• 1 World Bank, “Annual GDP Growth (%),” http://worldbankorg/indicator/NY.GDPMKPT.KD.ZG?locations=CN (...)

1The economic opening of China that Deng Xiaoping called for in the late 1970s contributed to strong and sustained economic growth (the average increase in GDP approached a regular 10% annually between 1980 and 20141), allowing a quadrupling of income per capita from 1990 to 2006, and a doubling between 2006 and 2014. This rate of income growth was accompanied by a spectacular reduction in poverty and an overall improvement in health, which was nonetheless not as marked as in the Maoist period when progress had been very impressive. For example, life expectancy at birth rose from 35 to 65 years between 1949 and 1975, but “only” rose from 67 to 75 years between 1981 and 2014. Over the latter period, other countries performed better, with lower rates of GDP growth. These developments were however accompanied by large increases in social inequality. In the field of health provision, the rural population’s access to care came up against a real financial barrier right from the early years of the establishment of “market socialism.”

• 2 46% in 2014. World Bank, “Rural Population (% of Total Population),” http://worldbankorg/indicator/ (...)

2In the 1960s, the poor health of the rural population and the bias towards the cities to the detriment of the rural areas led the government to launch a huge healthcare programme based on the short‑term training of “barefoot doctors” and the establishment of rural medical cooperative programmes. The nationwide emphasis on basic healthcare enabled China, with its essentially rural population (74% in 19902), to obtain health indicators well above expectations for its level of income per capita.

• 3 China Health Statistics Yearbook, 2005.
• 4 Martine Audibert, Yong He, and Jacky Mathonnat, “Two-Period Comparisons of Healthcare Demand with I (...)

3However, from the mid-1980s onwards, as a result of several factors such as the ageing population, the withdrawal of the state from the health sector, and the lower efficacy and efficiency of the health system, the gap between China and the middle income countries narrowed. Between 1990 and 2004, the share of public expenditure to total health expenditure fell from 25% to 17%, whereas the share of out‑of‑pocket expenditure by Chinese households rose from 36% to 53%.3 By studying the access to healthcare by the rural population of China over the periods’ 1989-1993 and 2004-2006, based on CHNS data, we have shown that consumption by the poorest households was significantly affected by the financial burden falling on them, thereby widening the inequality gap.4

• 5 Jacky Mathonnat, Martine Audibert, Xiezhe Huangfu, Anning Ma, and Laurène Petitfour, “Progrès et dé (...)

4Being aware of this situation and the underlying discontent of the population, since 2003 the Chinese government has undertaken a deep reform of health policy and of the health system. It will take place through several successive stages,5 along three main axes: the development of health insurance and movement towards universal coverage (UHC), the reform of the hospital sector, and the restructuring of drug policies. Thus, one can see a willingness to improve not only the quality of care and the effectiveness of the health system, but also for several years now, and especially since 2009, its efficiency. Indeed, in many areas and at different levels of the health pyramid, the inefficiency of the system is a matter of concern, even if the concept of efficiency does not necessarily appear to be a goal unanimously shared by all the stakeholders, nor always clearly understood at the local authority level.

• 6 The township is the first level of administration just above the village. Next comes the county (xi (...)

5This article is concerned with the effects of the vast field of reforms already mentioned: health insurance in rural areas (the “New Rural Cooperative Medical System - NRCMS), hospital and pharmaceutical policies, and the resources, performance, and efficiency of the township hospitals (xi-angzhen weishengyuan6). Township hospitals constitute the basic level of the Chinese healthcare system. In Chinese terminology, they form part of the first level in the health pyramid.

• 7 Catherine Cazals, Jean-Pierre Florens, and Léopold Simar, “Nonparametric Frontier Estimation a Robu (...)
• 8 Rowena Jacobs, Peter Smith, and Andrew Street, Measuring Efficiency in Healthcare, New York, Cambri (...)

10The technical efficiency of the township hospitals was assessed by using non-parametric Data Envelopment Analysis (DEA) to study the effects of health insurance in rural areas, and order‑m7 partial frontier efficiency for the study of the hospital and pharmaceutical reforms. The assumption of the variable returns to scale instead of constant returns to scale is retained, as it is more relevant when the size of the hospitals and the managerial capabilities8 are taken into account (for further methodological details, see Annex 1).

• 9 Arousing strong discontent in the population: IBM Business Consulting Services, “Healthcare in Chin (...)

