Prevalence, expenditures, and associated factors of purchasing non-prescribed Chinese herbal medicine in Taiwan

Background and aims Purchasing Chinese herbal medicine (CHM) without a physician’s prescription may have adverse effects on health. However, the recent status of purchasing non-prescribed CHM and the associated factors are not completely understood. We aimed to report the prevalence of purchasing CHM and associated factors. Methods Using data from the 2017 Taiwan Survey of Family Income and Expenditure, we identified 16,528 individuals (householders) aged 18 years and older. Sociodemographic factors, expenditures on medical services and health behaviours were compared between people with and without a history of purchasing non-prescribed CHM by calculating adjusted odds ratios (ORs) and 95% confidence intervals (CIs) in a multiple logistic regression analysis. Results The one-year prevalence of purchasing non-prescribed CHM was 74.8% in Taiwan. In addition to sociodemographics, marital status (OR 2.14, 95% CI 1.88–2.44), the use of traditional Chinese medicine (OR 3.62, 95% CI 3.30–3.97), the purchasing of non-prescribed biochemical medications (OR 3.09, 95% CI 2.75–3.48), the purchasing of health foods (OR 2.59, 95% CI 2.33–2.86), the use of folk therapy (OR 2.27, 95% CI 1.95–2.64), and a high level of expenditure on alcohol (OR 3.79, 95% CI 3.29–4.36) were strongly correlated with purchasing non-prescribed CHM. Conclusion The one-year prevalence of purchasing non-prescribed CHM is very high in Taiwan and is correlated with sociodemographics, health behaviours, and the utilization of medical care. The interaction of non-prescribed CHM with physician-prescribed herbal medicine and biomedical medications requires more attention.


Introduction
With the growing knowledge of and positive attitudes towards the self-management of health and medical care around the world, complementary alternative medicine (CAM) is increasing been widely used by people in Western and Asian countries [1][2][3]. It has been reported that the one-year prevalence of CAM use among the general population in Europe, the United States, and Japan was 25.9% in 2014, 33.2% in 2012, and 62.1% in 2009, respectively [1][2][3]. According to an estimation made based on a national survey, the total out-of-pocket expenditures for various types of complementary health approaches in the United States were as high as $30.2 billion [4].
Chinese herbal medicine (CHM) and acupuncture are the two most important treatment types in traditional Chinese medicine (TCM), which is a type of CAM. The high prevalence of the lifetime use of acupuncture (7.7%) and CHM (4.8%) in Australian women was investigated in the previous study [5]. In the United States, the prevalence of the lifetime use of acupuncture was 4.1% in 2002 and increased to 6.8% in 2007 among adults aged 18 years and over [6,7]. An internet-based survey showed that the one-year prevalence of TCM use was as high as 39.7% in 2018 in Taiwan [8]. TCM has beneficial effects among patients with various diseases, such as stroke, cancer, pain and allergic rhinitis [9][10][11][12]. Although the outpatient payments for TCM have been covered by Taiwan's National Health Insurance Program since 1995, it is still common for people to purchase CHM without a physician's prescription in traditional pharmacies or illegal markets [13,14].
Herb-drug interactions, unqualified herbal medicine, and purchasing non-physician-prescribed CHM may lead to adverse effects, such as serotonin syndrome, rhabdomyolysis, acute delirium [15][16][17][18]. Although the prevalence and characteristics of and the factors associated with purchasing non-prescribed CHM among adults in Taiwan have been reported, there were several limitations of these studies, such as the use of outdated data (surveys performed in 2005 or 2006) [13,14], a small sample size (641 participants) [14], a focus on short-term use (one month) [14], and the lack of CHM expenditure data [13,14]. Therefore, there is a need for recent information regarding the people who purchase non-prescribed CHM in Taiwan. Using data from the Taiwan Survey of Family Income and Expenditure (TSFIE) in 2017, we conducted a cross-sectional study to investigate the recent prevalence and characteristics of and factors associated with the purchase of non-prescribed CHM among adults in Taiwan.

