q The Author 2005. Published by Oxford University Press in association with The London School of Hygiene and Tropical Medicine. All rights reserved.
doi:10.1093/heapol/czi018
The impact of China’s retail drug price control policy on hospital
expenditures: a case study in two Shandong hospitals
QINGYUE MENG,1,2 GANG CHENG,1 LYNN SILVER,2,3 XIAOJIE SUN,1 CLAS REHNBERG4 AND GÖRAN TOMSON2,4
1
Center for Health Management and Policy, Shandong University, Jinan, China, 2Division of International
Health, Karolinska Institute, Stockholm, Sweden, 3Pharmaceutical Sciences, School of Health Sciences,
University of Brası́lia, Brazil and 4Medical Management Center, Karolinska Institute, Stockholm, Sweden
In China, 44.4% of total health expenditures in 2001 were for pharmaceuticals. Containment of
pharmaceutical expenditures is a top priority for policy intervention. Control of drug retail prices was
adopted by the Chinese government for this purpose. This study aims to examine the impact of this
policy on the containment of hospital drug expenditures, and to analyze contributing factors.
This is a retrospective pre/post-reform case study in two public hospitals. Financial records were
reviewed to analyze changes in drug expenditures for all patients. A tracer condition, cerebral infarction,
was selected for in-depth examination of changes in prices, utilization, expenditures and rationality of
drugs. In the two hospitals, a total of 104 and 109 cerebral infarction cases, hospitalized respectively
before and after the reform, were selected. Prescribed daily dose (PDD) was used for measuring drug
utilization, and the contribution of price and utilization to changes in drug expenditures were
decomposed. Rationality of drug use post-reform was reviewed based on published literature.
Drug expenditures for all patients still increased rapidly in the two hospitals after implementation of the
pricing policy. In the provincial hospital, drug expenditures per patient for cerebral infarction cases
declined, but not significantly. This was mainly attributable to reduced utilization. In the municipal
hospital, drug expenditure per patient increased by 50.1% after the reform, mainly due to greater drug
utilization. Three to five fold higher drug expenditure per inpatient day in the provincial hospital was
due to use of more expensive drugs. Of the top 15 drugs for treating cerebral infarction cases after the
reform, 19.5% and 46.5% of the expenditures, in the provincial and municipal hospitals, respectively,
were spent on drugs with prices set by the government. A large proportion of expenditures for the top
15 drugs, at least 65% and 41% in the provincial and municipal hospitals, respectively, was spent on
allopathic drugs without an adequate evidence base of safety and efficacy supporting use for cerebral
infarction.
Control of retail prices, implemented in isolation, was not effective in containing hospital drug
expenditures in these two Chinese hospitals. Utilization, more than price, determined drug
expenditures. Improvement of rational use of drugs and correcting the present incentive structure for
hospitals and drug prescribers may be important additional strategies for achieving containment of drug
expenditures.
Key words: China, drug price, policy, hospital, expenditure
Introduction
The rapid rise of medical and pharmaceutical expenditures has
become a critical barrier to health care for the poor and its
control is a key objective for health policy makers (Bloor and
Freemantle 1996; Centre for Health Statistics and Information
1998; Wu and Yang 1999; Maynard and Bloor 2003). China is
no exception. Pharmaceutical expenditures have grown at a
rate well above overall economic growth for the past two
decades (Wei 1999). Drug expenditure per capita in real terms
increased from 36.6 yuan in 1990 to 184.3 yuan in 2001
(US$1 ¼ 8.2 Chinese yuan), an annual rate of increase of
15.7% (Zhao et al. 2003).
Internationally, a number of approaches have been used for
containing drug expenditures. These approaches can be
divided between those which affect the supply of drugs on
the market and those addressing the demand generated by
health professionals and consumers for pharmaceuticals
(Figure 1) (Gross et al. 1996; Ess et al. 2003). Conflicts of
interest in the operation of health systems, such as the case of
physicians who both prescribe and dispense, have been
associated with higher drug utilization and expenditures, and
the separation of the two activities has been a longstanding
feature of codes of ethics of the medical profession in many
countries. It is, however, still the norm in other countries,
186
Qingyue Meng et al.
the average of the industrial sector (MoH 2000). The number
of pharmaceutical manufacturers rose from 3097 in 1990 to
5396 in 1997 (China State Economic and Trade Commission
1999). By the end of the 1980s, the shortage of pharmaceuticals that had existed during the planned economy era had
been fully resolved (Dong et al. 1999).
Figure 1. Internationally used approaches to controlling drug
expenditures
and may be increasing (Ess et al. 2003). Less research is
available on other institutional arrangements where hospital or
prescriber income is highly dependent on drug sales.
Development of straight salary compensation for pharmacists
and public interest drug distribution and retailing systems in
some countries, such as Sweden, have also helped to eliminate
conflicts of interests in the pharmacy setting.
In China, a number of measures have been taken by policy
makers to slow the growth rate of drug expenditure. These
have included the use of a drug list and consumer cost-sharing
in social health insurance schemes, capping the annual growth
rate of incomes of hospitals, controlling prices of pharmaceuticals, and regulating mark-ups. Drug lists and modified
payment systems have been recommended and implemented
since the mid-1990s, and have been confirmed as effective in
controlling rapid increases in drug expenditures for health
insurance schemes (Hu et al. 2001; Yip and Eggleston 2001).
However, the slower growth rates of drug expenditures in
health insurance schemes is of limited significance in the total
societal expenditures for drugs, due to the low coverage of
health insurance in China (Wu 1999).
Most recently, direct control of retail prices of pharmaceuticals has become the key government strategy for constraining drug expenditures (MoH 2000). This exploratory study
aims to examine whether the retail price policy was actually
implemented in the institutions studied, whether it was
effective in containing hospital drug expenditures, and the
role of rationality of drug use in influencing the expenditure
patterns identified for a tracer condition.
Overview of the Chinese pharmaceutical sector
and the context of the drug retail price control
policy
Extensive market-oriented economic reforms initiated in the
late 1970s in China substantially influenced both the
production and social sectors. In line with rapid economic
growth, the health care and pharmaceutical sectors expanded
rapidly. Between 1978 and 1997, the real gross pharmaceutical product increased by 17.6% annually, 4.4% higher than
Private finance has become the main source for health
expenditure over the past two decades. In 2001, 515 billion
Chinese yuan was spent on health in China, accounting for
5.4% of gross domestic product (Zhao et al. 2003). Of total
health expenditure, 15.5% came from the government health
budget, 24.0% from employers and 60.5% from private
individuals (Zhao et al. 2003). Health insurance schemes
cover only about 14% of the total population, and the
uninsured pay for medical services out-of-pocket (China
Health Economics Institute 1999). Pharmaceutical expenditures were 230.3 billion yuan in 2001, accounting for 44.4% of
total health expenditures (Zhao et al. 2003), which was much
higher than that in either developed countries (OECD
countries, 10.3– 18.5% in 1997) or middle-income countries
(for example, Argentina 20.7% and Chile 15.7% in 1997)
(Zerda et al. 2002; Maynard and Bloor 2003). Of total drug
expenditures in China, 54.1% were incurred in hospital
outpatient departments and clinics, 30.8% in inpatient care
and 15.1% in pharmacy stores (Zhao et al. 2003). Hence, the
hospital sector is the main retail supplier of drugs. Drugs have
become a major source for financing public hospitals and have
been the most profitable fee item in hospitals since the early
1980s. The net income for public hospitals from selling drugs
was 36.7 billion yuan in 1997 (China State Economic and
Trade Commission 1999). Given the widespread use of
revenue-based bonus systems for both clinical departments
and doctors in public hospitals, the latter have strong financial
incentives to prescribe greater quantities or more expensive
drugs.