18In order to compensate for their reduced resources after the withdrawal of the state from the health sector, health facilities had to resort to charging their users, if not to balance their accounts then at least to maintain a manageable level of deficit. Gradually a profit motive emerged in those establishments, leading the cost of care to rise faster than per capita income. Moreover, relying on market-inspired elements in the Chinese economic reforms led to the closure of many state enterprises and the disappearance of agricultural co‑operatives in favour of the family responsibility system, which in turn brought on the collapse of the social security scheme. By the end of the 1980s, less than 5% of the rural population were covered by health insurance, against over 95% in the late 1970s. The households were faced with an expensive health system, well beyond the reach of the poorest and severely degraded in terms of quality of care, which was especially noticeable in rural areas (in village health stations as well as township hospitals).9

• 10 Martine Audibert, Xiaoxian Huang, Jacky Mathonnat, Aurore Pélissier, Ningshan Chen, and Anning Ma, (...)

21The analysis of our data shows that the development of the NRMCs has greatly increased TH activities, both curative and preventive.10 The total number of outpatients in the 24 THs in the study remained stagnant between 2000 and 2005, but rose sharply (+77%) between 2005 and 2008, increasing from 650,000 to 1,115,000. The same applies to inpatients, whose number was relatively stable at the beginning of the period, but more than doubled, from 25,000 to more than 50,000 after 2006, when every district had fully introduced the health insurance measures. The effect of the insurance scheme shown on these indicators was confirmed by an econometric analysis and by a double‑difference impact analysis. Three indicators were selected to represent health insurance and its penetration: a binary variable to show whether health insurance was introduced in the township, the proportion of the township population covered, and the reimbursement rate of hospital care for external consultations and hospitalisations. Whichever indicator was selected, the results of the econometric analysis show that the introduction of NRCMS had a beneficial effect on the volume of activity undertaken by the hospitals. This effect was more marked for inpatients (44%) than for external consultations (16%).This result matches the expectations of the authorities, whose efforts on reimbursement were more focused on hospitalisations than on external consultations. The effect on outpatient visits was greater in the poor townships than in the rich ones, which is also what the authorities wished to achieve. But this difference in favour of the poor has not been observed for inpatients.

• 11 Winnie Yip and William Hsiao, “The Chinese Health System at a Crossroads,” Health Affairs, Vol. 27, (...)

25Although there was already an observed increase in preventive activity at the beginning of the period, this trend consolidated considerably after 2005, showing that the directives for public health adopted at that level of the pyramid were swiftly implemented (Graph 2). However, detailed analysis per hospital displays a mixed outcome, with some hospitals showing no change in their activities while there was a vigorous increase in activity in others. Our results are in line with those from other studies that confirm that the introduction of the new scheme health insurance has led to an overall increase in hospital activity by facilitating access to care.11

• 12 Aurore Pélissier, Martine Audibert, Jacky Mathonnat, Xiaoxian Huang, and Anning Ma, “Curative Activ (...)

26Contrary to our observations on activity, the results show that the NRCMS had no effect, or even a negative effect, on the technical efficiency of the THs, regardless of the type of activity (curative or preventive).The average efficiency score, which was already relatively low before the establishment of the NRCMS (for example, taken altogether the hospitals would have been able to increase their average curative activity by 30% and their preventive activity by 40% without additional resources), fell further with each sub-period: after the implementation of the insurance system, hospitals would have been able to increase their curative activity by an average of 50% and their preventive activity by 65% without additional resources.12 Moreover, as Graph 3 shows, the number of efficient hospitals fell over the period, while there was a rise in the number of relatively very inefficient ones (those with a score of less than 60%).

• 13 Kimberly Barbiaz, Grant Miller, Hongmei Yi, Linxiu Zhang, and Scott Rozelle, “New Evidence on the I (...)

28The main characteristics of the effects of the NRCMS that we have observed and assessed on the basis of our work in Weifang Prefecture are largely confirmed by studies conducted in other Chinese provinces on these issues.13

• 14 Although it was voluntary, joining the health insurance scheme was strongly encouraged. The local h (...)

29Although these reforms have been a success in several respects, they have been less so in others. The first success, and a remarkable one, was the rapid and decisive move by households to join the NRCMS, even though joining was voluntary:14 from 5% coverage of households before the reforms, the rate reached 95% in 2008. The second success was the large increase in TH activity, which as we have seen was an expected outcome; it raised the rate of patient contact through external consultations from 0.6 before the reforms to 0.83 afterwards.

• 15 Jacky Mathonnat et al., “Progrès et défis pour trois réformes majeures dans le système de santé chi (...)

