There is an increasing trend in the utilization of complementary or alternative medicine (CAM) worldwide 1. According to the estimation of the World Health Organization, the usage of CAM ranges from 9 to 65% in different countries 2. In the United States, CAM utilization increased from 34% in 1990 to 42% in 1997 34. CAM is also very popular in Europe 5, Canada 6, and Australia 7. Patients choosing CAM are not necessarily dissatisfied with conventional medical care. They are seeking more effective ways to improve their health and well-being and to relieve symptoms associated with chronic, even terminal, illnesses or the side effects of conventional treatments for them 8.

Among different forms of CAM, Chinese medicine (CM) is well known for the medicinal formulas and acupuncture, and is one of the most popular alternative medicines in many countries. In Singapore, CM is the most popular form of CAM and has been employed by 88% of the CAM users 9. Moreover, CM accounts for 40% of the health care in China 10. Although CM is commonly used in the countries of East Asia, this system of medical service is also growing in popularity and offers an important alternative or complement to biomedical care in the Western countries 11.

In Taiwan, one distinguishing feature of the national health-care insurance system is the co-existence of the modern Western medicine (WM) and CM. In addition, the social health-care insurance system covers CM since 1975. Although only bone fractures and dislocations were included in the initial coverage of the Labor Medical Insurance, the therapeutic items were then expanded to internal medicine, gynecology, and acupuncture in 1983. In 1988, the Public Employee Medical Insurance Program started to reimburse CM 12. After the implementation of the National Health Insurance (NHI) Program in March 1995, all citizens who have established a registered domicile in Taiwan are mandated to join the program. In June 2003, more than 99% of the 23-million population has benefited from this universal health insurance program. However, CM has only 5% of the total expenditure in the NHI coverage. The health care items covered physician consultation for diagnosis, drug prescription, acupuncture, and muscle strain therapy 13.

Although information on different aspects of CM in Taiwan is available 121415161718, these reports were based on small samples 12141516 or limited to one component of CM 1718. Moreover, some of these studies were conducted before the establishment of the NHI program 121415. Recently, a study on the use frequencies, characteristics of users, and the disease categories treated by CM in Taiwan was conducted based on the complete datasets of CM outpatient reimbursement claims from 1996 to 2001 19. However, except sex and age, other user characteristics were not analyzed and the utilization pattern of WM was only slightly explored. The purpose of this study is to gain a more complete picture of utilization of CM in the national health insurance system with dual medical systems (Western and Chinese medical services) by determining the extent of CM utilization from 1997 to 2003 using WM as a reference and the demographic factors and primary indications that predict the choice of the type of medical service.

Table 1 shows the demographic characteristics of the studied cohort. Adjusted ORs and 95% confidence intervals (95% CIs) resulting from the logistic regression model are displayed in Table 2. The odds of using CM (OR 1.48) and WM (OR 1.74) were higher for females than for males (OR 1.00). Compared with the youngest age group (OR 1.00), the odds of CM increased with age to a peak in the 45–54-year-group (OR 1.75) whereas those of WM became relatively high only in the elderly subjects (≥ 65 years) (OR 1.09).

The odds of CM (OR 1.05) and WM (OR 1.00) in the group with a monthly wage US$640-1,279 were higher than those in the low-income group (< US$640) (CM: OR 1.04; WM: OR 0.98), farmers and fishermen (CM: OR: 0.90; WM: OR 0.97) and the other group (CM: OR 0.86; WM: OR: 0.77). Although the odds of CM were similar in the groups with (OR 1.00) and without (OR 1.00) severe diseases, the odds of WM in the people with severe diseases (OR 2.40) was higher than those without these diseases (OR 1.00) (Table 2).

The odds of CM were higher among people in Central (OR 1.65), Southern (OR 1.18), and Kaohsiung and Pingtung (OR 1.09) than those in Eastern Taiwan (OR 1.00), Northern Taiwan (OR 0.96), and Taipei (OR 0.91). However, those of WM were higher among people in Taipei (OR 1.03), Northern Taiwan (OR 1.08), Central Taiwan (OR 1.15), Southern Taiwan (OR 1.20), and Kaohsiung and Pingtung (OR 1.21) than Eastern Taiwan (OR 1.00). The odds of CM in the general population (OR 1.00) was higher than those in the mountainous regions (OR 0.57) and offshore islands (OR 0.78) whereas those of WM were higher in mountainous regions (OR 1.03) and offshore islands (OR 1.21) than the general population (OR 1.00) (Table 2).