Under the planned economy system, drugs were directly
distributed from state-owned drug wholesalers. From the early
1980s on, pharmaceutical manufacturers became free to
develop a large sales force linking them to hospitals,
pharmacies, prescribers and consumers. The number of
pharmaceutical wholesalers and facilities increased from
4554 in 1993 to 16 519 in 1997 (Chinese Pharmacy Yearbook
Editing Commission 1998). Currently, hospitals and pharmacy stores can purchase drugs either from wholesalers or
directly from manufacturers.
Between 1980 and 2000, the government controlled the entire
cascade of drug prices, from manufacturers’ exit prices, to
wholesale and retail prices. Manufacturers’ exit prices were
based on production cost plus a 5% mark-up, to which a 15%
mark-up was added for the wholesale price, and addition of a
further 15% mark-up constituted the retail price (China State
Commission of Planning and Development 1998). However,
faced with the rapid expansion of the pharmaceutical sector
and asymmetry of access to cost information between price
regulators and manufacturers, the government was unable to
generate the necessary cost estimates for setting appropriate
exit prices. Furthermore, since mark-ups for both wholesalers
and retailers, including hospitals, were a fixed percentage,
China’s drug price controls and hospital drug expenditures
expensive drugs were preferred by both. In order to attract
wholesalers and hospitals to their products, manufacturers
requested higher prices. Under this system, drug prices were
thought to be unreasonably high (Hu and Li 2001; Du 2002;
Wang and Wei 2003).
187
containing expenditures might be blocked by competing
priorities, and that hospitals might naturally seek to maintain
or increase revenues.
Methods
In late 2000, seeking to contain the rapid increase in
pharmaceutical expenditures and to improve the rational use
of drugs, the Chinese government changed its drug pricing
policy from controlling the entire cascade of prices for all
pharmaceuticals to controlling retail prices for selected
products only. There were three key rationales for the new
drug pricing policy (China State Commission of Planning and
Development 2000). First, retail price is the final price charged
to health care users. If retail prices could be effectively
controlled, consumer access would improve. If retail prices
were capped, manufacturers, wholesalers and retailers,
including hospitals, would negotiate among themselves the
distribution of profits. Secondly, it was more feasible for the
government to set reasonable prices for the smaller number of
drugs selected, a sub-set of frequently used and relatively costeffective drugs. This strategy would also considerably reduce
the regulatory burden for the government. Policy-makers
believed that if the prices of the selected drugs, which account
for a large proportion of utilization, were controlled, users
would stand to benefit. Thirdly, it was believed that price
competition would be created with the new policy. Since
overall drug prices were thought too high, and the prices of
drugs set by the government would be decreased, it was
believed that the prices of drugs set by the market might also
decrease in order to compete. Given a constant drug utilization
pattern, decreases of overall drug prices would reduce
expenditures to users.
The State Commission of Development and Planning decided
to set retail prices for drugs listed under the Urban Health
Insurance Scheme, because these were believed essential and
frequently used. There are two parts to this list, A and B. Prices
of Part A drugs are set by the central government and are
definitive ceilings for retailers. The central government also
sets guiding prices for Part B drugs, which are used by the
provincial governments. Provinces can set price ceilings 5%
higher or lower than the central guiding prices for Part B
drugs. All retail prices charged to the users must be lower than
these ceilings. The government declared that retail prices
should be reduced by an average of 15% before the end of
2001 (China State Drug Administration 2003).
The drug pricing policy is monitored through periodic checks
in facilities and regulated by Departments of Price Administration at county, provincial and central government levels, in
collaboration with Departments of Health and of Drug
Administration. Drug retailers were asked to strictly follow
and to publicly post the new drug prices. Patients have the
right to question the prices of drugs prescribed and can report
abuses to the Department of Price Administration or to the
local Consumers’ Protection Association.
The survival of Chinese public hospitals relies on market
revenues and they have autonomy in allocating surpluses. It
was therefore of concern that the policy’s objective of
Selection of hospitals
This is a retrospective pre/post-reform case study. The study
was conducted in two public hospitals located in the capital
city of Shandong Province, a provincial hospital (one of the
largest in the province, with over 1 million outpatient visits
annually), and a municipal hospital. In China, public hospitals
dominate the provision of medical services and generate a
large proportion of drug expenditures.
The hospitals were purposively selected to represent different
hospital types. The provincial hospital is directly managed by
the provincial Department of Health and is one of 10 model
hospitals in China. Model hospitals are selected by the
Ministry of Health as examples for administration and service
quality. The municipal health authority directly manages the
municipal hospital.
Selection of a medical condition and patients’ records
Criteria for choosing a tracer diagnosis were that the condition
should, in the Chinese context, be relatively intensive in
consumption of drugs, with a wide range of types and
substitutes; have inpatient care with a relatively long hospital
stay; and have a stable composition of the case mix in a short
time period. A panel of medical experts was organized to
select the diagnosis. Cerebral infarction (ICD-10 code I63)
was selected as an appropriate tracer condition for drug
expenditure.
In both hospitals, the majority, but not all, cases were
confirmed by computerized tomographic examinations.
Because the numbers of patients with the sole diagnosis of
cerebral infarction were limited, cases with the secondary
diagnoses of hypertension (essential hypertension) (ICD-10
code I10) or atherosclerotic heart disease (coronary) (ICD-10
code I25.1) were also selected. Cases with additional
complications were excluded, to reduce variation in disease
severity between the two time periods.
The drug retail price control policy was implemented in late
2000. Because drug expenditure might not yet have been
affected in 2001, the cases were selected from the beginning of
2002 for the post-reform period and from the end of 2000
backwards for pre-reform cases. All cases meeting the
diagnostic criteria after 1 January 2002 were selected. In the
provincial hospital, records of 49 cerebral infarction cases
were selected, and in the municipal hospital, 55 cases were
selected. In the provincial hospital, 49 cases from prior to the
implementation of the policy were selected, and in the
municipal hospital, 60 such cases. A total of 213 medical
records of cerebral infarction patients were reviewed. The
study periods covered January to August 2002 (post-reform),
and May to December 2000 (pre-reform).
188
Qingyue Meng et al.
Data collection
Hospital medical records selected were copied by the
investigators from Shandong University and relevant data
were recorded. Indicators extracted included patients’ age,
gender, occupation, insurance status and source, condition
when admitted, diagnoses, expenditures by types of drugs
(Western and Chinese drugs), and the names and quantities of
all drugs prescribed.
In addition, hospital financial records were reviewed to extract
the number of outpatient visits and inpatient days, hospital
income and expenditures by items, and numbers of hospital
staff and beds. Local authorities from the government agencies
involved were also interviewed to obtain information on the
implementation of the new policy.
Data analysis
For examining the comparability of cases before and after the
policy implementation, severity of selected cases was
compared using four indirect indicators: age, health condition
when admitted, and the secondary diagnoses of hypertension
and of coronary artery disease.
