Background Polypharmacy represents a growing public health concern for its associations with drug-related adverse events, poor adherence, medication errors and increased healthcare costs.
Aim To assess the occurrence of polypharmacy prescribing and to identify the correlates in terms of doctors and patients characteristics in rural Western China.
Methods 20 125 prescriptions were collected from 680 primary health clinics in villages from 40 counties in 10 provinces of Western China. Two outcome variables are used in the analysis: number of drug items prescribed and use of five or more drugs (polypharmacy). A multilevel Poisson regression model and a multilevel logistic regression model were employed to examine the socio-economic, demographic determinants of the above two outcome variables, respectively.
Results The average drug number per prescription was 2.36, and the percentage of polypharmacy was 5.8%. The odds for polypharmacy were higher for village doctors with a heavier workload than those with less workload (OR=1.70; 95% CI 1.26 to 2.29). Patients with injections prescribed had higher odds for polypharmacy than patients without injections (OR=4.61; 95% CI 3.93 to 5.42). Village doctor's workload, subsidy for doctors from government and injection in prescriptions, patient's age and gender were also significant predictors of number of drug items prescribed in the studied areas.
Conclusions Polypharmacy seems to be associated with the village doctors' and patients' background and government policy. Intervention studies are needed to promote rational use of drugs in rural China.
- drug prescription
- public health
- rural China
- drug safety
- primary healthcare
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
- drug prescription
- public health
- rural China
- drug safety
- primary healthcare
Polypharmacy, the concurrent use of multiple drugs, is of major concern worldwide for its associations with drug-related adverse events, poor adherence, medication errors, underprescribing and increased healthcare cost.1 2 Polypharmacy is related to many personal, social and medical factors and indicates suboptimal prescribing—more drugs may be being prescribed than is justified on clinical evidence, and the risk of adverse drug interactions rises with the number of drugs taken by patients at one time.3–5 Generally, polypharmacy was defined practically as five or more drugs used at a time. Polypharmacy is more common among older people because of the higher prevalence of chronic diseases such as cardiovascular diseases in them,6 and many studies of polypharmacy in older people have been published.
The rural three-quarters of the Chinese population is served by a three-tier system of health services and referral with village clinics and village doctors as the first tier, township health centres as the second tier, and county-level hospitals as the third tier. In the first tier of the system, village doctors engage in both healthcare and farming, and often earn as much from farming as from medical practice. Many of them received rudimentary medical training and work independently in the villages of their birth. Village doctors diagnose and treat patients, prescribe medicines and refer patients to the higher levels of service as warranted. Hardly any studies have been conducted to investigate the levels and determinants of polypharmacy in rural areas of Western China, where the economy is underdeveloped, and the healthcare services are poor. The aim of this study was to assess the frequency and pattern of polypharmacy across 10 provinces of rural Western China and to identify characteristics of doctors and patients that may influence prescribing in village health clinics.
Materials and methods
Study population and design
Paper prescriptions were collected from village health clinics in 40 counties from 10 provinces of Western China. There were 748 townships and 10 170 villages in the 40 counties. The primary outcome in the original 40 county study was the prescription of antibiotics for a patient. Assuming that the population proportion of primary outcome=0.5, intraclass correlation coefficient=0.25, and type I error=0.05, a sample size of 21 790 prescriptions was required to obtain a margin of error no more than 0.02. Data-collection sites were selected using a three-stage probability sampling method. Five townships were selected from each county by the following steps. Using Dejiang county in Guizhou province as an example, we demonstrate below how townships were selected (see table 1). First, all the townships in the county were ordered from smallest to largest according to the population size. Second, the accumulated population size was calculated. Third, the sampling interval was calculated by dividing the total population of the county by five. In table 1 the total population size is 451 607, and the sample interval is 451 607/5=90 321. In the fourth step, a random number was determined by reading a serial number from a banknote of Renminbi (Chinese currency). The random number consisted of the last digits of this serial number but had the same length as that of the sampling interval. For example, if the serial number from Renminbi banknote is 85 523 225, and the random number is then 23 225, since the sampling interval had five digits. Then, the first selection number was 23 225. Finally, we looked up the township with the accumulated population size which contained the first selection number. That was the first township from this county to be included in this study. The second selection number was the first selection number plus the sampling interval, and thus the second township was selected. In the same way, the third, fourth and fifth township were selected. Similarly, four villages were selected from each selected township. The sampling interval for village selection was calculated by dividing the total population size of a township by four. Finally, 200 townships and 680 villages were selected from these 40 counties.7 At each selected village health clinic, the last 30 paper prescriptions up to 31 May 2005 were collected.
Village doctors were interviewed using a structured questionnaire to determine the doctor's age, gender, level of education, length of experience working as a village doctor, workload, whether they received subsidy from the government, whether they had a licence to practise and the amount of income they derived from their medical practice.
