Insurance covered and non-covered complementary and alternative medicine utilisation among adults in Taiwan

(1)

Insurance covered and non-covered complementary and alternative

medicine utilisation among adults in Taiwan

Shu-Fang Shih

a,b

_{, Chih-Yin Lew-Ting}

b,_*

_{, Hsing-Yi Chang}

a

_{, Ken N. Kuo}

a

a_{Center for Health Policy Research and Development, National Health Research Institutes, No. 35, Keyan Rd., Zhunan Town, Miaoli County 35053, Taiwan} b_{Institute of Health Policy and Management, College of Public Health, National Taiwan University, No. 17, Xu-Zhou Rd., Taipei 10020, Taiwan}

a r t i c l e

i n f o

Article history:

Available online 22 July 2008 Keywords:

Taiwan

Complementary and alternative medicine (CAM)

Traditional chinese medicine National Health Insurance Help-seeking

a b s t r a c t

Complementary and alternative medicine (CAM) has been characterized by prevalence, cost, and patterns of use. Factors determining its utilization have also been analyzed, either generally or for speciﬁc diseases, but few studies have considered the determinants of its frequency of use. Taiwan’s pluralistic health care system and comprehensive insurance program covering Western medicine and traditional Chinese medicine (TCM) provide an interesting case to explore what forms of CAM people use, why and how often they use them. By using Taiwan’s 2001 National Health Interview Survey and linking it with National Health Insurance (NHI) claims data, this study aims to investigate the socio-demographic, economic, behavioral and health determinants of use and frequency of use for both non-covered and covered CAM in Taiwan. The former consists of services practiced by non-regulated or non-NHI contracted practitioners and the latter includes those TCM covered by NHI. Our study showed that the determinants of using non-covered and covered CAM differ with respect to socio-demographic and behavioral factors, and health needs. Our evidence suggests that future CAM research must delineate between use and frequency of use to better understand the underlying factors contributing to initiation and continuity of CAM use.

Introduction

The rise of interest in Complementary and Alternative Medicine (CAM) research may come from either an aware-ness of the rising popularity of non-orthodox therapies, a concern over the dominance and monopoly of the bio-medicine (Lew-Ting, 2003), or the reimbursement issue of these services. Previous studies have reported the preva-lence, cost, and patterns of CAM use (Eisenberg et al., 1993; Ernst, 2000; Harris & Rees, 2000; Ni, Simile, & Hardy, 2002; Thomas & Coleman, 2004). Some have also investi-gated the factors associated with its use either generally or for speciﬁc diseases (Verhoef, Balneaves, Boon, &

Vroegindewey, 2005; Maskarinec, Shumay, Kakai, & Gotey, 2000; Garrow & Egede, 2006; Grzywacz et al., 2006). How-ever, from the health policy perspective, a study of whether these treatment modalities are used or not seems too naı¨ve. In addition, how often people use them and why, may be more important because they can help us to better identify the needs of the users on a long-term basis.

Taiwan’s medical pluralism is characterized by a dual professional system consisting of two great medical tradi-tions: Traditional Chinese medicine (TCM) and Western medicine (Kleinman, 1980). Both are covered under Tai-wan’s National Health Insurance (NHI). However, since Western medicine still dominates health care resources (Chi, 1994) and political jurisdiction (Lew-Ting, 2003), TCM is still regarded as CAM. Coverage of TCM under the NHI includes Chinese medicines derived from traditional medicinal herbs, acupuncture, moxibustion, and trauma-tology manipulative therapy. TCM services covered by

*Corresponding author. Tel.: þ886 2 3322 8058; fax: þ886 2 2351 5509.

E-mail addresses:annsshih@nhri.org.tw(S.-F. Shih),chihyin@ha.mc. ntu.edu.tw(C.-Y. Lew-Ting),hsingyi@nhri.org.tw(H.-Y. Chang),kennank@ nhri.org.tw(K.N. Kuo).

Contents lists available atScienceDirect

Social Science & Medicine

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / s o c s c i m e d

(2)

NHI are called covered CAM in this study. Outside the NHI, also part of the broader health care system (Kleinman, 1980), are some TCM treatments mentioned above prac-ticed by non-NHI contracted providers and other CAM modalities of healing art such as folk remedies or spiritual healing etc. These are called non-covered CAM, which are less examined in Taiwan because it is usually considered not legitimate enough for policy research. With so many health care choices, Taiwan provides an interesting case to investigate people’s healthcare-seeking behavior regard-ing both covered and non-covered CAMs.

