The cost of coverage: rural health insurance in China

HEALTH POLICY AND PLANNING; 11(3): 238-252 © Oxford University Press 1996

The cost of coverage: rural health insurancein ChinaLENNART BOGG,1 DONG HENGJIN,2 WANG KELI,3 CAI WENWEI,3 AND VINOD DIWAN1

'Division of International Health Care Research (IHCAR), Karolinska Institute^ Stockholm, Sweden,'Department of Health Statistics and Community Medicine, and 3Department of Maternal and ChildHealth, School of Public Health, Shanghai Medical University, People's Republic of China

China has undergone great economic and social change since 1978 with far reaching implications forthe health care system and ultimately for the health status of the population. The Chinese MedicalReform of the 1980s made cost recovery a primary objective. The urban population is mostly pro-tected by generous government health insurance. A high share government budget is allocated tourban health care. Rural cooperative health insurance reached a peak in the mid-1970s when 90%of the rural population were covered. In the 1980s rural cooperative health insurance collapsed andpresent coverage is less than 8%. The decline has been accompanied by reports of growing equityproblems in the financing of and access to health care. This article is the first in a four-year studyof the impact on equity of the changes in Chinese health care financing. The article examines the rela-tionship between rural cooperative hearth insurance as the explanatory variable and health careexpenditure, curative vs. preventive expenditure and tertiary curative care expenditure as dependentvariables using a natural experimental design with a 'twin' county as a control.

The findings support the hypothesis that cooperative health insurance will induce higher growth ofhearth care expenditure. The findings also support the hypothesis that cooperative hearth insurancewill lead to a shift from preventive medicine to curative medicine and to a higher level of tertiary curativecare expenditure. The empirical evidence from the Chinese counties is contradicting World Bank healthfinancing policies.

BackgroundSince 1987 the World Bank has promoted a healthfinance policy package consisting of user fees, healthinsurance, privatization and decentralization. Chinaimplemented a Medical Reform in the 1980s muchin line with the World Bank policy recommendations.

China has undergone great change during the recent15 or so years - the Deng Xiaoping era. Change isevident in the GNP growth, income distribution,health care utilization and in the health status of thepopulation. GNP growth has outperformed all othercountries in the world, present and past. It is generallyaccepted that the economic level of a society,alongside female education and well-designed publichealth interventions, is one of the strongest deter-minants of the health status of the population. Thus,it could have been expected that the health indi-cators of China would improve sharply over the sameperiod.

The health indicators have improved in the big citiesof China, but in the countryside, housing the vastmajority of the population, signs of improvement arefew. Instead, concern1 has been expressed that someof the previous gains in health status may be lost. Thedifferences between regions, and between urban andrural populations, are wide and widening. In a coun-try of China's dimensions average data are not suffi-cient to understand the multi-faceted reality. Yet, itis alarming to note that the average infant mortalityrate (IMR) was officially recorded at 34.7 per 1000in 19812 and at 37.0 in 1992.3 Even more alarmingis the growing difference in IMR between the poorerrural areas and the major cities in China, which wasof a factor of 4.6 or mirroring the difference in mort-ality levels between Lesotho and Hungary (Table 1).

According to UNICEF the true IMR for China isaround 52 per 1000 (1993).4 A survey of the healthservices and the health status of the population in

Rural health insurance in China 239

Table 1. Infant Mortality Rate - Trends of inequity in China(mortality within first year of life per thousand born)

Table 2. Per capita consumption - Trends of inequity in China:RMB Yuan (percentage of whole China average)

1981" 1992* 1983 1988 1992 1993

Whole ChinaMajor citiesPoor rural counties

34.7 37.015.871.8

•: 1981 data arc based on the Third National Census*•: 1992 data are based on the Fourth National Census

300 poor counties (12% of all China) was undertakenby the Ministry of Health (MoH) in 1989, withtechnical assistance and financial support by UNICEFand UNFPA. The survey documents a dismal stateof maternal and child health in the poor areas.3

Maternal mortality exceeds 300/100 000 in somepoor and remote areas, a level comparable to that ofRwanda. The IMR exceeds 100/1000 in 38 of the 300counties and the average IMR was 68/1000, com-parable to Kenya. At the same time the health in-dicators of the big cities, like Shanghai and Beijing,are on a level with or better than those of big citiesin the US.

Although there could be questions regarding thereliability of earlier health statistics, there is littledoubt that China of the 1960s and 1970s had achieveda high level of public health, thanks to successfulpreventive programmes and an emphasis on primaryhealth care and universal accessibility. The absenceof health improvements for the rural population inthe last 15 years could partly be explained by im-proved health statistics, but the disappointing healthdevelopment is likely to be more related to the com-bined effects of widening income gaps and the large-scale introduction of market elements in the Chinesehealth care system.

The income and consumption inequities have widenedin the last few years (Table 2). The average ruralChinese had a consumption level which was 80% ofthe national average in 1983, but in 1993 he/she en-joyed only 67% of the national average. In the sameperiod the average urban Chinese increased his/herconsumption from 189% of the national average to216%. The average per capita consumption in 1992of urban Shanghai citizens was more than 7 times thatof rural residents in Guizhou province.

The per capita expenditure on medicine and medicalservices in 1993 in urban Guangdong (RMB 103.70)

Whole China 289 (100) 635 (100) 947 (100) 1148 (100)

Urban areasRural areasShanghai:

Urban

547232

Guizhou Province:Rural

089)(SO)

1281473

(202)(74)

2032650

3050

421

Q15)(69)

(322)

(44)

2480774

(216)(67)

Source: China Statistical Yearbook 1994, current prices

was more than 33 times that of rural Tibet (RMB3.10).6 A report,7 however, from an 8 provincesurvey in China on equity in the utilization of healthservices concludes that China has achieved a verywide distribution of clinics and other services and thatthey are widely used by those who identify needs forthem. These findings contrast with a number of otherreports both in China and outside, which have pointedto a declining accessibility of services in recentyears.*"10

China has an impressive network of health servicesextended over the country. Up to the end of the 1970smost health services were provided free or at a smallcost and until recently the income differences weresmall. Starting from around 1986 Chinese health careunderwent a Medical Reform (Yiliao Gaige), theprime feature of which was increased emphasis oncost recovery. The health care system in China hasnow become extremely dependent on fee-for-service(FFS) revenue, mainly from drug sales. Bonus pay-ments to physicians provide an incentive to raiseservice revenue as much as possible. Targetedmarketing methods, previously unheard of in China,even bribes to treating physicians," are now quitecommonplace.