Prescribed daily dose (PDD) was employed to measure
utilization of drugs instead of defined daily dose (DDD).
DDD is the unit recommended by WHO for measuring drug
utilization. However, in this study, many Chinese medicines
were prescribed which had no ATC classification or DDD.
PDD is the actual number of daily doses prescribed per capita
or patient (Capella 1993). The PDD of each drug was
calculated using formula (1), where PDDi represents the PDD
of drug i, Qi represents the amount of drug i used per day in
prescription i which involves drug i, and Fi represents period
(days) of prescription i.
P
QF
ð1Þ
PDDi ¼ P i i
Fi
For examining the effects of changes in price and utilization
on changes in drug expenditures, the price and utilization
effects were decomposed. Compared with drug expenditures
in period 1 (prior to reform), the percentage change of drug
expenditures in period 2 was calculated using formula (2):
X% ¼ ½ðDE2 2 DE1 Þ=DE1 p 100%
ð2Þ
where X% is the percentage change of drug expenditures, DE1
represents the drug expenditures in period 1 and DE2 is the
drug expenditures in period 2.
Drug expenditure is the product of drug price and utilization.
The percentage change in drug expenditure between the two
periods would be determined by changes in utilization, price,
or both, and the introduction of new drugs. The effects of the
above can be decomposed with formulae (3) – (5), based on
formulae from William Cleverley (Cleverley 1992):
PI% ¼ ½ðSP2i p Q1i =SP1i p Q1i Þ 2 1 p 100%
ð3Þ
UI% ¼ ½ðSP1i p Q2i =SP1i p Q1i Þ 2 1 p 100%
ð4Þ
JI% ¼ X% 2 ðPI% þ UI%Þ
ð5Þ
where PI% measures the percentage change in drug expenditures due to change in drug prices; UI% measures the
percentage change in drug expenditures due to change in drug
utilization; JI% measures the residual of percentage change in
drug expenditures that are not explained by percentage
changes of price and utilization, including joint effect of
both price and utilization and introduction of new drugs; P1i
and P2i stand for prices of drug i in period 1 and 2; Q1i and Q2i
stand for utilization (PDDs) of drug i in period 1 and 2.
During the study period, hospitals stopped using some drugs
in period 2. Using formulae (3) and (4), if a drug was used
only in period 1, the effect on percentage change was
determined by its levels of price or utilization compared
with the average.
Since the computerized price systems were updated regularly,
prices of some drugs were no longer available at the time of
data collection. In the provincial hospital, it was possible to
verify prices for about 35% of drugs prior to the reform and
about 45% after the reform; and for about 25% in the
municipal hospital prior to the reform and 40% after the
reform. Drugs with prices were, however, the most frequently
used ones and accounted for the majority of the total drug
expenditures. In the provincial hospital, PDDs of drugs with
prices accounted for 89.5 and 89.1% of total PDDs, and
expenditures for the drugs with prices accounted for 86.5 and
87.8% of the total drug expenditures before and after the
reform, respectively. In the municipal hospital, PDDs of drugs
with prices accounted for 89.6 and 89.4% of total PDDs, and
expenditures of the drugs with prices accounted for 78.8 and
79.6% of total drug expenditures before and after the reform,
respectively. Calculations are based only on those drugs for
which data were available. Since the drug price index was
minus 4% between 2000 and 2002 (China State Commission
of Planning and Development 2002), the monetary values of
drug expenditures were not adjusted for inflation.
The rational use of drugs
To identify contributing factors to drug expenditures related to
rational use, the top 15 drugs prescribed after the new price
controls, by expenditure, were analyzed. They were annotated
as prescribed, by brand or generic name. Traditional Chinese
drugs were not analyzed regarding the rationality of their use.
The evidence base on safety and efficacy of the ‘Western’
drugs (including some products developed in China but based
on Western pathophysiological concepts) was analyzed and
classified as follows:
(1)
(2)
drugs and electrolyte solutions which are generally
recognized as safe and effective for cerebral infarction
and/or for common secondary diagnoses, such as
infection, in hospitalized patients were denominated
‘good’, and included glucose saline, ceftriaxone (as
Rocephin), azithromycin, clindamycin, isotonic sodium
chloride, cefotaxime and pantoprazole;
drugs which are useful in limited indications for cerebral
infarction but which may be overused, such as mannitol,
recommended only in cases of cerebral edema, were still
denominated ‘good’;
China’s drug price controls and hospital drug expenditures
(3)
(4)
drugs which have some evidence of effectiveness but
which are not currently considered consensual evidencebased treatment, such as Nicholin (citicholine)
(Adibhatla and Hatcher 2002) and defibrase (Bell
1997), were denominated ‘some’; and
drugs which are not recognized as safe and effective for
cerebral infarction, or for other common accompanying
diagnosis in hospitalized patients, were denominated
‘weak’ (Annex 1).
Drugs were considered as not recognized as safe and
efficacious if they were not part of consensus guidelines for
cerebral infarction (Coull et al. 2002; Adams et al. 2003;
European Stroke Initiative 2003), and if a review of
publications indexed in Medline, from 1966 on, did not reveal
a favourable balance of well-designed clinical trials supporting their use for this indication. The drugs cerebroprotein
hydrolysate, kakonein and defibrase, absent from Medline, but
primarily developed in China, were also reviewed on the
Chinese Qinghua University database, CNKI (see Annex 1).
Defibrase, a snake venom, had no articles on Medline;
however, 111 on its use in cerebral infarction were found on
CNKI, with some modestly sized, favourable RCTs. Ancrod,
a related but apparently not identical venom from snakes of
the same genus of Agkistrodon but a different species
(A. rhodostoma rather than A. acutus), has significant trials
cited in Medline with positive results. Defibrase was therefore
classified as having ‘some’ evidence. For combination drugs,
evidence on the combination was sought.
Results
Characteristics of expenditures for all patients and implementation of the drug control policy are first described, presenting
changes in drug expenditure growth rates for all patients
before and after implementation of the price control policy.
With the tracer condition, drug expenditures, effects of price
and utilization on drug expenditures, and rationality of use for
patients are then presented.
189
Expenditure trends for all patients and implementation
of the retail price control policy
From 1998 to 2000, prior to the pricing reform, total revenues
increased by 39.7% in the provincial and 42.0% in the
municipal hospital (Table 1). From 2000 to 2002, total
revenues increased by 46.6% in the provincial and by 69.1% in
the municipal hospital. From 1998 to 2000, drug revenues for
the provincial hospital increased by 32.5% and for the
municipal hospital by 37.9%. From 2000 to 2002, drug
revenues for the provincial hospital increased by 42.9% and
for the municipal hospital by 49.5%.
While part of these greater revenues was due to a higher
patient load, the corrected values still rose, most markedly in
the outpatient sector. From 1998 to 2000, drug revenues per
outpatient visit and per inpatient day increased by 57.9 and
3.2% in the provincial hospital and by 30.6 and 21.4% in the
municipal hospital, respectively. From 2000 to 2002, drug
revenues per outpatient visit and per inpatient day increased
by 37.9 and 22.2% in the provincial hospital and by 30.7 and
32.3% in the municipal hospital, respectively. Major differences in drug expenditures between hospitals were evident.
These were 5.2-fold higher in the provincial hospital at
baseline, and even when correcting for length of stay, were
4.3-fold higher.