Data were collected by investigators from Xi'an Jiaotong University College of Medicine from June 2005 to August 2005. In total, 20 125 prescriptions were collected; these were for outpatients (village health clinics in rural Western China do not have inpatient facilities). Signed informed consent was sought from the village doctors after the study was explained to them by the interviewer. The protocol was reviewed and approved by the Human Research Ethics Committee of the Xi'an Jiaotong University College of Medicine.
Definition of polypharmacy
In this study, prescriptions with five or more drugs were defined as polypharmacy. From a clinical viewpoint, the occurrence of adverse drug event increases with number of drugs taken, and increases steeply when the drug number approaches five or more. This practical definition of polypharmay is similar to most polypharmacy studies.8–11
Main study variables
The number of drugs prescribed per prescription and the occurrence of polypharmacy are two main outcome variables chosen in this study. The average number of drugs per prescriptions and the percentage of polypharmacy were calculated as their summary statistics. The average number of drugs per prescription indicates the extent of polypharmacy. The percentage of polypharmacy prescriptions indicates the proportion of encounters that obtain prescriptions with five or more drugs from village clinics. Chinese traditional medicines were included, but purely herbal remedies were not included in this analysis. Generally, a proprietary Chinese medicine is always made from many kinds of herbs, which work together to cure disease. Thus, what really counts is the number of proprietary Chinese medicines but not the number of herbs. The 20 125 prescriptions used in this study do not include herbal medicines. Most of the village doctors prescribed proprietary Chinese medicines and chemical medicines. Purely herbal remedies were rarely prescribed by village doctors.
The Kruskal–Wallis H test was used to compare the number of drugs per prescription among the provinces. Poisson and logistic regression models were used to identify the independent predictors of multiple drug description and polypharmacy, respectively. To take into account the possible correlation among patients in the same villages, we used the multilevel Poisson and logistic regression models.12 The intraclass correlation coefficient (ICC) was calculated to assess the cluster effect within a same village. The analysis was conducted in two steps. First, we identified the potential predictors using univariate regression. Variables with p value <0.05 were selected for multivariate regression in the second step of the modelling. In the final models, variables with p value <0.05 were entered. For a quantitative predictor, transformations (such as quadratic, log arithmetic, square root) were tested to find its best characterisation.
Data analysed in this study were derived from 680 of 800 village health clinic questionnaires and 20 125 of 24 000 prescriptions. There were 120 villages were excluded for the following reasons: 38 villages without village clinics, 55 villages without doctors, 11 villages without prescriptions, 15 villages where village doctors were not found and a township which had only three villages. The socio-demographic characteristics of the study sample are shown in table 2. The average age of the village doctors was 40.7 years, and 19.4% of them were female; 64.7% of villages doctors had full-time medicine education, and 52.6% of them had a subsidy from the government. On average, a village doctor's workload was 62.7 patient-visits per month. According to the prescription record, 56.8% patients were male, and 43.2% were female, with a range of ages from 1 month to 98 years. About one-half of all prescriptions were for upper-respiratory-tract infection, diarrhoea and gastritis.
Number of drugs prescribed per description
On average, there were 2.36 drugs per prescription, including 0.81 Chinese traditional medicines and 1.55 chemical medicines (table 3). Prescriptions with one drug item represented 26.4%. Those with two drug items comprised 35.4%, and 22.3% were for prescriptions with three drug items. There were 10.1% prescriptions with four drug items, and those with five or more than five drug items were 5.8%.
The average number of drugs prescribed is quite different among these 10 provinces. The Kruskal–Wallis H test demonstrated that the overall difference in drug number was significant (χ2=2822.5, df=9, p<0.001) among the 10 provinces. Sichuan province had the highest average number of drugs prescribed, and Inner Mongolia had the lowest level.
The overall percentage of polypharmacy was 5.8% in this study, and the percentage of polypharmacy was significantly different among these 10 provinces, as shown in table 3 (χ2=620.1, df=9, p<0.001). There were 8.4% outpatients aged over 60 years who obtained polypharmacy prescriptions from village doctors, and 5.5% outpatients aged <60 years obtained polypharmacy prescriptions.
Predictors of the number of drug items prescribed
Table 4 presents the results from a multilevel Poisson model analysis of the number of drug items prescribed. Provincial differences in the mean count of drugs prescribed still exist even after controlling for other covariates, with Chongqing, Sichuan, Guangxi and Gansu having a higher number of prescribing more drug items per prescription than Ningxia. Doctors with subsidies from the government tended to prescribe fewer drug items than those without. A heavier workload for village doctors tended to prescribe more drug items. Older patients had a higher number of drug items than the younger patients, and female patients had more than males. Patients with a prescription of injection tended to have more drug items than those without. Patients who were diagnosed as having an upper-respiratory-tract infection, diarrhoea or gastritis had more drug items than others. The ICC of number of drug items was 2.4%, indicating very little cluster effect in terms of prescribing more drug items among village doctors.