These facts, however, conceal considerable localvariations in a system that has become increasinglydecentralized. The direction of the recent changes isall the more surprising, given that they have theirpoint of departure in a health care system which wasdesigned with equity as a top priority and whereequity was achieved relatively successfully.12 Co-operative health insurance is seen by the Ministry ofHealth in China as an important part of the solutionto the growing inequities in access to basic healthcare.1314

240 Lennart Bogg et al.

The focus of this paper is on the allocative efficencyof cooperative health insurance, with implications forequity in health care utilization and health outcome.This represents the first phase of a SAREC-fundedstudy of the impact on equity of the changes in healthcare financing in China since 1986.

Rural health insurance - the Chinese experienceCooperative health care insurance has a relativelylong history in China. In the 1940s cooperative ruralhealth care was organized in the Shan-Gan-Ningborder area in China.15 In 1955 cooperative ruralhealth insurance was set up in Henan province. TheCommunist Party Central Committee expressed itsapproval of cooperative health insurance on 2February 1960. In December 1968 Chairman Maocommended the cooperative health insurance systemof Leyuan people's commune in Changyang countyin Hubei province. This created the impetus to startcooperative health insurance systems all over thecountry.

In 1976 about 90% of all the production brigades(villages) in China were covered by cooperativehealth insurance. By 1978 cooperative health in-surance was stipulated by law and on 15 December1979 the Ministry of Health, the Ministry of Financeand the Ministry of Agriculture jointly issued regula-tions governing the cooperative insurance systems.

The cooperative movement grew quickly during the1970s, but faded even more precipitously in thefollowing years. The development of rural coopera-tive health insurance coincided with the culturalrevolution (1966-1976). This association has left anegative imprint on the population in many areas,who continue to view cooperative insurance as a'product of the cultural revolution'. Moreover, poormanagement and frequent misuse by cadres duringthe early implementation added to the negative image.At present less than 8% of the population are coveredby cooperative or collective health insurance.16

In the past few years reports of decreasing access-ibility of health services have contributed to revivedinterest and active promotion by the Ministry ofHealth of rural cooperative health insurance schemes.

Rural health insurance in other developingcountriesVoluntary health insurance schemes for the ruralpopulations are now encouraged by the governmentsof China, Thailand, Vietnam and other countries. It

has to be noted, however, that the success stories arefew.

In Thailand17 the voluntary health card scheme hasbeen promoted since 1983, but has only reached acoverage of 5% of die population. There has beenreluctance from die population to purchase the cardunless a major expenditure is imminent, such as adelivery. As a consequence the scheme is heavilydependent on government subsidies. Another problemis that card holders sometimes find that they get lessattention from care providers than if they were pay-ing in cash.

Vietnam launched a voluntary health card scheme in1993." The coverage remains low, less than 5% ofthe population, of which more dian 90% are com-pulsory participants. There are many problemsreported; high dependence on government subsidies,poor quality of services, doctors asking for bribes anddiscrimination against card holders.

The low participation in Asian countries could be ex-plained by many factors in addition to die problemsreported, e.g. culturally determined risk evaluation.

Burundi has implemented a health card scheme since1984." In a survey it was found that 23% of house-holds held valid healdi cards. Women who wereinterviewed thought that the health card was good forpoor families, for families with seasonal income andfor women with husbands who drink. However, itwas also noted that the quality of services was low,that drugs were frequently out of stock and diat cardholders received less attention. A 50% higher utiliza-tion among card holders was observed. Twenty-sevenper cent of the households reported financial inabili-ty as the reason for non-participation. The revenuesof the card sales cover only one-third of the drugcosts.

The Bamako InitiativeThe Bamako Initiative refers to a variety of modelsfor community financing of essential drugs forprimary health care, an initiative endorsed in 1987by African leaders. With financial backing and pro-motion by UNICEF, it has been implemented in dif-ferent forms and to different extents by a number ofcountries in Africa and Asia. Although China is notofficially participating there are many similaritiesbetween the Bamako Initiative and die Chinese sys-tem of community financing by means of user fees,mainly for drugs.


A study of the early implementation of the BamakoInitiative20 in four countries concluded that there isconsiderable variation between the countries both inthe way the Initiative has been implemented and inthe country experiences. In most of the countriesproblems exist with respect to the incentives createdby the pricing structure. The failure of the exemp-tion mechanisms to protect vulnerable groups wasevident in all countries. The affordability of serviceswas cited as a serious problem in relation to traveland other costs involved in the use of higher levelfacilities, and to use of NGO-facilities. The evalua-tion suggested that a clearer picture of the relativeaffordability of the Bamako Initiative activities isneeded. The concerns in respect of the incentives,the exemption mechanisms and the affordability ofthe Bamako scheme could be equally relevant for theChinese rural health care system.

User fees and 'quality'Litvack and Bodart21 reported in 1993 a pre-postcontrolled experiment in Cameroon of introducinguser fees simultaneously with improved drug supply.Their conclusion was that utilization increased sig-nificantly, especially among the poorest quintile, asa result of the intervention. The report has beenfrequently quoted in World Bank publications asempirical evidence supporting the World Bank policypackage.22 The fact that people will be prepared topay according to their ability when drugs are availableand that they will not bother to visit clinics withoutdrugs, even if no charge is made, is not really sur-prising. This notwithstanding, the study design hasbeen repeated by others.23

These studies are based on only partial cost recoveryand tell us little of the long-term viability of suchschemes and even less of the impact in terms ofthe commercialization of services. To increase theutilization of health facilities by providing the drugsthat should have been there in the first place is notsufficient to convince that user fee financing is thepath to a better health for all.

Health insurance and health care expenditure -a theoretical frameworkThe World Bank has focused on health care financ-ing in developing countries as one of its central areasof interest since it started to take an active role in thehealth sector in the 1980s. In 1987 the World Bankpublished its policy paper 'Financing Health Servicesin Developing Countries: An Agenda for Reform'.2*The Agenda for Reform proposes four policies:

• to charge user fees at government health facilities,• to promote health insurance,• to promote the private sector, and• to decentralize government health services.

This World Bank policy document reflected themarket-oriented trends of the 1980s and has had con-siderable influence. Today the majority of govern-ments in developing countries are involved in healthsector reforms implementing part or all of the pro-posed policies. China, whether influenced by theWorld Bank or not, has followed the recommenda-tions closely.23'26

The four policies are interrelated in the sense thatprivatization will require user fees, which willfacilitate decentralization. The increased relianceon user fees is likely to create financial difficultiesfor the poor and for those suffering from catastro-phic or chronic diseases, which will prompt healthinsurance.

The Agenda identified three main problem areas inthe health sector, which the policies were targetedto deal with:

• allocation, insufficient spending on cost-effectiveprogrammes;

• internal efficiency, poor use of scarce resources;and

• inequity, inequalities in access to care.