Drug prices in the two hospitals were monitored by the same
level of the Departments of Price Administration as well as of
Health and Drug Administration, generally every 2 months.
When interviewing staff in Provincial and Municipal Departments of Price Administration, investigators were told that the
hospitals studied had properly adjusted the drug prices
according to the changes made by the government. The drug
price recording systems in the two hospitals were checked
with the updated price list of the government; this confirmed
that the prices of drugs regulated by the government had been
updated. It was found that the changes in prices of drugs for
treating cerebral infarction cases were consistent with
implementation of the drug pricing policy. In the provincial
hospital, on average, prices of drugs set by the central
government (Part A) decreased by 6.9% and prices of drugs set
Table 1. Characteristics of hospitals studied for all patients
Indicator
No. of hospital beds
No. of outpatients (1000 s)
No. of inpatients
Average length of stay (days)
Total revenue (million yuan)
Share of total drug revenues (%)
Share of outpatient drug revenue (%)
Share of inpatient drug revenue (%)
Drug revenue per outpatient (yuan)
Drug revenue per inpatient (yuan)
Drug revenue per inpatient day (yuan)
Provincial hospital
Municipal hospital
1998
2000
2002
1998
2000
2002
1 275
985.4
23 419
16
240.9
59
20
39
49
4 026
271
1 380
974.3
24 579
17.5
336.5
56
22
34
76
4 632
280
1 392
1 005.1
29 699
16
493.4
54
21
33
104
5 490
342
110
86.3
5 989
12
18.34
42
17
25
36
775
63
112
104.2
6 362
12
26.0
41
19
22
47
909
76
112
126.6
6 658
12
44.0
36
18
19
61
1 230
100
190
Qingyue Meng et al.
by both central and provincial governments (Part B) decreased
by 17.6%. In the municipal hospital, prices of Part A and Part
B drugs decreased by 17.4 and 11.3% on average, respectively. These changes in prices were close to the declaration
made by the government that an average of 15% reduction
would be expected after the control of retail prices.
In the provincial hospital, a computerized price checking
system was available at the hospital main entrance, with which
users could check prices for drugs and professional services.
In the municipal hospital, prices of common drugs were posted
on the wall near the cashier’s window. In addition to external
price monitoring, internal control systems had been created at
both hospitals. Two staff members in the provincial and one in
the municipal hospital were responsible for verifying prices
charged to patients and assuring implementation of official fee
schedules. From early 2002, the provincial hospital started to
use a ‘daily accounting system’, which meant inpatients would
be informed of their expenditures on drugs and other services
every day of their stay and could question expenditures. The
municipal hospital still used the traditional method, presenting
bills at discharge.
Drug prices and expenditure per patient for
cerebral infarction
Before comparing the drug expenditures on cerebral infarction
cases, indicators of potential severity of cases in the two time
periods were evaluated (Table 2). Except for age in the
provincial hospital, no significant differences in severity
related indicators were found between the two time periods.
Drug expenditures between age groups in the provincial
hospital were compared using the Spearman Test method,
which showed no significant relationship between age and
drug expenditures (R ¼ 20.0574, p ¼ 0.405). While there
was a marked difference in length of stay between hospitals, it
is not clear if this was due to variation in hospital practice or to
patient severity.
To examine the impact of implementation on hospital drug
expenditures, expenditures per patient for cerebral infarction
patients before and after implementation of the new policy
were examined (Table 3). There were no statistically
significant changes in drug expenditures per inpatient in the
provincial hospital between the two periods. In the municipal
Table 2. Comparison of cerebral infarction cases in the two time periods,
2000 and 2002
Indicators
Age (mean)
Condition when
admitted (%)
(1) Emergency
(2) General
With hypertension (%)
Yes
With coronary (%)
Yes
Provincial
Municipal
2000
2002
p value
2000
2002
p value
59
64
0.045
63
65
0.418
53
47
49
51
0.838
50
50
36
64
0.169
61
65
0.675
60
55
0.577
12
19
0.408
17
9
0.276
Table 3. Drug expenditures (yuan) and length of stay per patient with
cerebral infarction
Indicators
Provincial hospital
Length of stay (days)
Total expenditure per patient
Drug expenditures per patient
Western
Chinese
Municipal hospital
Length of stay (days)
Total expenditure per patient
Drug expenditures per patient
Western
Chinese
2000
2002
20
10 327
7 931
6 215
1 716
8
6
5
1
10
2 193
1 392
879
483
Difference
p value
19
873
673
312
361
21
21 454
21 258
2903
2355
0.589
0.358
0.287
0.387
0.205
11
2 878
2 090
1 756
334
1
685
698
877
2149
0.355
0.014*
0.002**
0.001**
0.066
*Significance at 0.05 level; **significance at 0.01 level.
hospital, total expenditure per inpatient rose by 50.1%, driven
by skyrocketing expenditures for Western drugs, up 97.9%
between the two periods. Expenditures for Chinese medicines
decreased, but not significantly.
As in the overall data for all patients, the provincial hospital’s
care for cerebral infarction patients was associated with far
higher drug expenditures than at the municipal hospital, 5.7fold higher per patient before the reform, with the difference
falling to 3.2-fold higher per patient after the reform. In the
provincial hospital, length of stay, total expenditures and drug
expenditures for cerebral infarction cases declined, a pattern
distinct from the hospital’s overall trends, but these changes
were not statistically significant. In the municipal hospital, the
trends for total expenditures and drug expenditures for
cerebral infarction cases were consistent with the trends for
the general inpatient population.
Relation of expenditures to insurance status of cerebral
infarction patients
Before and after the reform, the proportion of cerebral
infarction patients covered by health insurance was 77.5 and
72.3% in the provincial hospital and 10.0 and 12.7% in the
municipal hospital, respectively. There were no significant
shifts in health insurance coverage within each hospital
between the periods, though far more provincial hospital
patients were insured. In the provincial hospital, drug
expenditures per insured and uninsured inpatient were
7325.2 and 8199.0 yuan (t ¼ 20.501 and p ¼ 0.6379),
respectively, in 2000. Conversely, in 2002, the drug
expenditure per insured patient (5163.8 yuan) was lower
than that of the uninsured (7313.3 yuan) (t ¼ 21.141 and
p ¼ 0.001). There were no significant differences in drug
expenditures between insured and uninsured patients in the
municipal hospital in either time period.
To explore explanations for the significant difference in drug
expenditure per inpatient between the insured and uninsured in
the provincial hospital in 2002, prices and utilization of drugs
and hospital length of stay were compared. The average prices
of drugs were 31.1 yuan per PDD for the insured and 31.3
China’s drug price controls and hospital drug expenditures
191
Table 4. Utilization of drugs per patient and per inpatient day and price per prescribed daily dose (PDD), cerebral infarction cases
Indicator
Total PDDs (all drugs)
% of total PDDs represented by
top 15 drugs
PDDs per patient
PDDs per inpatient day
Number of types of drugs used
Average price per PDD (yuan)
Provincial hospital
Municipal hospital
2000
2002
Change (%)
2000
2002
Change (%)
12 066
10 217
215
7 328
8 081
10
52
246
12
200
32
59
209
11
207
31
6
215
211
4
22
67
122
12
155
11
60
147
13
181
14
28
20
9
17
25
yuan per PDD for the uninsured, not a significant difference
(t ¼ 20.136 and p ¼ 0.893). There was also no significant
difference in hospital length of stay between the insured (18
days) and uninsured (19 days) (t ¼ 20.592 and p ¼ 0.557).