Table 2 also shows a steady increasing trend in the odds of CM from 1997 (OR 1.00) to 2003 (OR 1.15). Moreover, the odds WM utilization also increased from 1997 (OR 1.00) to 2002 (OR 1.28) but decreased in 2003 (OR 1.17). Of all seasons, the odds of either CM or WM were the highest in winter.

The crude utilizations of CM and WM are shown in Table 3. The number of CM users increased from 36,372 in 1997 to 41,823 in 2003. However, the number of WM users from 115,833 in 1997 to 121,605 in 2002 and decreased to 120,926 in 2003. The average number of patients who had used any CM service in one year and the number of visits to CM providers were 39,562 and 191,612, respectively. The corresponding figures for WM were 119,915 and 1,542,342. Most of the patients had one ambulatory visit a year. However, the utilization frequency of WM (mean = 12.55–13.34, median = 8–9) was higher than that of CM (mean = 4.63–5.03, median = 2–3). Most of the ambulatory CM service was provided by clinics (93.4–96.1%) and only 3.9–6.6% was provided by hospitals. Moreover, the frequency of ambulatory visits in hospitals steadily increased from 3.9% in 1997 to 6.6% in 2002. On the other hand, clinics provided about two-thirds (60.5–67.8%) of the ambulatory WM service, outpatient department of the hospitals took the remaining one-third (32.2–39.5%) of ambulatory service.

Diseases of the respiratory system (CM 22.1%, WM 35.6%) and the musculoskeletal system and connective tissue (CM 18.1%, WM 7.5%) were the top two primary indications in the ambulatory health care of both CM and WM. In CM, the remaining common primary indications were injury and poisoning (16.2%), diseases of signs, symptoms and ill-defined conditions (14.2%), and diseases of the digestive system (11.4%). In WM, diseases of the genitourinary system (7.2%), diseases of the digestive system (7.2%), and diseases of sense organs (7.1%) were also common primary indications (Table 4).

Herbal medication was the most important component of CM. This component accounted for more than two-thirds of the ambulatory visits (68.4–72.7%). The utilization rate decreased from 70.8% in 1997 to 68.7% in 2003. The other two major components were muscle strain therapy (including dislocation therapy) and acupuncture. Muscle strain therapy accounted for 15.6–17.5% and acupuncture for 9.2–13.0% of the ambulatory visits. These two components showed increasing utilization trends: muscle strain therapy increased from 16.4% in 1997 to 17.2% in 2003 and acupuncture from 9.4% in 1997 to 13.0% in 2003. In addition to these main components of CM, general consultation services accounted for 3.4% in 1997. However, the utilization of these consultation only services decreased to 1.1% in 2003 (Table 5).

In this study, the logistic regression method was used to identify patient characteristics associated with the utilization patterns of CM and WM over time. Females used both health care services more than males. The higher utilization of CM in females has also been reported in Singapore 9 and the Western countries 1129303132 as well as in Taiwan 12141619. In practice, CAM has been used to treat postpartum conditions, menopause, and chronic diseases among women 16192932333435. Moreover, the age distribution of CM utilization peaked at 45–54 years. This finding is similar to those reported previously 1112161929303132. Since WM provides health-care services such as vaccinations for the children 36, the utilization was revealed to increase with age and peak in youngest group (2–7 years). Moreover, WM utilization was highest in the elderly people (≥ 65 years).

The utilization of CM was higher in the regular salary income group than the low-income group, farmers and fishermen and the other group. These results are similar to the previous findings that CAM users are those with higher education and in the middle to upper socioeconomic status 343738. WM was more frequently used by patients with severe diseases. These diseases require long-term supportive treatment and may not be suitable for CM. Although CM usage was more prevalent in the central and southern parts of Taiwan, CM was not frequently used in mountainous areas and offshore islands. One possible explanation for this phenomenon is the uneven distribution of CM providers, since there are only 6 CM hospitals and 24 CM clinics in eastern Taiwan 13.