Linking Taiwan’s 2001 National Health Interview Survey (NHIS) with NHI claims data, this study describes the prev-alence, types and patterns of use for both non-covered CAM and covered CAM and investigates socio-demographic, economic, behavioral and health determinants of use and frequency of use for both non-covered and covered CAM in Taiwan. Results from our study not only suggests that delineating between use and frequency of use can help to better understand the underlying motivations for CAM use but also provide more insights of the health needs of the people and healthcare-seeking behavior within and outside the health insurance system.

Methods Data source

This study uses two data sets: the 2001 NHIS and 2001 NHI claims data. The 2001 NHIS collected health, health care, and socio-demographic information from a nationally representative sample of Taiwan’s non-institutionalized population (Shih et al., 2003). Survey data were linked to the 2001 NHI claims data for those participants who signed the informed consent allowing the linking of their NHI data. The utilization of covered CAM was drawn from the NHI claims data. We excluded the people with incomplete information. Only those people who are 20 years and older were included for their independent healthcare-seeking behavior. Therefore, the ﬁnal samples consist of 10,600 and 9187 adults for the multivariate analyses of non-covered and non-covered CAM, respectively.

Measures

In this study, CAM is defined as treatment modalities that are not fully recognized by Western medicine in Tai-wan such as TCM and other treatment modalities practiced by non-qualified or non-regulated practitioners or practi-tioners in folk sector, including both sacred and secular healing modalities (Kleinman, 1980). We classified CAM into ‘‘covered CAM’’ and ‘‘non-covered CAM’’ categories based on whether the modalities are practiced by practi-tioners with official licenses or the practipracti-tioners are NHI contracted providers. Taiwan’s NHI covers TCM including traditional Chinese medicines derived from medicinal herbs, acupuncture, moxibustion, and traumatology ma-nipulative therapy. However, some TCM services provided by non-NHI contracted or non-qualified practitioners are defined as non-covered CAM in this study. Information about utilization of covered CAM was drawn from the

insurance claims data for each individual. The advantage of using linked data is that it provides clear and reliable records of which insurance covered CAM services the participants had used and how often they sought those ser-vices. It also reduced some possible recall bias in answering survey questions on past utilization.

The information about non-covered CAM was mainly drawn from the survey data. In the 2001 survey, respon-dents were asked to answer the question: ‘‘In the last one month, have you ever undergone any of therapies in this list in any non-medical setting?’’ The interviewer then read from a list covering the following seven CAMs: (1) shou-jing, a process in which a ‘‘demon’’ is compelled to leave a person who is physically or emotionally upset by an actual or imagined event; (2) ciou shen wun bu, a process in which the patient either draws lots from a bundle of bamboo sticks with special instructions or prays which is interpreted by an intermediary to ask for divine advice to treat physical or mental maladies; (3) siiang hui, eating the ash of temple incense; (4) scrapping, a process in which the skin is scrapped with a thin, ﬂat object to unblock clogged energy in the body; (5) cupping, a process in which a suction cup is placed on a body part to suck out unclean energy; (6) bodywork, e.g., tuei-na massage and chiroprac-tic care; and (7) bone setting and die-da sun shang, pro-cesses in which bones or ligaments are manipulated to treat musculoskeletal problems. In addition, the partici-pants were asked an open-ended question regarding whether they had used any other similar therapies during the month leading up to the interview. In answering this question, additional CAMs were reported such as acupunc-ture, qigong (a type of relaxing, meditative body/mind/ spirit exercise), bloodletting, foot massage, the drinking of water blessed by a psychic, fortune changing (a process in which things are done or items are worn to change ones fate or reduce unlucky inﬂuences), prayers for divine inter-vention or advice and drawing from a bundle of sticks inscribed with various instructions on how to solve one’s problems. If respondents reported having used any of these therapies, they were then asked how frequent in the past month they used them, why did they seek those types of therapy and how much they had spent on their latest ther-apeutic session.

A total of 14 types of non-covered CAM was reported as above. These modalities were further classiﬁed into 6 do-mains based on their nature and functions: (1) spiritual healing or healing involved supernatural forces (anything including rituals); (2) folk remedies (anything applied to or remove from the body using some accessories to relieve stagnation), (3) chiropractic care and massage (anything involving a practitioner manipulating the subject’s body); (4) osteopathy (traumatology manipulative therapies): bone setting and die-da sun shang; (5) acupuncture by non-NHI contracted practitioners; and (6) qigong (any type of meditation or exercise).