However, drug utilization, at 166 PDDs per inpatient for the
insured and 234 PDDs for the uninsured, showed a significant
difference (t ¼ 21.141 and p ¼ 0.001). The difference in
drug expenditure between the insured and uninsured mainly
resulted from a difference in the quantity of drug utilization.
Changes in drug utilization and price for cerebral
infarction patients
In the provincial hospital, both total utilization (PDD) and
utilization per patient and per inpatient day for cerebral
infarction fell in the second period (Table 4). PDDs used per
inpatient day and per patient decreased by 10– 15% in the two
periods. Price per PDD decreased slightly.
In contrast, utilization of drugs increased by 10% in the second
period in the municipal hospital. Utilization of drugs per
patient and per inpatient day increased by 9 –20% between the
two time periods. Price per PDD increased by nearly 25%.
The utilization levels in PDD per day were relatively similar
between the two hospitals, but the average price per PDD was
very different, a result which may explain the higher overall
expenditures per inpatient day in the provincial hospital. The
provincial hospital charged 2.8 times more than the municipal
hospital per PDD at baseline, falling to 2.2 in the second
period. In the provincial hospital, price per PDD decreased by
1.2 yuan; in the municipal hospital, it increased by 2.8 yuan. In
short, after implementation of the policy, unit price had
decreased slightly in the provincial hospital and increased in
the municipal hospital for cerebral infarction cases.
factors (Table 5). It is clear that for both hospitals and for
changes in either direction, the expenditure level was more
sensitive to changes in utilization than to price changes.
Drug utilization patterns and the evidence base for drugs
used in the two hospitals
The preceding section suggests that utilization rather than
price was more influential in determining drug expenditures
for the tracer condition. The utilization patterns of drugs were
further analyzed to examine what types of drugs were used for
treating cerebral infarction patients and differences between
the two hospitals. Table 6 shows the top 15 drugs in terms of
expenditures, by hospital, for the post-reform period, with
their prices per PDD. Expenditures on these drugs accounted
for 78.4% of total drug expenditures for the provincial hospital
and 62.4% for the municipal hospital.
Of the top 15 drugs used in the provincial hospital, one drug’s
price was set by the central government and the prices of four
drugs were set by both central and provincial governments.
Prices of the rest (10 drugs) were set by the market.
Expenditures on drugs with prices set by the government
accounted for 19.5% of expenditure on the top 15 drugs in the
provincial hospital. In the municipal hospital, five of the top 15
drugs’ prices were set by the central government, three by both
central and provincial governments, and the remaining seven
by the market. The drug prices set by the market increased by
0.96% in the provincial hospital and by 35.8% in the
municipal hospital, in contrast to the decreases in prices set
by both central and provincial governments. In the municipal
hospital, expenditures on drugs with prices set by the
government accounted for 46.5% of the top 15 drugs’
expenditures.
Table 5. Price and utilization effects on drug expenditures per inpatient
day, cerebral infarction cases
Price and utilization effects on drug expenditures
Of the total 12.3% decrease in drug expenditures for cerebral
infarction cases per inpatient day in the provincial hospital,
7.45% was due to a decrease in drug utilization, price changes
accounted for 5.26%, and other factors, including introduction
of new drugs, were responsible for 0.39%. In the municipal
hospital, of the total 36% increase in drug expenditures per
inpatient day, 25.04% was attributable to the increase in drug
utilization, 8.8% was due to price increases, and 2.2% to other
Total change of drug
expenditures
Change due to utilization
Change due to price
Change due to both
price and utilization
Provincial hospital
Municipal hospital
212.3%
36.0%
27.45%
25.26%
0.39%
25.04%
8.80%
2.20%
192
Table 6. Top 15 drugs by expenditure in two Shandong hospitals: price setting method, price per PDD and strength of evidence base, after the pricing policy reform
Provincial hospital
Drug name
Price per
PDD
Expenditure
(yuan)
Evidence
base
246.5
149.3
193.7
70.7
55.5
12.1
160.2
400.9
75 195
21 505
20 655
20 334
14 258
13 207
13 170
12 135
Weak
Weak
Weak
NE
Some
Good
Weak
Good
98.1
198.8
60.4
256.0
88.6
74.9
117.2
11 021
9 143
8 727
8 449
7 944
6 525
6 446
248 714
Weak
Weak
NE
NE
Weak
NE
Good
65%
Drug name
Cerebroprotein hydrolysate
injection (M)
Panax notoginseng (CP)
Xiangdan injection (C)
Safflower injection (M)
Aceglutamide (M)
Cefotaxime sodium (C)
Pantoprazole sodium injection (M)
Mannitol (C)
Isotonic sodium chloride (C)
Glucose saline (C)
Troxerutin (M)
Clindamycin phosphate (M)
Xueshuantong (CP)
Defibrase for injection (M)
Kakonein injection (CP)
Expenditures for top 15 drugs
% for Western drugs with weak
evidence base
(M): Prices set by the market; (C): prices set by the central government; (CP): prices set by both central and provincial governments.
NE: Chinese medicine, not evaluated.
Price per
PDD
Expenditure
(yuan)
Evidence
base
61.4
60.3
27.2
62.0
67.4
33.6
93.1
6.9
14 979
8 094
6 220
5 146
4 922
4 016
3 630
3 595
Weak
NE
NE
NE
Weak
Good
Good
Good
5.9
7.2
33.8
49.2
52.3
65.1
52.5
3 167
3 125
2 903
1 870
1 760
1 732
544
65 703
Good
Good
Weak
Good
NE
Some
Weak
41%
Qingyue Meng et al.
Deproteinized calf blood extractives
injection (M)
Buflomedilhydrochloride injection (M)
Cerebroprotein hydrolysate injection (M)
Kudiezi injection (M)
Nicholin (citicholine) (M)
Glucose saline (C)
Fraxiparine (nadroparine calcium) (M)
Rocephin (CP)
Ginkgo leaf extract and dipyridamole
injection (M)
Cytidine 50 -triphosphate Na2 (M)
Panax notoginseng [Bnrk] F.H.Chen (CP)
Shexiang injection (M)
Piracetam (CP)
Breviscapine (M)
Azithromycin (CP)
Expenditures for top 15 drugs
% for Western drugs with weak
evidence base
Municipal hospital
China’s drug price controls and hospital drug expenditures
In the post-reform period, the municipal hospital incorporated
within the top 15 drugs six drugs not previously used for
cerebral infarction patients; expenditure for these drugs
accounted for 31.6% of the top 15 drugs’ expenditures.
These drugs had a mean price per PDD of 24.1 yuan,
compared with the overall mean price per PDD of 9.8 yuan.
The provincial hospital incorporated five drugs not previously
used within its top 15 drugs, with a mean price per PDD of
93.4 yuan compared with the overall mean of 31.6 yuan. The
expenditure on these newly-used drugs accounted for 45.6%
of total expenditures for the top 15 drugs. These findings
suggest a shift to newly incorporated (although not necessarily
new) drugs with higher prices. Of these drugs, only one in the
provincial and three in municipal hospitals were subject to
price regulation. This shows the difficulties of regulating
prices when new products are launched.