32In 2009, the government initiated a new wave of reforms aimed at universal health coverage by 2020. Care quality continued to preoccupy the authorities. The care provided by the township hospitals was judged to be poor by households who avoided them in favour of second‑level hospitals, leading to their being overloaded. Moreover, despite the controls over the price of medicines set up between 1992 and 2007, the financial withdrawal of the state from the health sector had gradually pushed health facilities towards the profit motive, leading to widespread over-prescription and irrational use of medicine. The strategy adopted by the hospitals for bypassing the controls was to limit the use of essential fixed‑price medicines (listed in 1992) in favour of medicines without price restrictions. This made these medicines more expensive for the patients and more profitable for the hospitals, which were both prescribers and suppliers.15

• 16 National Health and Family Planning Commission, Guojia jiben yaowu mulu (National list of essential (...)
• 17 Users pay the costs of care in advance and claim repayment later.

34This reform has four aspects: the production of medicines, the fixing of prices, the modalities for providing and paying for the care structures, and the rationalising of prescriptions. A new list of essential medicines has been drawn up. It came into force in 2012 and includes 317 medicines produced by the pharmaceutical industry, 203 by traditional Chinese medicine, and a certain number of plants with “recognised” medicinal properties.16 In order to take account of the specific needs of the different provinces, local governments may add a supplementary list. For example, in Shandong Province a further 216 medicines were added. Essential medicines are sold at cost price,17 because the primary healthcare facilities no longer have the right to sell them for a profit ‘(“zero mark‑up” policy). Moreover, these medicines are the only ones that such facilities are allowed to prescribe. When care delivery specifically requires the use of non-essential medicines, the patient has to be referred to the district hospital or to buy drugs on the market. In this way the de facto link between hospital income and staff income, which had emerged because of hospitals’ ability to increase their revenue through the sale of medicines without price controls, was broken once staff were no longer motivated to over‑prescribe. To make up for the loss of resources linked to the sale of medicines, different mechanisms for compensation have been put in place, each district being able to choose the modality that suits it best. They are: compensation paid exclusively by government subsidies, an incentive system such as payments related to measurable results, the development of supporting activities (laboratory tests, and X‑ray examinations) from which the hospital gets a share, or a mixture of several of these methods. Finally, in accordance with the desire to emphasise public health, the functions of the township hospitals were strengthened by the development of preventive activities (targets for vaccination numbers, followed by targets for children and the elderly to be given health checks and, where necessary, home visits) and more demanding supervision of health stations.

• 18 Martine Audibert, Xiezhe Huangfu, Jacky Mathonnat, and Laurène Petitfour, The Chinese Drug Policy R (...)
• 19 One of the reasons brought up during our interviews is that the salaries offered by the government (...)

35Descriptive analysis of the data we gathered18 highlights shifts in the township hospitals’ resources, both human and physical, over the period studied (2006-2012). The township hospitals’ staff establishment falls into two categories: “in quota” and “extra quota.” The number of staff in the quota is calculated and approved jointly by the Ministry of Human Resources and the Social Security Office at the provincial level, and salaries are paid out of the government budget. Staff outside of the quota are recruited by the hospitals, either because the quota could not be filled19 or because the hospitals want to recruit additional staff, and these are paid out of hospital funds. Our analyses show that overall staff numbers did increase, but more before the reforms than during them, with gaps between the different hospitals widening under the reforms (Graph 4). The analysis of this evolution in terms of staff status shows that, with some exceptions, the growing numbers of in‑quota staff followed overall growth in staff numbers. The same applies to the number of beds, which rose sharply (from 44 in 2006 to 77 in 2012) (Graph 5). It should be noted that differences between hospitals increased as much in staff numbers as in the provision of beds.

• 20 Karen Eggleston, Li Ling, Mengue Qingyue, Magnus Lindelow, and Adam Wagstaff, “Health Service Deliv (...)

38These results, which throw light on the strategy of township hospitals, are also emphasised in other studies.20 Beyond medical reasons, there are financial considerations that encourage investment in hi-tech equipment whose costs are not framed.

• 21 Martine Audibert et al., The Chinese Drug Policy Reform: Analysis of the Activity and the Efficienc (...)

39One of the aims in establishing a list of essential medicines and in forbidding the sale of medicines for profit was to reduce the cost of care. That aim seems to have been achieved. In the hospitals studied, the income from medicines per outpatient and inpatient fell by around 50% between 2009 and 2012. This shift caused by the reform is quite clear (Graph 7). All things being equal, the result is a potential reduction in expenditure on medicine, that is to say savings (equal to a positive gain) for the patients and for the health system. We have estimated the savings at 60% of what the expenditure would have been if there had been no reform. In parallel with this, the amount of reimbursement for hospitalisation has risen greatly. In 2012 it was two to four times higher, depending on the district, than in 2008.The situation is a bit different for external consultations because there is huge range of variation according to district: in some districts the amount of reimbursement made great progress over the period, but in others it stagnated immediately after 2009.

• 22 Owing to a lack of data, preventive activity, apart from vaccinations, could not be analysed. Our d (...)