In a previous study, 62.5% of the valid beneficiaries in the NHI Program have been reported to use CM at least once from 1996 to 2001 19. By controlling the co-variant factors, we found an increasing trend in the utilization of CM and WM. Moreover, 39,562 patients used CM services in each year whereas the corresponding figures for WM were 119,915. Most of patients had a single ambulatory visit annually. However, the mean utilization frequency of WM (12.55–13.34) was higher than that of CM (4.63–5.03). The numbers reported in this study may seem higher than those reported previously 19 because these values were based on only users, not total population. In addition, we revealed that over 90% of the CM services were provided by clinics whereas clinics only provided about two-thirds (60.5–67.8%) of the ambulatory WM services. These findings were consistent with those reported previously 151939. Under the NHI Program, copayments vary with the provider type. They are highest for ambulatory health care at medical centres and lowest for clinics 40. Moreover, the copayments in CM are lower than WM. Therefore, more than 90% of the CM users visited clinics and less than 10% CM hospitals.

Based on the ICD-9-CM code, we found that diseases of the respiratory system, diseases of the musculoskeletal system and connective tissue, injury and poisoning, signs, symptoms and ill-defined conditions, and diseases of the digestive system were the primary indications in CM. Similar patterns have been reported previously 1214. However, the top two primary indications were also the same as those in WM. We have also found that the utilization of these two health-care services was higher in winter than in the remaining seasons. These findings may be due to the fact that diseases of the respiratory system occurs more frequency in winter. These findings indicate that CM may be a feasible complementary choice for the patients with respiratory diseases.

Herbal medication is the major component of CM accounting for more than two-thirds of the ambulatory visits. In the United States, herbal products have an annual market of US$5.1 billion 41 and 38.2 million adults have used herbs and supplements 42. The herbal prescriptions in the Western countries are usually with a single herb whereas the majority of prescriptions in Taiwan are composed of 3–6 herbs and are often prescribed for three times a day 17. Under the NHI program, herbal formulas were provided as standard pharmaceutical products in standard powder forms. This strategy not only can control the quality of the prescriptions but also enable the standardization of herbal formulas. Acupuncture is another popular form of CAM 34567. It is mainly employed in the treatment of neurologic and musculoskeletal diseases 184344. In Taiwan, about 23% of people used acupuncture from 1996 to 2002 18. We obtained similar results in this study. In addition to an increasing trend in the utilization of acupuncture, the prevalence of muscle strain therapy was also increased in recent years. This increase may be due to the increase in the number of providers 13. However, further investigations are needed to confirm this suggestion.

This study is the first population-based investigation to determine the utilization patterns of CM and WM in Taiwan under the NHI Program. Since we obtained the data from a sufficiently large national representative sample from the NHI sample files, some of the common shortcomings in interview or questionnaire surveys may be avoided. The large sample size and the comprehensive datasets allow us to study a wide array of factors in the general population. However, we have excluded enrolees who were newborns aged below 2 years in 1997, those died in the study period, aliens, those with incomplete data, and those without continuous enrolment information. This manipulation of the data may lead to some selection bias. Since we used only the NHI claims data, we are unable to determine the utilization of CM services not covered by the NHI program. Therefore, the utilization of CM in Taiwan might be underestimated.

The utilization of CAM has rapidly increased in many countries during the last two decades 4567. CAM is rarely covered in national health systems. However, 75% of the Dutch population wanted hospitals to provide CAM 45 and 74% of the people in UK felt that complementary therapies should be available on the national health system 46. In Taiwan, the NHI Program is a comprehensive and universal health insurance program. This program not only covers conventional WM services, but also traditional CM services. Moreover, CM is popular and more than 60% of all beneficiaries of this health insurance system had used CM at least once a year 19. In this study, we found that the rate of increase in the use of CM was smaller than that in WM (Table 2) and that the increase was not apparently shown in the number of visits per user. Taiwan's experience of covering CM services under its national health insurance system may serve as an important reference to other countries. This strategy offers people another choice for medical care services with mandated coverage.

The authors declare that they have no competing interests.

LCC, NH, YJC, CHL and YTH contributed equally to this paper. FYK carried out the analysis. LCC, NH, YJC, CHL, FYK and HYT pictured the idea and drafted the manuscript. All authors have read and approved the final manuscript.