Source of data and study design
We used the database from the 2017 TSFIE that was conducted by the government (the Directorate-General of the Budget, Accounting and Statistics, Executive Yuan, Taiwan). The sources of the data, sampling methods, and measurement details were described in a previous study [19]. In brief, the researchers used a stratified random sampling method to choose the sampled households from the whole population in Taiwan, and these households were interviewed annually to determine their major sources of income and expenditures. The householder was the representative of the family, and he or she was asked to answer questions regarding daily income and expenditure activity within the previous year. The population of Taiwan is approximately 23 million and is distributed throughout 7 cities and 18 counties. The universal sampling rate was approximately 0.20%, with 16,528 sampled households. The survey content is categorized into four domains: household members, household facilities and housing conditions, income and expenditures, and consumption expenditure. The 2017 TSFIE covers the one-year period from January 1 to December 31, and the interview was administered from January 1 to February 28 in 2018. At the beginning of each interview, the interviewer explained the survey purpose and asked if the interviewee was an income earner aged 20 years or over who lived in the household. Because the 2017 TSFIE was administered by face-to-face interviews with account-keeping collection methods, it was more accurate than studies conducted solely based on interviews.

PLOS ONE
Using the data from the 2017 TSFIE, we conducted this cross-sectional study to investigate the prevalence of purchasing CHM without a physician's prescription and the associated factors. Information from the 2017 TSFIE database used in this study included a series of personal and family characteristics, including sociodemographics (age, sex, marital status, education level, employment status, and geographic locations), health care utilization, individual and household out-of-pocket expenditures, etc.

Ethical approval
At the initial interview during the 2017 TSFIE, the participants were asked for permission for access to the interviewed database for research purposes, and all study participants signed the informed consent form. To protect personal privacy, the electronic database was encoded, and the identifiers of participants were scrambled prior to further academic access for research. The current study was also evaluated and approved by the Directorate-General of Budget, Accounting and Statistics, Executive Yuan, Taiwan

Definition and measures
In Taiwan, TCM is legal, and it is covered by the National Health Insurance Program for ambulatory care but not inpatient care. TCM practitioners are registered TCM doctors who have been certified by the Ministry of Health and Welfare, and they can only provide TCM in a hospital or clinic. Western medicine doctors cannot practise TCM in Taiwan. According to medical law in Taiwan, TCM doctors can advertise the medical benefits of TCM. TCM includes herbal medicine, acupuncture, moxibustion, bone reduction, traditional trauma treatment, traditional dislocation treatment, traditional fracture treatment, Tuina, Baguan, and other therapies. However, not all TCM modalities are covered by the National Health Insurance.
The core question in the 2017 TSFIE was "Have you ever purchased Chinese herbal medicine for yourself or your family members without a physician's prescription in the past year?" In this study, we defined purchasing non-prescribed CHM as purchasing CHM without a physician's prescription. The purchased non-prescribed CHM included Chinese medicine materials (such as Dang Guei, Ren Shen, Sih-Wu-Tang, Si Shen Tang, Shi Quan Da Bu Tang, Gou Ci Zih), tinctures, traditional formulas, and concentrated preparations. People can purchase nonprescribed CHM from biochemical pharmacies (such as over-the-counter CHM, Good Manufacturing Practice products), CHM pharmacy (such as Chinese medicine materials and concentrated preparations), and illegal markets.