According to the Agenda, health insurance wouldhelp to improve allocative efficiency and to reduceinequities, in access to care.

The World Bank's Agenda for Reform linked thesuccessful introduction of user fees in developingcountries to widespread popular participation inhealth insurance schemes. Four main systems weresuggested:

• Social insurance, which is related to a governmentsponsored health insurance, often compulsory forgovernment staff.

• Employer-based schemes, where the employereither maintains his own on-site health facility orcontracts with an external facility to provide healthcare for his employees.

• Prepayment schemes, including, for example,health maintenance organizations (HMOs),systems where the participants have formed orjoined a group to regularly pay a fixed amount and


in turn receive care according to the rules of theorganization.

• Private insurance, which involves voluntary pay-ment of premiums to an insurance company to getthe agreed coverage with respect to health carefees.

All four systems can be found in Chinese rural healthcare financing.

Health insurance can involve both a savings aspect,i.e. risk sharing across the time span of an individ-ual's life, and an equity aspect, i.e. risk sharing acrossthe population. If the risk sharing is limited to thelocal community, which is often the case with ruralhealth insurance, the equity implications of the in-surance scheme will obviously be more limited. Theobjective can be to cover smaller expected expen-diture as a kind of prepayment or savings systemand/or to cover major unexpected expenditure of amore catastrophic nature.

What constitutes a particular challenge in the designof health insurance systems, and especially in ruralhealth insurance, is the fact that health care needs aregreater among those with the least ability to pay,while for most other insurance systems the demandfor insurance increases with the ability of the insuredto pay.

One of the purposes of introducing health insuranceis, obviously, to improve access to services forunderserved groups. However, improving access mayincrease costs more than expected. For health in-surance to be viable, long-term expenditure must notexceed long-term revenue. Health insurance is some-times introduced with expenditure projections basedon past expenditure patterns. This is likely to bemisleading, since large previously unmet needs canbe anticipated, particularly among low-incomegroups.

The concept of 'moral hazard' refers to the risk thatthose individuals who are aware of their poor healdiand have a high likelihood of medical expenditure willbe more prone to enrol in voluntary health insurance,and that those who have enrolled will be less hesi-tant to use the available services since they only haveto pay part of the cost. There is a risk that providersmay induce extra demand that will consume the ad-ditional resources mobilized by health insurance. Itis well established that in the health care sector theproviders, e.g. the physicians, have a considerable

influence on the demand for health care by die-patients, something which is usually referred to as'supplier-induced demand' (SID).

Public subsidies of health insurance are motivated byme public interest to ensure the entire population'saccess to basic health services, arising both from thewish to prevent or reduce disease spreading and fromaltruistic motives to avoid human suffering as far aspossible. These characteristics have led most of theEuropean countries to establish tax-financed or com-pulsory social insurance systems of financing healthcare.27

In China, as in most of the developing countries, tax-financing has not been seen as a feasible short-termsolution to healm care financing. An efficient taxsystem requires legal, accounting and auditingsystems, and, even more critical, a body of trainedprofessionals to operate the systems.

Finance and ethicsTwo broad but rival ediical bases on which systemsof health care finance can be based appear from theliterature. The first considers access to health careas essentially similar in ethical respects to access toodier goods. The second ethical basis regards accessto health care as a right of citizenship, like me rightto vote, which should not in any way depend on in-dividual income or wealth.

This second ethical basis is clearly the predominantviewpoint in most European countries and in Canada.This was the ediical basis of Chinese health care upto 1978, even if at all times there have been dif-ficulties in living up to the high ideals. Article 45 ofthe Chinese Constitution stipulates: 'Citizens of thePeople's Republic of China have me right to materialassistance from the State and society when they areold, ill or disabled. The State develops die socialinsurance, social relief and medical and healm ser-vices that are required to enable citizens to enjoy thisright.'

The Medical Reform (Yiliao Gaige) in ChinaDuring die 1980s me Chinese government had to facea shrinking revenue situation. Both central and localgovernment revenue decreased from more dian 34%of the GNP in 1978 to less than 20% in 1992.28-29 In1987-88 the provinces had to cut their expenditureon healm care by 18%.30 The central governmentsaw its power slip along with the shrinking tax base.The MoH's already small support for village doctors


declined by 45% in real terms from 1979 to 1987."At the same time the cost to the government of the'free' government health insurance (Gongfei) systemincreased by an average of 30% annually.32

In May 1982 the central government allowed someenterprises to try new methods to reduce medicalcosts. From around 1985 the previous system offinancing rural health care by a cooperative medicalservice (health insurance) was replaced by user feefinancing.33 In March 1988 a reform research groupwas set up to review the (government) medical in-surance system. The group produced a draft titled 'APlan to Reform the Medical Insurance System'. Fourcities were chosen in March 1989 as trial areas.

According to a statement by Dr Chen Minzhang,Minister of Health: 'the way to reform the freemedical care system is to change it into a comprehen-sive social medical insurance system, strengtheninggradually people's 'sense of cost' and getting moreefficient use of the state's medical funds.' The im-plication of this statement seems to be that the 'free'government health insurance system (gongfei) willeventually cease to exist and will be replaced by asocial health insurance system, with considerableco-payments and/or co-insurance. A key objectiveappears to be a better control of the rapid costescalation.

Health care expenditure and provision of servicesin rural ChinaThe rural health care system in China is organizedin a three-tier system in the counties; county hospital,township hospital/health centre and village healthstation. A survey34 in 1987-1988 of the financingand provision of MCH services in six counties inChina by the Shanghai Medical University and theMCH Project Office of MoH demonstrates that thelower the level of care, the heavier the reliance onfee-for-service and especially drug sales for revenue(Table 3).

In five of the six surveyed counties an average of71.5% of the total health care revenue was fee-for-service revenue. The county government provided onaverage 24.8% and province and prefecture (betweencounty and province-level) governments contributedan average of 3.7%. Sales of drugs made up no lessthan 53.1 % of the total health care revenue. Surgeryaccounted for less than 2% and other clinical fees forless than 7% of total revenues. This reflects theperverse pricing of health services in China, where

Table 3. Sources of health care financing in five counties* inChina, 1988 (percentage of total financing)

County Township MCH Paediatric(Total) (PHC)

Public budgetFee-for-serviceTotalof which drug sales

28.571.510053.1

15.884.2100

60.8

7.093.010057.5

5.095.010056.3

* Anfu County in Jiangxi Province, Fuyu County in Jilin Province,Hu County in Shaanxi Province, Mingshan County in SichuanProvince and Tonghai County in Yunnan Province

Source: MCH Services and Expenditure Research 1990, ShanghaiMedical University/MCH Project Office of MoH, Beijing

clinical services are priced below costs, which againare kept low by central salary controls.

