongQing project office showed that: in Wuxi County, MFA coverage rate in H8 towns was 5.1% while in H8SP towns was 7.5%; in Qianjiang, it was 4.3% in H8 and 8.1% in H8SP respectively. Both of them didn't reach the highest extended target coverage in 2004. However, this extension had partially resulted in significant difference in characteristics of the two samples, and the potential beneficiaries of MFA were enlarged significantly in H8SP towns. In H8SP towns, physician visits among the respondents with illness in last two weeks had increased to some degree compared with H8 towns, whereas, substantial differences existed in frequency of MFA use and hospitalizations after controlling for the need and other confounding factors. The poor families in H8SP towns had made better use of MFA package and less use in- patient services than that of H8 towns.
The indicator of frequency of MFA use was a more sensitive indicator than physician visits in the last two weeks (which often subjected to the influence of acute disease), which could better reflect the overall outpatient services utilization of poor families during last year. Frequency of MFA use had much stronger linkage with the presence of chronic disease in this study.
Chronic diseases are not easily curable, cost much to treatment and bring a lot of economic burden to patients . In poor rural areas, most of the poor families primarily relied on out-
patient treatment and accepted hospitalization only as a last resort because of its high medical cost . Introducing out-patient reimbursement into H8SP benefit package played an important role in encouraging the poor families to use out-patient services when they needed. Adding out- patient reimbursement to the benefit package of H8SP towns significantly increased MFA enrollee's accessibility to the basic health services.
Findings of this study showed that poor people in H8SP had much higher frequencies of MFA use and less use of in-patient services than those in H8. The significant difference between two groups could be partially explained by the reason that poor families' frequent use of out-patient services in H8SP towns could prevent their diseases from getting worse and thus reduced their potential need for hospitalizations. That maybe also contributed to the reduction of percentage of non-use services of MFA cardholders when needed, especially of hospitalization services.
In general, the new package of H8SP had improved the accessibility of MFA enrollee to some degree. Non-use physicians among MFA cardholders with illness in last two weeks of H8SP (26%) was lower than that of H8 (34%), and also lower than the rural average level (43 %) of 2003 national health services survey .
However, percentage of non-use but ought-to-use hospitalization in H8SP (54%) and H8 (67%) were much higher than the average rural level (31 %) of 2003 . The reasons for non-use of health services showed that financial difficulty was the leading cause for their giving-up in seeking medical services when needed. Findings of other studies also indicated: financial burden of poor families was still the main barrier to their access of health services [33–36]. Analysis results of the large medical debt of MFA cardholders showed that: in H8SP towns, large amount medical debt was less likely occurred than that in H8 towns, but there is no statistical
significance between them, the debts was strongly associated with hospitalization services and presence of chronic disease, which suggested that, in spite of the extended benefit package of MFA, its financial assistance to poor MFA cardholders was still quite limited.
Although new benefit package of MFA had made great efforts to help poor families to overcome financial obstacle, Reimbursement rate of hospitalization increased from 40~70% to 60 ~ 80%, and for some special cases, services were free. But for most ordinary MFA cardholders, 20~40% co-payments for hospitalization costs still brought a huge economic burden to these poor families.
In addition, regulations of setting ceiling for reimbursement and setting limitations on disease eligibility of MFA had limited poor families to benefit more from MFA. Besides, the poor MFA cardholders with diseases in both H8SP and H8 towns often chose to stand the illness until they became deadly ill, then hospitalization had become their last but also a very expensive resort. This meant the poor families had to borrow large amount of money to afford the co-payment and additional costs for their serious disease due to the ceiling limit of MFA. So unsurprisingly, among those poor families, medical debts had strongly associated with hospitalization use although they had gotten more financial assistance from the benefit package of H8SP than that of H8 towns.
This finding indicated that: there is still space to improve the design of benefit package of MFA Scheme for the policy maker in the future to increase the utilization.
Besides the financial reason, no family member's accompaniment, transportation, and remote distance to health facilities, also played obstructive roles in the accessibility of non-users when they needed. Results of regression analysis showed that: except for type of benefit package (H8SP/H8), among the other variables associated with frequency of MFA use and hospitalization, awareness of MFA was one easily changeable factor. The more poor people learned of this favorable policy, the more likely they used the MFA services. Therefore, promoting access to health care should focus on publicity of these pro-poor policy and program among the poor enrollees.
Our study had some important limitations. The data were collected by survey and therefore subjected to respondent's error in recall. Due to lack of accurate income measuring, we could not directly identify distribution of services utilization among MFA cardholders with different income levels, and couldn't figure out if there exist unequal service uses among MFA cardholders in terms of their ability to pay for co-payment services. Some literature argued that the relatively less poor people were more likely to benefit from services than extremely poor people under the co-payment mechanism of services [8, 37, 38]. In this study, we introduced an indirect measurement on income level with availability of food for subsistence, and measured financial burden caused by services utilization with medical debt, these variables should need more academic discussion. In addition, due to lack of considering CMS insurance system, we could not able to learn about effectiveness of combination of these two healths security systems on the poor families in project areas.