Statistical analyses
We used chi-square tests to compare the differences in sociodemographic factors (such as age, sex, occupation, education, marital status, level of income, urbanization), lifestyle factors (including smoking and alcohol consumption), and medical care behaviours (receiving Western medical care, TCM and dental care) between people who had and had not purchased nonprescribed CHM.
The crude odds ratios (ORs) and 95% confidence intervals (CIs) of factors associated with purchasing non-prescribed CHM were calculated by univariate logistic regression. These factors included age, sex, occupation, education, marital status, level of income, urbanization, smoking expenditure, alcohol expenditure and the use of Western medicine, TCM and dental care. The significant factors (p<0.05) were then entered into multivariate logistic regression analysis to calculate the adjusted ORs and 95% CIs of the factors associated with purchasing non-prescribed CHM.
For each covariate, we assigned a predictive score as a risk index according to the significant adjusted OR, and thepredictive score was proportional to the OR. The purchasing predictive score was defined as follows: when 1.0 � OR<1.5, the purchasing predictive score was 1; when 1.5 � OR <2.0, the purchasing predictive score was 2; when 2.0 � OR <2.5, the purchasing predictive score was 3; and when 2.5 � OR <3.0, the purchasing predictive score was 4. All analyses were performed using Statistical Analysis Software (SAS), version 9.2 (SAS Institute Inc., Cary, North Carolina, USA). A two-sided p-value less than 0.05 was considered statistically significant.

Results
The one-year prevalence of purchasing non-prescribed CHM was 74.8% among 16,528 adults aged older than 18 years (Table 1). A higher percentage of middle-aged people (50-59 years) purchased non-prescribed CHM than people aged 18-29 years (p<0.0001), and more males purchased non-prescribed CHM than females (p<0.0001). Higher percentages of people who had an occupation in the agriculture/forestry/fishery industry, were currently married, and had high level of income purchased non-prescribed CHM than their counterparts.
As shown in Table 2, a higher prevalence of purchasing non-prescribed CHM was found among people who purchased non-prescribed biomedical drugs (p<0.0001), purchased health food (p<0.0001), spent a relatively large amount of money on alcohol, and used folk therapy (p<0.0001), dental care (p<0.0001), Western medicine (p<0.0001), and TCM (p<0.0001). Compared with the reference group in the multiple logistic regression (Table 3), the adjusted ORs for purchasing non-prescribed CHM in the groups who were 70-97 years of age, were female, had 0 years of education, and had agriculture/animal occupations were 1.51 (95% CI 1.17-1.94), 1.25 (95% CI 1.14-1.