In maternal preventive work fee-for-service revenueaccounted for an even higher share, 93.0% of totalrevenue, with drug sales providing 57.5%. Pre-vention in this context is defined on the basis ofadministrative structure and includes all services per-formed through the maternal and child health and theanti-epidemic (immunization) units. In preventivehealth care for children fee-for-service revenue pro-vided 95.0% of total revenue, of which 56.3% wasfrom drug sales.

The bulk of preventive work in Chinese rural healthcare is performed at township or village level. Attownship level (including villages) the county budgetprovided 14.5% and the township budget another1.3% of the total revenue. Fee-for-service chargesgenerated the remaining 84.2% of the revenue. Drugsales contributed 60.8% of the total revenue at thetownship level.

Most of the public budget funding (60.2%) isallocated to the county level (tertiary care) while mostcurative work takes place in the villages where almostno public budget funds (0.4%) are allocated. Mostpreventive work is based at the township healthcentres (64.8%), which receive only 39.4% of thepublic budget funds (Table 4).

Purpose and hypothesisIn 1993 we initiated a research project focusing onthe impact on equity in the financing of and accessto health care resulting from the medical reform in


Table 4. Allocation of public budget expenditure, curative andpreventive services in Rudong County, Jiangsu, China, 1988(percentages)

Level of care Public budgetexpenditure

Curativeservices

Preventiveservices

TertiarySecondaryPrimaryTotal

60.239.40.4100

9.136.454.5100

0.864.834.4100

Source: MCH Services and Expenditure Research 1990 ShanghaiMedical University/MCH Project Office of MoH, Beijing

China. For the present paper we have used a con-trolled natural experimental design to analyze thehealth expenditure trends in two 'twin' counties inJiangsu province of China. One of the counties hasa health financing system built on health insuranceand the other county has a predominantly user feefunded system. The analysis is based on macro-levelhealth care expenditure data. We believe this to beone of the first reports using this study design.

The Rand health insurance experiment33 was an im-portant (and very costly) randomized experimentwhich demonstrated in an American setting that co-insurance will reduce total health expenditure. TheRand study also found that 'well-care' services areabout as price responsive as other services. Similarexperiments have not been reported from other coun-tries and are perhaps unlikely for cost reasons.

The Agenda for Reform postulates that health in-surance will improve allocative efficiency and helpto reduce inequities in access to care. Our assump-tion is that different financing systems yield differenthealth care expenditure profiles and will differ in theirimpact on equity in terms of who bears the cost ofand who enjoys access to care. The interrelationshipbetween the financing system and the provision ofhealth services is determined by a complex interac-tion of the political sphere, third party-payers, pro-viders and patients. Their interaction determines theavailability of funds for health services, the alloca-tion of funds for different services, and the equityand efficiency of the system. Exactly how this interac-tion develops is determined by the general socio-economic environment, as is suggested by our model(Figure 1).

One of the fundamental problems is to understandexactly which parameters in the financing system are

health fetatu

Source. Lennart Bogg, 1HCAR

The present studyconcerns the relationshipbetween the payers ofhealth care, the patientsand the providers asreflected in the financialflows, i e the funding ofhealth care

Figure 1. Model of health financing and health service deliveryin a social context

the explanatory variables; the relative mix of userfees, budgets and insurance, the payment modes andlevels or the relative price of different services. Fewempirical studies have attempted to shed light onthis question. Most of the studies so far have beenexploring only one explanatory variable, for exam-ple the GDP level or the introduction of user fees orhealth insurance. We are examining cooperativehealth insurance versus user fees as explanatoryvariables, with inflation-adjusted per capita healthcare expenditure, curative expenditure, preventiveexpenditure and tertiary curative expenditure as thedependent variables. We have recently collectedmicro-level data in household surveys, provider in-terviews and patient exit-interviews which will makemultivariate analysis possible in the near future.

The provider and patient incentives of the reformedChinese health care system, with its fee-for-servicesystem, could be expected to result in a high propor-tion of expenditure on curative medicine relative topreventive medicine and to a shift towards higherlevel curative care. This is due to the fact that theprice level of services is kept below cost except


for Pharmaceuticals and high-tech diagnostics36-37

and that providers have financial incentives (bonusincome) to increase service revenue, at the same timeas insurance-covered patients may demand or beinduced to demand excessive services or unnecessarydrugs.

Our hypothesis is that a system based on voluntaryhealth insurance will contribute to a higher level ofhealth care expenditure, to an increasing share ofcurative care expenditure and to an increasing shareof tertiary care expenditure.

Setting and methodsWe have reviewed the health care revenue andexpenditure in Jintan county and in Jurong county.Jintan county has developed a comprehensive healthinsurance system covering the majority (more than60%) of its inhabitants, while Jurong has decided torefrain from developing; health insurance and hasmainly user-fee based financing of health care. Bothcounties are relativley well-off by Chinese standards.Jintan county has a slightly smaller population of541 000 (9.1% less) with an average GNP per capitaof RMB 4210 (1993). Jurong has a somewhat largerpopulation of 598 000 and a slightly lower GNPlevel, RMB 3833 per capita (minus 9.0%). Both are

above the national average GNP per capita in Chinaand around the average of Jiangsu province.

Jintan and Jurong counties were purposely selectedas two neighbouring 'twin' counties with similarsocioeconomic environment, demographic and healthstatus, geographic, cultural and educational condi-tions as well as political and regulatory environment,but with different health financing systems (Table 5).

We have reviewed health sector accounts data andcompared the funding sources, expenditure levels andtrends of the two counties. The time period reviewedwas 1986 to 1994. This period was chosen as 1986was approximately the time that emphasis was shiftedfrom cooperative medical services to cost recovery,by means of user fees, in Chinese rural health care.From that time only part of the health care expendi-ture, indeed only part of the staff salaries, has beenpaid from public budget contributions.