Regarding rationality of use post-reform, a high proportion of
expenditure went on Western drugs with a weak evidence base
in the published literature. Of the top 15 drugs in use in the
provincial hospital, the majority of expenditure—158 633
yuan or 65%—was for poorly evidence-based Western drugs.
Only 13% (31 788 yuan) of expenditures for the top 15 drugs
was for Western drugs considered to be safe and effective, and
this was primarily for antibiotics to treat co-existing infections
and intravenous fluids. Drugs considered to be highly
effective, such as acetylsalicylic acid, were not prominent in
either PDD or expenditures. In the municipal hospital, not
only were expenditures markedly lower, but a smaller
proportion of the Western drugs employed had a weak
evidence base (41%). However, these still accounted for more
than one-third of expenditures.
An average of 147 yuan per PDD of the top 15 drugs was
prescribed in the provincial hospital compared with only 48
yuan in the municipal hospital. The prices of the top 15 drugs
used in the provincial hospital were a mean of three times
higher than prices of the top 15 drugs used in the municipal
hospital. The markedly higher drug expenditures in the
provincial hospital, even when corrected for length of stay,
mainly resulted from use of more expensive drugs for cerebral
infarction cases.
Discussion
The study found evidence of de-facto implementation of the
retail price control policy in the two study hospitals. The
findings show that formal actions were taken by different
regulatory agencies and that the regulatees were aware of the
new system. In contrast, the desired effect of reduction or
stabilization of hospital drug expenditures was not identified
in either hospital in the analysis of data for all patients.
Increasing expenditures were due to a combination of factors.
These included: increased production of services, increased
utilization of drugs, and increased prices for some drugs. For
the tracer condition of cerebral infarction, there may have
been some stabilization of expenditures in the provincial
hospital, but expenditures increased markedly in the municipal
hospital. Utilization was a more potent determinant of
expenditures than price.
193
Because the overall growth of drug expenditures was as rapid
as before in the provincial hospital, we cannot conclude a
positive impact of the new drug pricing policy on drug
expenditure containment, even though drug expenditure per
inpatient seemed to decline (no statistical significance) for
cerebral infarction. The different trends in drug expenditures
for all patients and those of the tracer condition may result
from several factors, including the small size of tracer cases
lacking representation of overall drug expenditures and the
tracer condition being a unique case with a real decline in drug
expenditures. Further study through expanding the sample size
or the scope of tracers is needed in order to provide definitive
explanations.
A number of factors appear to have contributed to increasing
hospital drug expenditures. These include: incorporation of
drugs not previously employed, with higher average prices;
irrational use of drugs; inter-hospital differences such as use of
more expensive drugs, and possibly the social or insurance
characteristics of a hospital’s patient population. A very high
proportion of expenditures on cerebral infarction cases were
for Western drugs without an adequate evidence base of safety
and efficacy.
While there is evidence that Chinese public hospitals have
similar expenditure, income and incentive structures throughout the country (Meng et al. 1998), caution should be used in
generalizing the findings from this small sample of two
Shandong hospitals to China’s 15 700 other public hospitals at
and above county level. The different regulatory, economic
and administrative contexts may influence the implementation
and results of the drug pricing policy. In addition, the modest
number of cerebral infarction cases due to the short time
period studied may constrain the validity of the analysis. We
did not adjust drug inflation when comparing the drug
expenditure changes between 2000 and 2002, given that the
drug price index was claimed to be negative by the official
statistics. However, the price index may not truly reflect the
changes in drug prices if only drugs managed by the
government were used for calculation. Drugs with prices
that were set by the market and which increased during the
study period may not be adequately included in calculations
for the drug price index.
Differences in severity of patients between the two hospitals
are unlikely to explain the major inter-institutional differences
in drug utilization and expenditures. Xu et al. (2001) found
similarly for acute appendicitis; expenditure per case was
1163 yuan for provincial hospitals and 690 yuan for municipal
hospitals in 1999, mainly resulting from differences in drug
expenditures. It is well recognized in China that expenditures
in tertiary hospitals are much higher than those in community
hospitals for the same health conditions (Liu and Cheng 2002).
Explanations put forward have included costlier inputs and
higher quality provided in tertiary hospitals as well as the
relatively higher ability to pay of users of tertiary hospitals
(Centre for Health Statistics and Information 1998; Xu et al.
2001). For the case of cerebral infarction in this study, it
appears that the provincial hospital used more expensive
drugs, more of which lacked evidence of safety and efficacy,
than the municipal hospital. Besides the abovementioned
194
Qingyue Meng et al.
explanations from other studies, the drug prescribers in the
provincial hospital may have higher expectations of income or
other advantages from selling the drugs than prescribers in the
municipal hospital, even though drug prescribers in both
hospitals faced the same financial incentives.
The study evaluated the impact of the policy only as regards
hospital expenditures and did not evaluate whether the policy
had an effect on improving access to essential drugs. The
number of PDDs per outpatient visit was not studied. Lower
prices due to the policy could, for example, have improved
access to medicines, without having the desired effect on
hospital drug expenditures.
The rationality analysis, in spite of its striking findings, is
likely to present a very conservative underestimate of the
impact of irrational use on expenditures. This is because
rationality of use was examined looking only at the drugs
selected, and not the indications for individual patients. Had
this latter approach been used, it is likely that an even greater
proportion of expenditures would be considered unjustified,
such as those due to overuse of antibiotics, although other,
effective drugs, might also have been identified as underused.
Furthermore, drugs with some, but not yet solid, evidence
were not counted as irrational, and Chinese drugs were not
evaluated. Additionally, the analysis was only performed for
the post-reform period, and while it clearly demonstrates the
important contribution of irrational use to drug expenditures,
this situation may have preceded the policy.
Why then, was the rapid increase in drug expenditures not
constrained, given implementation of the policy? This may be
explained from four dimensions. First, hospitals could shift to
more expensive drugs not covered in the price control list. The
data provide some preliminary support for this possibility. It
has been reported that in some hospitals, drugs for which
prices had been reduced were no longer available; instead,
more expensive drugs were prescribed (Lu 2002). Secondly,
public hospitals would attempt to maintain drug income levels
in order to maintain their overall revenue level. While the
government reduced prices of listed drugs, alternative sources
of financing for hospitals were not created. To keep drug
revenues stable, hospitals could increase drug utilization or
irrational drug use. They may believe this to be necessary,
given the widely held view that official prices of professional
services set by the government are not high enough to cover
costs of hospital services (Meng et al. 2002). Thirdly,
corruption in drug purchasing and prescribing within hospitals
could contribute to the rapid rise in drug expenditures. Many
pharmaceutical companies send medical representatives to
lobby doctors and drug purchasing managers to use their
drugs, offering financial incentives. The returns to the drug
prescribers and managers are usually based on the quantities
of the drug sold. This would stimulate hospital staff to use
unnecessary and expensive drugs without strict regulations on
drug prescriptions.
Finally, there is a lack of coordination of action by different
government agencies on controlling drug expenditures—three
different government departments are responsible for work
related to drug expenditure containment. This situation can be
exemplified by the lack of standard treatment protocols for
common diseases such as stroke, or by the lack of costeffectiveness evaluation when approving new drugs for
marketing. While regulators formally implemented the price
regulation, important complementary actions were missing,
such as setting criteria for drugs to be included on the price
list, evaluation and price-setting of new drugs, their indications for use, and so forth.
Effective price competition between regulated and unregulated drugs, a desired output of the policy, was not identified.