Discussion
Using nationwide survey data from the 2017 TSFIE, we found a very high one-year prevalence (74.8%) of purchasing non-prescribed CHM among adults. Purchasing non-prescribed CHM was associated with sociodemographic factors, expenditures on unhealthy lifestyle habits, the use of medical care, the use of folk therapy, and the purchase of biochemical medications and health food. Sex is an important factor associated with medical care and CAM [14]. In general, women were found to have better knowledge, attitudes, and practices regarding self-management of health than men, and they were also more willing to address health problems by trying multiple therapies, including conventional medicine, CAM, biochemical medications, health foods, and herbal medicine [14]. In addition, social networking and being a housekeepers make it easier for women than men to access non-prescribed CHM for themselves or family members  [19]. Similar to previous study findings [14], our results showed that women were more likely to purchase non-prescribed CHM than men. Compared with young people, elderly individuals have more coexisting medical conditions, and they are relatively more willing to try multiple therapies to treat their illnesses [13]. A previous study also suggested that older people are more likely to purchase non-prescribed CHM to improve their well-being and disease-related symptoms than young people [14]. The survey regarding CAM use in Tokyo showed that older people had a higher prevalence of CAM use than people aged younger than 60 years [20]. Older people are the major purchasers of TCM products not covered by Taiwan's National Health Insurance [21]. Therefore, our finding of a higher likelihood of purchasing non-prescribed CHM among elderly individuals than among young people is reasonable.
In the present study, we found that participants who lived in highly urbanized areas had a higher likelihood of purchasing non-prescribed CHM. In general, residents in urban areas have more access to information on conventional and unconventional therapies than rural residents because there is more information on health management in cities than in rural areas [22,23]. Our study also found that people who work in agriculture, with animals or in bluecollar jobs had a relatively high likelihood of purchasing non-prescribed CHM because many of these people live in rural areas and have low levels of education and income. These findings were consistent with those of previous studies [14].
People living in urban areas have higher levels of income than in people living in rural areas, and income is one of the determinants of purchasing CAM [2]. We hypothesized that people with high income levels would have much more money to purchase conventional medicine, traditional medicine, folk therapy, health food, and non-prescribed medicine not covered by insurance. This viewpoint also helps explain the association of purchasing of biochemical medications, purchasing health food, using of folk therapy, and high expenditure levels on alcohol were factors associated with purchasing non-prescribed CHM in this study. Having a stable income is an important factor in a marriage. We also found that being married was associated with purchasing non-prescribed CHM in this study. Caring about each other, paying more attention to each other, and expressing concern prompt people with illnesses and their family members to purchase non-prescribed CHM. A previous study indicated that marital status was a factor influencing the use of dietary supplements in South Korea [21]. Under the coverage of Taiwan's National Health Insurance Program, people with illnesses or diseases can easily receive Western medical care and TCM. It is understandable that people with a disease or illness might try multiple therapies to improve their well-being and diseaserelated symptoms [14]. This may explain why the use of outpatient Western medical care and TCM were associated with the purchase of non-prescribed CHM. However, our study also showed that people who were not hospitalized had a greater likelihood of purchasing non-prescribed CHM than those who had experienced hospitalization. In general, patients who need inpatient care in the form of hospitalization have relatively more severe medical conditions than those who need only outpatient care. Hospitalized patients have less opportunity to purchase non-prescribed CHM, and most of these patients had better compliance with Western medicine inpatient care.
A Palestinian survey found that the prevalence of potential drug-herb interactions was 24.9% among 237 patients with chronic disease who had used at least one herb and that male sex, older age, a higher number of biochemical drugs, and more chronic disease were associated factors [15]. A large scale study based on data of Taiwan's National Health Insurance suggested that Ma Huang, Dang Gui, and Baizhi were the most common Chinese herbs interacting with biochemical medicines in 1998-2011 [24]. Interactions between herbal medicines and prescribed drugs should be considered when using herbal medicine [15,16,24,25].
Some limitations existed in this nationwide survey. First, due to the cross-sectional study design, we could not provide information regarding the long-term trend in purchasing nonprescribed CHM. The causal inferences between purchasing CHM and the associated factors are also limited in this study. Second, recall bias is always possible because our data were obtained with face-to-face interviews. Elderly people and those with brain-related illnesses may struggle to recall the information from the past year. Third, we have no further information regarding the dosage or frequency of use of the purchased non-prescribed CHM. Thus, we could not evaluate these details among people in Taiwan. In addition, previous studies suggested that medical conditions and underlying diseases are determinants of seeking TCM, purchasing herbal medicine, and using folk therapy [13,14,23,26]. One of the study limitations is that the information regarding the history of diseases and medical conditions of participants was not available in the database of the 2017 TSFIE. Finally, we tried to use scoring model to investigate whether predictive score is associated with the purchase of non-prescribed CHM. However, we have no other database for further validation and this is one of our study limitations. Our predictive model needs further improvement, validation, and evaluation by future well-design studies.
In conclusion, a very high one-year prevalence of purchasing non-prescribed CHM was observed among adults in Taiwan, and the purchasing of non-prescribed CHM was associated with sociodemographic factors and the use of medical care. The findings of our study remind health policy makers, clinical Western medicine physicians and TCM physicians that the purchasing of non-prescribed CHM is a serious problem. The interactions of non-prescribed CHM with physician-prescribed herbal medicine and biomedical medications needs more attention.
Supporting information S1 Table. Expenditure on non-prescribed Chinese herbal medicine by baseline characteristics (N = 12,374). (DOC)