The GNP and the health expenditure data have beenadjusted by the rural general price index, using 1986as the base year. We did not use the rural medicalprice index for reasons of comparability with GNP.We were also not so much concerned with determin-ing the volume changes but rather the change in costto the patients, whether caused by volume or price

Table 5. Health systems environment of two 'twin' counties in China (1994)

Jintan county Jurong county

Financing system

Health insurance coverageIncome per capitaService deliverySocioeconomic characteristics

Population

Health status

Traditions/culture

Health care regulatory system

Geography

Educational system

Political system

Fce-for-service payments to providerswith health insurance refunds of part ofthe cost for the majority>65% of the populationRMBY2522public ownershipagriculture-based village and townshipenterprises, 25.3% farmers541 000 pop. (1992), 11.8% urban,one-child policy, 5.8% < 5 years old,14.2% >60 years oldgrowing problems with chronic anddegenerative diseases, gaining controlover epidemic diseasesHan Chinese populationcentral government price controls,central salary systemsouthwest Jiangsu, mainly flat landwith some hilly areascentralized educational system,>99% school enrolmentCCP-control

Fee-for-service payments toproviders without insurancerefunds for the majority< 5 % of the populationRMBY 1847public ownershipagriculture-based village and townshipenterprises, 47.2% farmers598 000 pop. (1992), 15.2% urban,one-child policy, 5.0% <5 years old,10.4% > 60 years oldgrowing problems with chronic anddegenerative diseases, gaining controlover epidemic diseasesHan Chinese populationcentral government price controls,central salary systemsouthwest Jiangsu, mainly flat land withsome hilly areascentralized educational system,>99% school enrolmentCCP-control


level changes. For 1994 we used the general priceindex, since the rural general price index was notavailable.

The data sources for this paper are interviews andaccounts data extracted from the public healthsystem in the two counties. It is our impression thatthe quality and detail of public health accounts havegreatly improved in recent years. During a visit inOctober-November 1993 to both Jurong and Jintancounties we interviewed health officials at county,township and village levels. Health facilities and prac-titioners outside the public system were not included.The private sector, NGOs and other non-MoH oper-ators remain very small in the Chinese health sector,with the exception of hospitals operated by the armyof other units, e.g. the Railway Ministry. The villagedoctors in Jurong county can be viewed as semi-private; they are, nevertheless, included in this study.

Findings and discussionJintan county - rural health insurance systemJintan county has developed, in collaboration withmedical universities and encouraged by the MoH, arural health insurance system, which by some is seenas a possible model for national level implementa-tion. The health insurance system as of the late 1980swas described in an ILO report.38

Several health insurance systems combine to builda relatively comprehensive system. First, there arethe national insurance systems: the GovernmentHealth Insurance (Gongfei) and the State EnterpriseWorker Insurance (Laobao). Gongfei and Laobaomainly cover the urban state-employed population.In Jintan county about 3.7% of the population arecovered by either Gongfei or Laobao. Gongfei canbe characterized as a social insurance with coverageof government staff and limited coverage for theirimmediate dependants. Laobao can be classified asan employer-based scheme. Previously, the entitle-ments were similar under the two systems, which bothguaranteed free examination, treatment, hospitaliza-tion and drugs, except for a minimal registration fee,board expenses during hospitalization and nutritiousmedications. Direct dependants receive 50% cover-age of expenditure.

The Jintan cooperative health insurance systemcovers more than 60% of the county population. In1969 the coverage was more than 75%. Later it

decreased until in 1983 the system again started toattract participants. In 1992 there was yet anotherdecrease in participation. The growth of private enter-prise has affected participation negatively. It wasreported that healthy people generally were less in-clined to participate in the system. The Jintan co-operative health insurance is built on an individualcontribution of 1.0-1.5% of the per capita income,or RMB 8-10 per year. The villages contribute fromtheir welfare funds and the county contributes fromthe previous year's excess service revenue. There isa list of 120 essential drugs which are covered by thesystem. Fifty per cent of the funds are expended atvillage health stations (primary care). The participantspay the registration fee (RMB 0.20) only, treatmentand drugs are covered. Drug demand is controlledby issuing only one day's need at a time, except forsome chronic diseases. A list defines the ailmentscovered by the scheme. The system will reimburseoutpatient visits by RMB 3-5, inpatient visits attownship hospitals up to RMB 50 and county hospitalinpatient visits up to RMB 100. Inpatient visits willrequire referrals. Outpatient visits to higher levelfacilities are not reimbursed. The ceilings for reim-bursement appear modest in view of the cost of, forexample, a CT scan (RMB 400), tuberculosischemotherapy (several hundred RMB) or cancer care(several thousand RMB).

A high-risk insurance system was started in 1989 inJintan county. The premium is RMB 2-3 per personper year. The insurance is intended to cover hospital-ization and will reimburse 40% of charges aboveRMB 300 up to RMB 1000, 50% of charges aboveRMB 1000 up to RMB 2000 and 80% above RMB2000, up to a ceiling of RMB 3000. Treatmentcharges are refunded, but not drugs and not high-techdiagnostics.

Jurong county - user fee based system

In Jurong county, health officials reported thatalthough the MoH advocated cooperative rural healthinsurance there was considerable resistance both fromthe doctors and from the public. The officials didnot believe that cooperative insurance was a viableoption since the healthy were reluctant to pay thepremiums and the doctors preferred more market-oriented solutions and the bonus payments that gowith it.

Patients covered by Laobao are often required to seekcare at the factory health station or hospitals con-tracted by the work unit. Gongfei patients are, in


principle, subjected to referral restrictions, but inpractice appear to have been freer in their choice offacility, although restricted within the boundaries oftheir county. Since the medical reform, co-payments(flat-rate contribution) and co-insurance (percentagecontribution), and expenditure ceilings, have been in-troduced according to local rules and applications.Gongfei is administered by the county health bureau,while Laobao is administered by the respective workunit. In Jurong county approximately 1.5% of thetotal population are covered by Gongfei.

The Jurong Office of Government Health Insuranceunder the County Health Bureau allots RMB 74 perinsuree/year of which RMB 40 is allocated from thegovernment health budget. The work units providethe remaining RMB 34 per insuree. Claims aboveRMB 40 per person/year will be reimbursed by a cer-tain percentage depending on the patient category.Expenditure above RMB 74 per person/year will becovered up to 70% by the county government andup to 30% by the work unit. The patient has to payfirst and reclaim later from the work unit.

The Laobao insurance is controlled by die work unit,which applies its own rules, but in general followsthe Gongfei system. Expenditure up to RMB 40 perperson/year is free. A co-insurance is applied foramounts above. Since the rules are decided by thework unit the system varies between townships andbetween work units. In Shishi township in Jurongcounty only inpatient visits were reimbursed. Thereimbursement rates varied from 0-100%.

Prepayment schemes for preventive services

In addition to the insurance systems there are severalprepayment systems which are interesting exam-ples of Chinese health finance innovations. There areprepayment schemes for immunizations, child healthand for perinatal care. These schemes function mainlyas savings systems, but they also include an insuranceelement. The main purpose, however, is to stimulatedemand for preventive services. In Jurong county theimmunization prepayment scheme was initiated in1987. It involves a premium of RMB 40 per child,which will guarantee free treatment should thechild fall ill with any of the target diseases, providedthat the vaccination schedule has been followed.The immunization coverage is reported to be above99%.