This may be due to the higher returns from use of more
expensive drugs in a setting of fixed markups, driving
pharmaceutical manufacturers and wholesalers to increase
prices that are not regulated by the government. The
assumption of the drug pricing policy design is that drug
consumers would prefer low-price drugs. However, this does
not take into account that drug prescriptions are written by
doctors, not by the patients who will pay. Price competition
would be favoured if drug consumers or third party payers had
access to comprehensive information about the quality, cost
and efficacy of the drugs, as discussed below, and even then it
may be difficult. Furthermore, the study illustrates a problem
with price regulation found in other studies. Expenditure is a
function of price and volume, but the regulator has control of
only one component, the price. Control of utilization implies
control of prescribing behaviour of doctors, restriction of
consumers’ demand and other factors. Hence, the classical
way of circumventing price regulation is to increase the
volume component. In addition, the pharmaceutical market is
characterized by introduction of new products with an initial
high price. Overall, this study confirms the difficulties of using
price regulation as an isolated strategy to curb pharmaceutical
expenditures.
Given a price-elasticity on the demand side, the use of price
controls is unlikely to lower demand. People who pay out-of
pocket are currently the dominant financiers of the Chinese
health system and they are not well organized enough to exert
sufficient pressure to contain the gaming responses to price
controls, such as shifts to more expensive medicines. Recent
studies have shown that unification of individual health
insurance schemes in a city can help to control hospital
charging behaviour and constrain hospital costs (Zhang et al.
2000; Meng et al. 2004). In the city studied, the insurance
operators still work in isolation from each other, resulting in
the absence of a strong, single third party payer.
Irrational use of pharmaceuticals has been an increasing
concern in China. Yang and colleagues reported that more
than 98% of outpatients with common colds were given
antibiotics by physicians (Yang et al. 1993). Zhan et al. (1998)
also reported on the unnecessary prescribing of drugs in rural
health facilities, resulting in safety problems and rising drug
costs. Chen et al. (2003) analyzed rationality of use of the top
10 antibiotics for 571 inpatients using a drug utilization index,
and found that four of the top 10 drugs were used irrationally.
The underlying motivation behind the extensive irrational use
of drugs may be financial incentives that stimulate hospitals to
purchase, and physicians to prescribe, unnecessary and more
China’s drug price controls and hospital drug expenditures
expensive drugs. An aggravating factor may be the lack of
widely accepted and used standard treatment protocols for the
main pathologies, associated with a weakly evidence-based
culture of medicine in China. This is reflected in the Chinese
scientific literature, where it has been said that many articles
on drug safety and efficacy were not rigorously designed
and that more than 50% of Chinese medicines used lack an
evidence base (Guo 2001). The findings of this study provide
further evidence of the problem, and importantly, they
link irrational use to its enormous opportunity costs in
taking scarce health resources away from more useful
applications.
Summarizing, the price controls, while effective in lowering
prices for regulated drugs, were implemented without
fundamental changes in the incentive structures. In such a
setting, the tendency may be to move away from, rather than
towards, evidence-based medicine and cost-containment
goals. This study provides an analysis of the impact of the
policy in two specific contexts in Shandong province. Further
research with larger and more representative samples would
be useful.
Conclusion
The introduction of the new retail price control policy, while
implemented, was not found to have a positive impact on
containment of hospital drug expenditures in the two Chinese
hospitals studied. The effects of the policy on access and on
patient outcomes were not analyzed. Utilization more than
price determined drug expenditures in this model. For cerebral
infarction cases, more rational use of pharmaceuticals would
be crucial to constrain the rapid growth of drug expenditures.
Price control regulation should be supplemented with other
measures, including improvement of rational use of drugs
and measures addressing the present incentive structure
for hospitals.
References
Adams HP Jr, Adams RJ, Brott T et al. 2003. Guidelines for the early
management of patients with ischemic stroke: a scientific statement
from the Stroke Council of the American Stroke Association. Stroke
34: 1056– 83.
Adibhatla RM, Hatcher JF. 2002. Citicoline mechanisms and clinical
efficacy in cerebral ischemia. Journal of Neuroscience Research 70:
133–9.
Bell WR. 1997. Defibronegating enzymes. Drugs 54 (Suppl 3): 18 –30.
discussion 30-1.
Berge E, Sandercock P. 2002. Anticoagulants versus antiplatelet agents
for acute ischaemic stroke. Cochrane Database of Systematic
Reviews 4: CD003242.
Bloor K, Freemantle N. 1996. Lessons from international experience in
controlling pharmaceutical expenditure II: influencing doctors.
British Medical Journal 312: 1525– 7.
Bossi E, Bossi L. 1989. Acute and chronic cerebrovascular diseases.
A therapeutic approach. Minerva Medica 80: 991 –8.
Capella D. 1993. Descriptive tolls and analysis. In: Dukes MNG (ed).
Drug utilization studies: methods and uses. WHO Regional
Publications, European Series No.45. Copenhagen: WHO Regional
Office for Europe, pp. 64 –5.
195
Centre for Health Statistics and Information. 1998. National Health
Service Survey Report. Beijing: Ministry of Health.
Chen HY, Han LP, Yao XL, Ji B. 2003. Analysis of the frequency and
rationality of use of antibiotics in 571 patients. China Pharmacy 14:
159– 61.
China Health Economics Institute. 1999. Distribution of health
expenditures in China. Beijing.
China State Commission of Planning and Development. 1998. Method of
drug price management. Policy Document No. 2196. Beijing.
China State Commission of Planning and Development. 2000. Reforms of
drug pricing policy. Document No. 961. Beijing.
China State Commission of Planning and Development. 2002. Analysis of
decreasing index of drug prices. Price report. Beijing.
China State Drug Administration. 2003. National actions in reducing
drug prices. Publicity Report. Beijing.
China State Economic and Trade Commission. 1999. Statistics of
pharmaceutical industry. Beijing.
Chinese Pharmacy Yearbook Editing Commission. 1998. Chinese
Pharmacy Yearbook. Beijing: People’s Health Publishing House.
Cleverley W. 1992. Essentials of health care finance (third edition).
Caitherburg, MD: Aspen Publishers Inc., pp. 323–7.
Coull BM, Williams LS, Goldstein LB et al. 2002. Anticoagulants and
antiplatelet agents in acute ischemic stroke: report of the Joint Stroke
Guideline Development Committee of the American Academy of
Neurology and the American Stroke Association (a division of the
American Heart Association). Stroke 33: 1934– 42.
Dong HJ, Bogg L, Rehnberg C, Diwan V. 1999. Drug policy in China:
pharmaceutical distribution in rural areas. Social Science and
Medicine 48: 777– 86.
Du DQ. 2002. Difficulties for costing pharmaceuticals. Price Theory and
Practice 3: 10 –11.
Ess SM, Schneiweiss S, Szucs T. 2003. European healthcare policies for
controlling drug expenditures. Pharmacoeconomics 21: 89– 103.
European Stroke Initiative. 2003. Recommendations. Ischaemic stroke:
prophylaxis and treatment. Accessed 1 December 2003 at: [http://
www.strokecenter.org/prof/guidelines.htm#Acute].
Greener J, Enderby P, Whurr R. 2001. Pharmacological treatment for
aphasia following stroke. Cochrane Database of Systematic Reviews
4: CD000424.