We used 36-Item Short-Form Health Survey Taiwan ver-sion 1.0 (SF-36) to investigate the associations between the multi-dimensional health status and CAM use so that we could observe more speciﬁc dimensions associated with CAM use. The 8 dimensions covered by SF-36 include physical functioning, role limitation-physical, role

(3)

limitation-emotional, mental health, social functioning, bodily pain, vitality, and general health. To take into account Taiwan’s pluralisms in health system, we also constructed a variable called typical healthcare-seeking behavior from the open-ended question in the survey: ‘‘What do you usu-ally do when you feel ill?’’ Most participants reported two behaviors. We used this question to categorize people into 4 types: (1) people seeking Western medicine only, (2) people seeking both Western medicine and others, (3) people who seek TCM and therapies others except West-ern medicine; and (4) other combinations of care. Other vari-ables found to be associated with CAM use in previous studies such as age, gender, average annual household in-come, education level, ethnicity, marital status, number of having certain chronic conditions, religion, residential area, urbanization level and unhealthy lifestyle behaviors (such as smoking, drinking or betel nut chewing) were included in this study.

Statistical analysis

We used a two-part model to distinguish the factors related to use and use frequency for both non-covered and covered CAM. In the ﬁrst part, we used logistic regres-sion to analyze the factors associated with use of two sets of CAMs. In the second part, due to the over-dispersion of the number of visits to non-covered CAM and covered TCM, we used a negative binomial model to analyze factors associ-ated with use frequency (Mullahy, 1986). Robust estimation was applied to take into account the heterogeneity. Results

Proﬁle of CAM use

Table 1presents the prevalence rate and the number of visits per month for covered and non-covered CAM. In the survey, 6.05% of the respondents reported using at least one type of non-covered CAM in the month prior to the survey. The most frequently used were chiropractic care and mas-sage (44.70%), followed by folk remedies (35.63%), spiritual healing (23.39%), and bone setting and die-da sun shang (10.05%). The average number of monthly visits per user

for all non-covered CAM was 3.26, the highest was for bone setting and die-da sun shang (4.02) and the lowest for spiritual healing (1.44). Others averaged between 2 and 4 visits. With regard to results from the claims data, 6.67% had ever used covered CAM one month before the in-terview. Under covered CAM, 4.95% of respondents had ever used traditional Chinese herbal medicine, and 2.25% ever used acupuncture, moxibustion, and traumatology manipulative therapy. Overall, the average number of visits to covered CAM was 2.59 per month. The average number of visits that traditional Chinese medicinal herbs were used was 2.47 while either acupuncture, moxibustion, or traumatology manipulative therapy had been used 2.25 times.

Eighty-one percent of those using non-covered CAM used a single type, 14.32% a combination of two types, and 4.48% three or more types. The most used combination of modalities was scraping and massage (23.66%), and the second involved three modalities: cupping, chiropractic and massage (21.37%). When considering users of non-covered CAM and non-covered CAM altogether, around 85.90% could be categorized as single type users and around 10.56% combination users. The most common combina-tions of the two were chiropractic care and massage with the covered CAM (41.25%), followed by a combination with scraping with covered CAM (28.75%) and cupping with covered CAM (23.75%).

Based on self-reported data from the NHIS, people most often sought non-covered CAMs for aches (12.92%), peace of mind (9.46%), and unexplained symptoms (8.95%), followed by heat exhaustion (6.90%), lower back pain (5.33%), insomnia and sleeping disorder (3.78%), inﬂuenza (3.76%), tiredness and exhaustion (3.68%), and headaches (3.17%); 4.65% refused to express their reasons for using non-covered CAM. Those most likely not to give a reason were users of folk remedies (32.82%) and spiritual healing (8.41%).

Factors associated with use and use frequency of non-covered and covered CAM

Table 2shows the results from multivariate analysis of the determinants of decision to use and frequency of use for both non-covered and covered CAMs. Most likely to use non-covered CAM were middle-age people, women, people with higher education levels, those following tradi-tional religions or Buddhism, those with more unhealthy behaviors, those with chronic conditions, those lived in central area, mixed users of health care services, and those with bodily pain. Our results showed that factors associated with use of non-covered CAM were also found to be associ-ated with frequency of use. However, other factors such as ethnicity, urbanization level, and other health dimensions also determined the frequency of usage. In particular, those who had better physical functioning, and less bodily pain were less likely to be frequent users while those with better social functioning were more likely to make more visits to practitioners of non-covered CAM.