The immunization prepayment system in Jintan wasstarted in 1985 in Shetou township. The first system

broke down financially, however, due to an outbreakof measles in the same year. In 1986 Houyangtownship restarted the system and in 1987 it was im-plemented over the whole county. In 1986 thepremium was only RMB 5. In 1993 it had risen toRMB 60 per child. Around 95% of the eligiblechildren are enrolled. The funds, of which 20% arefor the anti-epidemic station, are managed by dietownship hospitals. The anti-epidemic station uses thefunds for cold chain operation, equipment, insuranceand bonus payments.

Jintan has a perinatal prepayment system, whereprospective parents pay RMB 37 which entitles memto care from marital advice to post-partum care.More than 80% of die couples elect to participate.Up to 12 antenatal visits and three postnatal visitsplus post-partum examination of modier and child 42days after delivery are covered. Some health instruc-tion and information materials will be provided.The cost of the delivery itself is not included, but thescheme covers insurance of the mother against deathfrom obstetrical problems (the amount in case ofdeath is RMB 2000). The fund is managed by thetownship hospital. Out of the RMB 37, RMB 8are for the MCH institute for the insurance com-ponent, for die training of doctors and for doctors'bonuses.

There is also a child health prepayment system inJintan county. Parents pay RMB 37 from the timeof die delivery and will be entided to free healdicheck-ups four times in die first year of the child'slife, two check-ups from one to diree years of ageand one check-up from four to seven years of age.The insurance component is RMB 7. The prepaymentschemes do not cover die cost of drugs, laboratorytests or treatment charges.

Escalation of health care revenueThe real per capita fee-for-service healm care revenuein Jintan county (insurance county) has grown fasterdian in Jurong county (user fee county) in the periodfrom 1986 to 1994, which is consistent witii ourhypomesis diat health insurance will reduce financialbarriers to access to care.

As shown in Figure 2, the deflated GNP per capitahas increased moderately in both counties in dieperiod 1986-1994, whereas die growm of die deflatedhealm care fee-for-service revenue (= patients' directhealdi care expenditure) per capita has clearly out-paced economic growdi in Jintan county (insurance


county) in the study period. When we compared thedevelopment of the two counties we found that thegrowth trends of the GNP per capita were similar,although the level of the deflated GNP per capita hasbeen slightly higher in Jintan (insurance county).

5000 n

4SO0

4000 '

3900 -

3000

2900-

2000

1900-I

1000 {

900

• • • JHan OMnkap

- • • Jiroig ONP*»p

—•— JHm FFSAnp

• JmwgFFSAap

f

/

/

_ ^ — — • "

90

49

40

39

30

26

20

19

It

9n

19M 1997 19U 1969 1860 1M1 1992 1993 1994

Ywr

Figure 2. Trends of real GNP and health care fee-for-servicerevenue in Jintan county (insurance) and Jurong county (user fees)

There was only a small difference of RMB 3.42 (US$0.41) per capita in 1986 in the health care fee-for-service revenue. This difference has six-doubled in1994, in constant (1986) money value. The fee-for-service revenue per capita in 1994 was 87% higherin the health insurance county than in the user feecounty. The difference is statistically significant(P<0.02, double-sided Wilcoxon test).

More curative and less preventive servicesExamining die composition of the care provision, ourfindings are serious.

The World Bank argued in its 1987 policy documentthat full-cost fees could not be introduced until healthinsurance was widely available and that the revenuesfrom user fees can be used for vital preventive andbasic curative programmes. What we see from Chinais not only that user fees have been implemented ona large-scale without health insurance being availableto the rural population, but also that the revenuesgenerated by user fees have gone to curative medicineand not to preventive medicine (Figure 3). We cansee that in both counties the (tertiary level) deflatedfunding of preventive medicine dropped until 1992,although in the same period the funding of curativemedicine grew rapidly. After 1992 the county govern-ments increased allocations to preventive care.

A benevolent interpretation of these findings wouldbe that, indeed, health insurance has contributed to

isee 1M7 19m ieee 1990 iaei 1992 1993 1994

Yur

Figure 3. Trends of real tertiary preventive funding in Jintancounty (insurance) and Jurong county (user fees)

removing financial barriers to access to care and thatthis is reflected in the higher curative care expenditurein the insurance county. It remains to be investigatedwhether the increase in health expenditure has beenfor 'desirable' care consumption, i.e. for the basicpackage of vital preventive and curative services aspromoted by the World Bank, or if it has gone toclinically unmotivated consumption of vitamins,multi-antibiotics and CT scans. From our visits tohealth facilities in the counties and from interviewswith households and providers we have reason to fearthat the expenditure growth is primarily related to thelatter category. We already know, as noted above,that real per capita expenditure on preventivemedicine decreased by 55 % in the insurance countyfrom 1986 to 1992. Figure 4 gives an indication ofwhat type of care expenditure has grown.

Figure 4. Trends of real per capita expenditure on obstetricalcare and laboratory tests, Jintan county (insurance) (RMB, 1986money value)

Rural hearth insurance in China 249

The real per capita (fee-for-service) expenditure onobstetric care in Jintan (insurance county) droppedslightly (14.1 % decrease) from 1990 to 1992, whilereal per capita expenditure on laboratory tests nearlydoubled (86.0% increase). Comparable data fromearlier and later years, and from Jurong county, werenot available.

In Jurong county (user fees) 21.1% of total county-level health care funds were used for preventiveservices in 1985 against only 6.4% in 1992. InJintan county (insurance) the share of county-levelpreventive services dropped from 8.6% in 1986 to4.0% in 1992.

The total health care funding in Jintan (insurance) in1992 (except unofficial payments and funding ofarmy and other non-MoH facilities) consumed 2.4%of the GNP of the county. Total health care fundingin Jurong (user fees) in 1992 (except funding ofarmy and other non-MoH facilities and exceptunofficial payments) was 1.4% of the GNP of thecounty, a drop from 1.6% in 1991.

Expanding tertiary curative care servicesThe funding of county (tertiary) level curative carewas increasingly higher in Jintan (health insurance)relative to Jurong (user fees). The difference in-creased by 473% in real terms and is statisticallysignificant at P<0.02 (double-sided Wilcoxon test)(Figure 5). The level of tertiary curative care revenueper capita was 286% higher than in Jurong in 1994.The growth of tertiary curative care was very rapidin Jintan (insurance) from 1986 to 1989, when itexceeded that of Jurong by 147%. Apparently thisled to some restrictive action in Jintan county, becausein 1990 the fee-for-service revenue level was broughtdown. After 1990, growth of tertiary curative carehas again shot up in Jintan.