Gross DJ, Kile J, Ratner J, Thomas C. 1996. Prices for prescription drugs:
the role of market forces and government regulation. In: Davis P
(ed). Contested ground: public purpose and private interest in the
regulation of prescription drugs. New York: Oxford University
Press.
Guo X. 2001. Use of evidence-based medicine in appraising drugs.
Chinese Journal of New Drugs 10: 641– 3.
Hu SL, Chen W, Cheng XM et al. 2001. Pharmaceutical cost-containment
policy: experiences in Shanghai, China. Health Policy and Planning
16 (Suppl 2): 4 –9.
Hu YP, Li LP. 2001. Analysis of drug cost escalation. Journal of Theory
and Reform 6: 69– 70.
Karoutas G, Milonas I, Artemis N et al. 1987. The pharmacological
effects of buflomedil in patients with multi-cerebral infarcts
dementia: an open, preliminary assessment. Current Medical
Research and Opinion 10: 380– 9.
Kobayashi S, Yamaguchi S, Okada K et al. 1988. Effect of long-term
administration of buflomedil on regional cerebral blood flow in
chronic cerebral infarction. Angiology 39: 934–41.
Liu H, Cheng X. 2002. Costing of community health services. Chinese
General Medicine 5: 626– 8.
Lu SQ. 2002. Why the consumers cannot enjoy the benefit after the
implementation of drug price control? China Consumers 7: 44 –6.
Maynard A, Bloor K. 2003. Dilemmas in regulation of the market for
pharmaceuticals. Health Affairs 22: 31– 41.
Meng Q, Ge R, Bian Y, Ma Y, Liu X. 1998. Analysis of fixed and
variable cost in hospitals. Chinese Health Service Management 7:
383– 6.
Meng Q, Bian Y, Ge R et al. 2002. Adjustment of health care prices for
correcting the incentive structure for hospitals. Chinese Journal of
Health Economics 5: 34– 6.
196
Qingyue Meng et al.
Meng Q, Rehnberg C, Zhuang N et al. 2004. Impact of urban health
insurance reform on hospital charges: a case study from two cities in
China. Health Policy 68: 197 – 209.
Ministry of Health. 2000. Regulation of drug pricing policy for health
providers. Policy Document No. 391. Beijing.
Ricci S, Celani MG, Cantisani AT, Righetti E. 2002. Piracetam for acute
ischaemic stroke. Cochrane Database of Systematic Reviews 4:
CD000419.
Wang Q, Wei JF. 2003. The causes of and solutions to unreasonable drug
prices. Journal of Lanzhou Commercial College 19: 96 –8.
Wei JW. 1999. Drug expenditures in China. Chinese Journal of Health
Economics 22: 1 –3.
Wu R. 1999. Cost-containment policy in China. Chinese Journal of
Health Resources 2: 28– 30.
Wu R, Yang H. 1999. Methods of containing drug expenditures in
China and other countries. Chinese Primary Health Care 13:
46 –7.
Xu L, Forbes I, Qu J et al. 2001. Assessing hospital financial reform.
China-Australia Joint Research Report. Jinan City and Sydney:
Shandong University and University of New South Wales.
Yang YH, Fu SG, Peng H et al. 1993. Abuse of antibiotics in China
and its potential interference in determining the etiology of
pediatric bacterial diseases. Pediatric Infectious Disease Journal
12: 986– 8.
Yip W, Eggleston K. 2001. Provider payment reform in China: the case of
hospital reimbursement in Hainan Province. Health Economics 10:
325– 39.
Zerda A, Velasquez G, Tobar F, Vargas JE (eds). 2002. Health insurance
systems and access to medicines. Geneva: World Health Organization.
Zhan SK, Tang SL, Guo YD, Bloom G. 1998. Drug prescribing in rural
health facilities in China: implications for service quality and cost.
Tropical Doctor 28: 42 –8.
Zhang Y, Li J, Zhang X. 2000. Application of settlement in line with
different diseases to the health insurance schemes. Chinese Journal
of Health Economics 19: 41– 2.
Zhao YX, Wan Q, Gao G, Du L. 2003. Health expenditures in China
2001. Chinese Journal of Health Economics 22: 1 – 3.
Acknowledgements
The research team is grateful to the Alliance for Health Policy and
Systems Research for the financial support. They thank the staff of the
two hospitals studied for their contribution in data collection. The
Swedish Institute is acknowledged for supporting the first author to
prepare the manuscript at Karolinska Institute and the Swedish
Foundation for Cooperation in Research and Higher Education for
supporting the third author’s work. Thanks to Associate Professor Dr Nils
Wahlgren of the Department of Neurology of the Karolinska Institute,
and secretary of the European Stroke Council, for reviewing the
classification on rationality.
Biographies
Qingyue Meng is Professor of Health Economics and directs the Center
for Health Management and Policy of Shandong University. He is a
member of the Advisory Committee of Management and Policy to
China’s Ministry of Health.
Gang Cheng is a lecturer in the Center for Health Management and Policy
of Shandong University, China.
Lynn Silver is Professor of Pharmaceutical Sciences, School of Health
Sciences, University of Brası́lia, Brazil.
Xiaojie Sun is currently a master’s student in the Center for Health
Management and Policy of Shandong University, China.
Clas Rehnberg is Associate Professor of Health Economics, Medical
Management Center, Karolinska Institute, Sweden.
Goran Tomson is Professor of International Health, Division of
International Health and Medical Management Center, Karolinska
Institute, Sweden.
Correspondence: Professor Qingyue Meng, Center for Health Management and Policy, Shandong University, Wenhua Xi Rd 44, Jinan,
Shandong 250012, China. Email: [email protected].
Annex 1. Drugs with a weak evidence base
The drugs considered as having a weak evidence base for use
in acute cerebral infarction were cerebroprotein hydrolysate (0
articles in Medline, 2 small randomized controlled trials
(RCTs) identified in CNKI); deproteinized calf blood
extract—also searched as Actovegin and Solcoseryl (1 article
in Russian with no abstract on the mechanism of Actovegin
after ‘cerebral ischaemia’, 1 study in rats); buflomedilhydrochloride injection—also searched as Loftyl (1 clinical trial
for acute cerebral ‘ischaemia’ with no placebo group, against
nicergolin, and 2 studies of cerebral blood flow in a small
number of patients after cerebral infarction) (Karoutas et al.
1987); Kobayashi et al. 1988; Bossi and Bossi 1989; gingkodypiridamole combination (0 articles); cytidine 50 triphosphate
(although this is a precursor of citicholine, no specific
literature was located for this substance); piracetam—this
drug has a number of published trials, however metanalysis by
the Cochrane collaboration considered the evidence of
effectiveness insufficient and raised concern of possible
increases in mortality, a second review considered that there
was weak evidence for effectiveness for aphasia post-cerebral
infarction (Greener et al. 2001; Ricci et al. 2002); aceglutamide (0 articles); troxerutin (0 articles); kakonein (0 articles
on medline, and 3 small studies on CNKI); and Fraxiparine
(this is nadroparine, a low molecular weight heparin
(LMWH))—use of LMWH in acute cerebral infarction has
been studied and while some trials were favourable,
metanalysis of the literature by the Cochrane collaboration
concluded that use of low molecular weight heparins increases
the risk of death and haemorrhage in relation to aspirin,
without concomitant benefits (Berge and Sandercock 2002).