40As was the case in the previous period, the curative activity22 of the hospitals expanded greatly. The average annual number of external consultations per hospital doubled from 30,000 in 2006 to 60,000 in 2012. The number of hospital admissions likewise increased steadily in the same period, rising from an average of 1,500 per hospital in 2006 to nearly 3,400 in 2012. However, our analysis of changes in the share of admissions shows an increasing inequality between hospitals (Graph 8). One part of this growth in demand arose from the transfer of patients who, instead of being handled by the more expensive district hospitals as in the past, were now channelled more towards township hospitals. This was because of the lower cost of medicines, the better management of insurance, and the perceived (and objective) improvement in the quality of care. Another part of this increase in demand arose because some patients, who for reasons of affordability would have given up seeking health care before the reform, now became “new entrants.” Finally, our interviews have revealed that village health units, having been placed directly under the township hospital authorities, are referring more patients to them.

• 23 Laurène Petitfour, Xiezhe Huangfu, Martine Audibert, and Jacky Mathonnat, “Estimating and Explainin (...)

42We have seen that reform of rural health insurance did not have any clear influence on the efficiency of the hospitals. Did drug reform enable the hospitals to perform any better?

• 24 The majority of hospitals are in deficit every year.

43The results show that the efficiency of THs remains low. On average, they could actually improve their efficiency (that means delivering more services) by at least 40% without using more resources. Furthermore, the reform seems rather to have had a negative effect on their technical efficiency, the average of which declined over the period studied (2009-2012). In addition, a greater number of hospitals had an efficiency score equal to 60 or less (Graph 9, which likewise shows a lower average score). This is partly explained by the fact that in the greater part of the hospitals in our sample, the increase in subsidies more than made up for the loss of resources from the sale of medicine, creating what is generally called a situation of “soft budget constraint” mentioned previously. In fact, the arrangement permitting the awarding of subsidies, when compared with the previous situation, allows township hospitals to reduce their deficit,24 all things being equal, without having to make any particular effort to do so. It would seem, as we have emphasised already, that efforts to invest were somewhat disproportionate to needs, just as the recruitment of non-quota staff was left to the initiative of the hospitals. We also noted that some hospitals intensified the delivery of profitable activities (injections, laboratory tests, sonar scans, etc.). It is therefore quite likely that the cost of care borne by patients has fallen less than it might have done.

• 25 Adam Wagstaff and S. Yu, “Do Health Sector Reforms have their Intended Impacts? The World Bank’s He (...)

49The third positive result is that there has been improvement in the way the NRCMS handle patients’ interests: insurance reimbursements for sickness have greatly increased (rates and amounts), together with the moderating effect of the pharmaceutical policy on the cost of medicines. These latter two results have helped considerably in lowering financial barriers to access to healthcare. But the drop in the number of households facing major health expenditures (“catastrophic costs”) is probably smaller than expected. A study by Wagstaff and Yu shows that in Gansu Province, households now covered by health insurance for a list of specified treatments are seeking treatments whether they are covered or not, which they would have avoided in the absence of an insurance scheme.25 One consequence has been an increase in private health expenditure, creating in some cases a counter-intuitive but not paradoxical link between the development of the insurance scheme and a large increase in costs borne privately by households. But there is no evidence of a nation-wide magnitude of this kind of situation.

• 26 According to our interviewees at the National Health and Family Planning Commission.

51The first is the quality of health care. Admittedly, our studies do not allow us to prejudge the effect of the reforms on the quality of prescriptions or of the care delivered, because our lack of data meant that we could not take them into account. That said, however, our interlocutors tell us that care quality has actually improved with the reforms, albeit unequally. In general, major underlying problems still persist, creating major public discontent. In extreme cases, this has led to acts of violence towards medical staff in different parts of the country.26 It is difficult to know the extent of this problem, but it is taken very seriously by the health authorities and by the Party. Moreover, patients have become increasingly demanding and are attracted by hi-tech treatments, and this has led certain establishments to develop strategies of investment in top‑of‑the‑range technology, enabling them to increase their income even when it is not always clear that there is any gain in care quality for the patient.

• 27 We believe that there is a need to reconsider the health map by matching some township hospitals an (...)

52The second problem is that the efficiency of the hospitals has remained low, and even deteriorated over the final years of the study, partly because a number of them suffered from overcapacity,27 and partly because the reforms contributed to a soft budget constraint for the THs. In addition, the system still works in a very bureaucratic manner. The managers are not really in a position to manage their establishments, and they are not sufficiently incentivised to do so. So there are questions to be asked about rearranging the incentives in order to improve efficiency, which is a very real preoccupation of the directors of the National Health and Family Planning Commission.