Government health budget and the growing costof Gongfei insurance

While the government health budget in Jurong countyhas been kept at a stable nominal level since 1985,the government health budget of Jintan county hasexpanded. There could be a number of reasons forthis difference, but the considerable above-countylevel contributions to Jintan, which is not paralleledin Jurong, provides part of the explanation. Thestronger political support in Jintan county is probablyfuelled by a desire to demonstrate the success of itshealth care system.

- Jhttn tartivy

18M 1M7 1»t 1SS9 1M0

Imi

1092 IKS

Figure 5. Trends of real tertiary curative revenue in Jintancounty (insurance) and Jurong county (user fees)

The budget allocation in Jurong county for Gongfei(government) insurance was RMB 1 500 000 (1993)corresponding to RMB 74 per person per year. ActualGongfei expenditure has, however, exceeded budgetallocations consistently and increasingly over theyears. In 1992 current value Gongfei expenditure perenrolee in Jurong county was RMB 278, which isextremely high compared to the current value percapita health expenditure of RMB 34. After adjustingfor the 50% partical coverage of dependants, the percapita expenditure, RMB 170, exceeds the averageper capita expenditure by a multiple of five.

Gongfei expenditure in Jurong has grown propor-tionally more than the total health expenditure dur-ing the years 1988 to 1992. Gongfei expenditureincreased by 97.1% compared to 77.2% growth oftotal per capita health expenditure.

Both in 1992 and in 1993 the Gongfei system facedhuge deficits in Jurong. These deficits were event-ually covered by the county government. But, sincethe funds are handled by the individual work unitsthere are differences among them in financial via-bility. Not infrequently work units will resort todelays in refunding claims, with delays of up to oneyear reported. Co-insurance may also be raised.

i

Reported problems, outside the chronic deficits in thesystem, are cheating by the patients, for example byreceiving medicine which is for use by other (non-covered) persons.

Coverage by the national insurance systems is verylow in the two counties, less than 5% of the popula-tion, yet the systems consume a disproportionate


share of the total available health care resources. InJurong county 12% of total health care resources wereconsumed by Gongfei participants who constitutedless than 2% of the total population.

Drag prescriptions'Da chu fang' (big prescriptions) is an expressiondescribing the frequent overprescription practicesmotivated by revenue and bonus considerations. Thefindings have confirmed a difference in the level ofdrug sales between the two systems (Figure 6).

Figure 6. Trends of real pharmaceutical fee-for-service revenuein Jintao county (insurance) and Jurong county (user fees)

The trends of drug sales were similar until 1992, butin 1993 and 1994 growth of per capita drug sales inJintan (insurance) has been markedly higher. The percapita consumption of drugs was 67% higher in Jin-tan than in Jurong (user fees) in 1994. For both coun-ties drugs contribute a large share of total servicerevenue. In Jurong, 47.3% of the total health carefunding came from drug sales, while drug sales con-tributed 38.3% in Jintan in 1994. Salaries are low,while the price level of drugs to some extent is global.The high reliance on drug sales for revenue raisesquestions about the pricing of health services inChina; it also raises ethical questions.

SummaryWe believe this to be one of the first reported con-trolled natural experimental studies examining theimpact of voluntary health insurance on allocativeefficiency. The expenditure data from the two coun-ties confirm the hypothesis that health care expend-iture will grow faster in a fee-for-service (FFS)system with voluntary health insurance. The findings

further confirm that curative care expenditure andtertiary curative care will grow faster in a fee-for-service system with voluntary health insurance. Realper capita curative care, tertiary level curative careexpenditure and FFS revenue have grown in bothcounties, while the real per capita expenditure onpreventive programmes has remained low, in con-trast to the World Bank policy suggestions thatrevenue raised by user fees can be allocated to basiccurative and preventive care.

County (tertiary) level health care consumes most ofthe public funding in both the counties, but is likelyto be used mainly by the higher income urban popula-tion and by the small number of civil servants coveredby government insurance. Village and townshiphealth care (PHC) serves the larger rural population,but receives a small share of the public funds.

Previously reported natural experiments were mostlylacking a control group.39 They have usually com-pared demand before and after the introduction ofhealth insurance. The absence of a control group leftthe results open to the influence of various confoun-ding factors. For example, it is impossible to dis-entangle the effects of general GNP changes from theimpact of financial reform. Clearly, any study whichis not a randomized controlled trial will be subjectedto the influence of confounding factors, but webelieve that our study design has allowed us to reducesuch problems to a minimum. We believe that themost important influencing factors, such as changesin GNP per capita levels, educational levels and pro-vider incentives, have been controlled by the studydesign.

Paretian welfare economics theory postulates that thesum of individuals' health care demand will add upto the optimal societal utilization of health services.The paradox of health care demand is, however, thatwhat is perfectly rational for the individual is notnecessarily rational for the society. This is becausehealth care involves life-death decisions or at leasthealth-ill health decisions.

If a patient suffers from a fatal condition where thechance of successful surgery is very small (e.g. livercancer), it will, nevertheless, be perfectly rational forthe patient to invest all the money that he/she can layhis/her hands on, to increase the chance of survivalfrom, say, 1% to 3%. What is there to lose if thealternative is to die? From the society's point of view,investing in liver cancer surgery is not likely to be


high on the list of cost-effective health interventions,precisely because of the small likelihood of survival.Paradoxically, the smaller the chance of survival andthus ceteris paribus the smaller the cost-effectivenessfrom societal point of view, the more the individualwill be prepared to pay. If not for consideration ofthe loss to dependants, there would be no upper limitto what the patient would be prepared to pay, evenfor an infinitely small chance of survival.

This, in combination with the risk of moral hazardin healm insurance systems and the risk of supplier-induced demand in fee-for-service systems, mayexplain why the World Bank policy package of userfees, privatization, decentralization and health in-surance leads to an escalation of health care expend-iture and a shift from preventive medicine to curativeand to tertiary curative care, as witnessed in Jintancounty.

References1 Bumgarner JR et al. 1992. China Long-term Issues and

Options in the Health Transition. A World Bank CountryStudy. Washington DC: World Bank.

2 Li CR. 1992. A Study of China 's Population. Beijing: ForeignLanguages Press.

3 Ministry of Health. 1994. Guojia Weisheng Fuwu Yanjiu(Research on National Health Services), China. Beijing:Ministry of Health.

4 Parker R. 1993. personal communication.3 Piazza et al. 1992. China: Strategies for Reducing Poverty in

the 1990s. A World Bank Country Study. Washington DC:World Bank.

6 China Statistical Yearbook, 1994. Beijing.7 Henderson G, Akin J, Li ZM, Jin SG, Ma HJ, Ge KY. 1994.

Equity and the utilization of health services: report of an eight-province survey in China. Soc. Sri. Med. 39(5): 687-99.