Predictors of polypharmacy
The results from multilevel logistic model analysis of polypharmacy are presented in table 5. Chongqing and Sichuan had higher odds of polypharmacy than the other provinces. The odds of polypharmacy for village doctors with a higher working load were higher than those with less workload. Patients with an injection had higher odds of polypharmacy than those without injection, suggesting that polypharmacy was more likely to occur in prescriptions with an injection. In more than one-half of prescriptions with injection, two or more than two drug items were injected at a time. This could contribute to the large number of drugs in these prescriptions and result in polypharmacy. The ICC of polypharmacy was 15.18%, suggesting that the odds of polypharmacy for village doctors are higher than others.
The average number of drugs per prescription is one of the important indicators of multiple drug prescription recommended by WHO. This figure varies around the world: 3.16 in Nigeria, 2.15 in a medical outpatient department in western Nepal and 2.8 in western Norway among 70–74-year-old people.13–15 In the present study, there were 2.36 drugs per prescription on average in village clinics of rural Western China. It seems that the average number of drugs per prescription in rural Western China is slightly lower than in other countries.
It is found in this study that female patients had more drugs than males, a result consistent with the finding that females were associated with a higher risk of inappropriate prescription and polypharmacy.16 17 Physiological changes with ageing make older people more susceptible to drug-related problems, and older people are at a higher risk for suffering polypharmacy.18 Twelve per cent of the population aged 65 to 74 years in the USA used five or more drugs.19 In South Wales, 9% of persons from a community survey aged 56 to 75 years used five or more drugs.20 It seems that the proportion of polypharmacy in older people in rural Western China is not higher than that of the USA and South Wales.
This report shows that a village doctor's workload influences their decision in prescribing multiple drugs significantly. This finding coincides with Bjerrum et al's study in Denmark.21 Consultation time is a patient care indicator for rational use of drugs advised by WHO. Adequate consultation time is essential for a rational decision of drug prescription.22 Heavy workload is closely related to a lack of consultation time. Most village doctors in rural Western China were part time mainly because they cannot depend on incomes from their medical practices alone. Unlike doctors in hospitals who had a fixed income, village doctors are self-employed: they are half-peasant and half-doctor. In general, their incomes are composed of medical practices, agriculture production, stock raising and others. Part time job was a great disadvantage for village doctors to practise well. Subsistence from the government can be regarded as the income of medical practice, an effective way to reduce drug number as found in this study.
The route of injection in drug administration is found to be positively correlated with polypharmacy in village clinics in the present study. Although hardly any studies have investigated the causal relationship between drug administration route and polypharmacy, the fact remains that complex drug regimens given by injecting often involve multiple drugs. Both polypharmacy and overuse of injections are part of the irrational use of drugs, and they are inseparable. It is therefore imperative that some interventional measures be taken to control these irrational drug utilisations.
Upper-respiratory-tract infection, diarrhoea and gastritis were commonly diagnosed in the studied areas. Although these diseases were not statistically correlated with polypharmacy, more drugs were prescribed for them than for other diseases.
The advantage of this study is that it is by far the largest study on polypharmacy in Chinese rural areas, filling a gap on polypharmacy in this geographical part of China. Data were collected across 151 548 km2 in Western China, including 14 532 400 village dwellers. The authenticity of the findings was warranted by original paper prescriptions collected from village clinics. The limitation of this study is that it is clinically based, in which the self-medication and medications for inpatients have not been taken into consideration. This is an observational study, and it is based on prescriptions, not dispensed or consumed medications. As a result, the true level of polypharmacy may have been underestimated. The study also has the limitation that the drug-utilisation outcome could not be analysed in the present study.
In conclusion, this study has investigated the multiple drug prescription and its predictors in village clinics in rural areas of Western China. The results demonstrate that polypharmacy prescriptions account for 5.8% of all prescriptions. Polypharmacy prescribing may be associated with the village doctor's workload and route of injection for drug administration. These results will have important policy implications for the Chinese government to promote the rational use of drugs in rural China.
What is already known on this subject
Polypharmacy is associated with drug-related adverse events, poor adherence, medication errors and increased healthcare costs.
In all probability, polypharmacy indicates suboptimal prescribing—more drugs may be being prescribed than is justified on clinical evidence, and the risk of adverse drug interactions rises with the number of drugs taken by patients at one time.
Polypharmacy is more common among older people because of the higher prevalence of chronic diseases such as cardiovascular diseases.
What this study adds
Polypharmacy prescriptions account for 5.8% of all prescriptions in rural Western China.
Our study suggests that the proportion of polypharmacy in older people in rural Western China is no higher than that of the USA and South Wales.
Our findings suggest that village doctors with heavy workloads are associated with a higher risk of polypharmacy prescribing in rural areas of Western China.
The route of injection for drug administration is found to be positively correlated with polypharmacy in village clinics in the present study.
We would like to thank all general practitioners, who participated in this study, and all investigators, for data collection. We are also grateful to P Milligan, for his comments on the earlier draft of this manuscript.
Funding Chinese Ministry of Health (MOH) the United Nations Children's Fund (Unicef).
Competing interests None.
Ethics approval Ethics approval was provided by Human Research Ethics Committee of the Xi'an Jiaotong University College of Medicine.
Provenance and peer review Not commissioned; externally peer reviewed.