8 Shi LY. 1993. Health care in China: a rural-urban comparisonafter the socio-economic reforms. BulL World Health Organ.71(6): 723-36.

9 Sidel VW. 1993. New lessons from China: equity and econ-omics in rural health care. Am. J. PubL Hlth 83(12): 1665-6.

10 Liu G, Liu XZ, Meng QY. 1994. The privatization of themedical market in Socialist China: A historical approach.Health Policy 27: 157-74.

11 Qi H. 1993. Shichang Jingji Yu Yide Yifeng Jianshe (MarketEconomy and the Upholding of Medical Ethics). JiankangBoo (Health News), 10-15. Beijing.

12 Sidel VW. 1982. Medical care in China: equity vs moderniz-ation. Am. J. Publ. Hlth. 72(11): 1224-6.

13 Feng XS, Tang SL, Bloom G, Segall M, Gu XY. 1995.Cooperative medical schemes in contemporary rural China.Soc. Sri. Med. 41(8): 1111-8.

14 Chen XM, Hu DW, Lin ZH. 1993. The rise and decline ofthe cooperative medical system in rural China. Int. J. HealthServ. 23(4): 731^*2.

13 Zhou SQ, Gu XY. 1991. Guanyu Fazhan Woguo Nongcun JiziYiliao Baojian Zhidude Jianyi (Concerning the Proposal toDevelop Our National System for Rural Collective Health

Insurance). Zhongguo Nongcun Yiliao Baojian Zhidu Yanjiu,Shanghai Yike Jishu Chubanshe, Shanghai.

16 Gong YL. 1994. personal communication.17 Piyaratn P, Janjaroen WS. 1994. Health Insurance in Thailand.

Paper presented at the National Seminar on Health SectorDevelopment, 4 -7 October 1994, Hanoi, Vietnam.

18 Abel-Smith B. 1993. The Health Insurance System in Vietnam.Assignment Report for SIDA, Stockholm.

19 Arhin D. 1994. The health card insurance scheme in Burundi:A social asset or a non-viable venture? Soc. Sri. Med. 39(6):861-70.

20 McPake B, Hanson K, Mills A. 1993. Community financingof health care in Africa: an evaluation of the Rflmnlrn Initiative.Soc. Sri. Med. 36(11): 1383-95.

21 Litvack IJ, Bodart C. 1993. User fees plus quality equalsimproved access to health care: results of a field experimentin Cameroon. Soc. Sri. Med. 37(3): 369-83.

22 World Bank. 1994. Better Health in Africa. Washington DC:The World Bank.

23 Wouters AM. 1995. Improving quality through cost recoveryin Niger. Health Policy Plan. 10(3): 257-70.

24 Akin JS, Birdsall N, de Ferranti D . 1987. Financing HealthServices in Developing Countries: An Agenda for Reform.Washington DC: The World Bank.

23 Kutzin J, Barnum H. 1992. Institutional features of health in-surance programs and their effects on developing countryhealth systems. Int. J. Health Plann. Manage.

26 Gu XY, Bloom G, Tang SH, Zhu YY, Zhou SQ, Chen YB.1993. Financing health care in rural China: Preliminary reportof a nationwide study. Soc. Sri. Med. 36(4): 385-91 .

27 OECD. 1992. The Reform of Health Care: A ComparativeAnalysis of Seven OECD Countries. Paris: OECD.

28 The Economist. 1992. A Survey of China. November 28th.London.

29 Agarwala R. 1992; China: Reforming Intergovernmental Fiscalrelations. World Bank Discussion Papers; 178. WashingtonDC: The World Bank.

30 UNICEF Office for China. Annual Reports 1986 and 1987.Beijing.

31 Piazza et al. 1992. op. cit.32 China in Brief: State Employees' Medical Care System. 1992.

Beijing: New Star Publishers.33 Zheng X, Hillier S. 1995. The reforms of the Chinese health

care system: county level changes: The Jiangxi Study. Soc.Sri. Med. 41(8): 1057-64.

34 Shanghi Medical University. 1990. Fuyou Weisheng Fuwu YuJingfei Yanjiu (MCH services and expenditure research).Research report of MoPH MCH Project Office, Beijing.

33 Manning WG et al. 1987. Health insurance and the demandfor medical care: evidence from a randomized experiment.The American Economic Review 77(3):251-77.

36 Peabody JW, Schmitt JM. 1992. Medical equipment in thePeople's Republic of China: A survey of medical equipmentprocurement, utilization, and maintenance in university-affiliated hospitals between 1976 and 1987. International Jour-nal of Techno Assessment in Health Care 8(1): 138-49.

37 Banta HD, Chen J. 1993. Medical Technology in China. In:Banta HD, Luce BR. Health Care Technology and its Assess-ment: An International Perspective. Oxford: Oxford MedicalPublications.

38 Ron A, Abel-Smith B, Tamburi G. 1990. Health Insurance inDeveloping Countries: the Social Security Approach. Geneva:ILO.

39 Manning. 1987. op. cit.

252

AcknowledgementsThis study was financially supported by the Swedish Agency forResearch Cooperation with Developing Countries (SAREC).Comments on an earlier draft by Professor G6ran Sterky, DrMonica Johnsson and by an anonymous referee are gratefullyacknowledged.

BiographiesLennart Bogg, MBA, BA (Sinology), is a Financial Controller atSIDA, Stockholm, and is registered with MCAR as a doctoralresearch student in international health systems research. He iscurrently coordinator of research collaboration with China andThailand. He served with UNICEF in Myanmar and China for6 years in the mid-1980s and with UNRWA in Gaza in 1990-91.

Dong Hengjin, MA, MMSc, is an associate professor in the Depart-ment of Health Statistics and Community Medicine, School ofPublic Health, Shanghai Medical University. He has published

Lennart Bogg et al.

more than 20 articles in China on health economics and healthfinance, and has participated in World Bank funded health finan-cing and health care manpower planning research.

Wang Keli, MD, MS, is an associate professor and Deputy Directorof the Department of MCH, School of Public Health, ShanghaiMedical University. He has participated in UNICEF funded MCHcare utilization research in 300 poor counties in China.

Cai Wenwei is a professor and Director of the MCH Departmentand Vice Dean of the School of Public Health, Shanghai MedicalUniversity. She is principal investigator in UNICEF funded MCHcare utilization studies in 300 poor counties in China.

Vinod Diwan, MD, DTMedSc, MPH, is an associate professorat IHCAR, Karolinska Institutet, Stockholm. He is principal in-vestigator in international health systems research collaborationwith Vietnam, Laos, Thailand, China, India and other countries.

Correspondence: Lennart Bogg, IHCAR, Karolinska Institutet,S-17177 Stockholm, Sweden.

Documents

The cost of coverage: rural health insurance in China