Only age, residential area, typical healthcare-seeking behavior, and bodily pain were associated with the decision to use covered CAM. Those aged 20–24 or 35–64 years old,

Table 1

Prevalence and average number of visits per month for covered and non-covered CAM

Type of CAMs Prevalence

rate (%)

Average number of visits

Non-covered CAMs 6.05 3.26

(1) Spiritual healing or healing involved supernatural forces

1.41 1.44

(2) Folk remedies 2.16 2.51

(3) Chiropractic care and massage 2.70 3.52

(4) Osteopathy 0.61 4.02

(5) Acupuncture 0.19 2.39

(6) Qigong 0.01 1.50

Covered CAM 6.67 2.59

(1) Chinese medicines derived from traditional medicinal herbs

4.95 2.47

(2) Aupuncture, moxibustion, traumatology manipulative therapy

(4)

those living in central and northern Taiwan, mixed users of health care, and those with bodily pain were more likely to use covered CAM. Similar to the ﬁndings from non-covered CAM, these factors were not only associated with decision

to use but also found to be associated with frequency of use. Other factors such as education, ethnicity, religion, and general health conditions were also found to be associ-ated with frequency of covered CAM use. Those with junior

Table 2

Multivariate analysis of factors associated with use and use frequency of CAM

Characteristics Use of CAM Frequency of CAM use

Odds ratio Coefﬁcients

Non-covered Covered Non-covered Covered

Age (ref ¼ 65 or over)

20–24 1.56 2.01** 0.78*** 0.62**

25–34 1.90** 1.51 0.87*** 0.53**

35–49 1.92*** 1.50* 0.99*** 0.44**

50–64 1.66** 1.59** 0.99*** 0.63***

Gender (ref ¼ male)

Female 1.80*** 1.07 0.53*** 0.03

Education (ref ¼ elementary or illiteracy)

Junior High School 0.97 1.19 0.10 0.26*

Senior High School 1.44** 0.97 0.41*** 0.03

College or above 1.47* 0.99 0.30** 0.11

Ethnicity (ref ¼ Mainland and others)

Fukianese 1.10 1.19 0.39*** 0.27*

Hakka 1.10 1.39 0.60*** 0.33*

Indigenous 1.54 0.87 0.69* 0.40

Average annual household income

(ref ¼ below 30,000 NT dollars)

30,000–49,999 0.89 1.00 0.07 0.06

50,000–69,999 0.80 0.94 0.08 0.09

70,000–99,999 1.11 1.04 0.06 0.02

100,000 or above 1.05 1.14 0.16 0.04

Religion (ref ¼ Christianity, Catholicism or Moslemism)

No religion 1.07 1.12 0.07 0.23

Traditional religions 1.57* 1.14 0.40* 0.33

Buddhist 1.81** 1.26 0.79*** 0.44*

Marital Status (ref ¼ otherwise)

Married 1.14 1.13 0.01 0.11

No. of unhealthy behavior (ref ¼ no)

One 1.51*** 1.09 0.54*** 0.00

Two or three 1.99* 0.63 1.28*** 0.26

Chronic condition (ref ¼ no)

Yes 1.25* 1.12 0.31*** 0.14

Urbanization level (ref ¼ medium)

Highest 1.20 1.00 0.25** 0.07

High-medium 1.15 1.08 0.20* 0.14

Lowest 1.06 0.87 0.18* 0.06

Residential area (ref ¼ Southern)

Northern 1.19 0.79* 0.06 0.11

Central 1.48** 1.30* 0.10 0.29**

Eastern 0.95 0.80 0.62*** 0.23

Typical healthcare-seeking behavior (ref ¼ Western medicine only)

Western and others 1.59*** 2.18*** 0.55*** 0.77***

Traditional Chinese medicine and others except Western medicine

2.53*** 4.72*** 1.27*** 1.59***

Others 1.53** 0.76 0.70*** 0.26

SF-36 health status measure

Physical functioning 1.02 1.00 0.08*** 0.03 Role-physical 0.98 1.00 0.01 0.00 Role-emotional 0.99 0.99 0.01 0.01 Social functioning 1.02 1.05 0.04* 0.04 Bodily pain 0.76*** 0.91*** 0.30*** 0.07*** Vitality 1.03 1.05 0.01 0.02 Mental health 1.00 0.98 0.00 0.01 General health 0.98 0.97 0.01 0.05* Note: *p < 0.05, **p < 0.01, ***p < 0.001.

(5)

high education level, those of Fukianese or Hakka heritage, those following Buddhism, those with worse general health were more likely to make frequent visits to providers of CAM covered by the NHI program.

Discussion

Using a national representative sample from household survey and insurance claims data, this study assesses the determinants of use and frequency of use for both non-covered and non-covered CAM. This study moves beyond investigating whether CAM is used or not, but instead also examines the determinants of use frequency in the pursuit of deeper understanding about CAM use. This study is based on the assumption that frequency implies continuity and might, therefore, indicate preference for or effectiveness of certain modalities. More predictors were found to be associated with use frequency of both non-covered and covered CAM, suggesting that further examining the frequency of CAM use may provide a more precise picture of the phenomena.

According to our study, although almost all people in Taiwan are covered by insurance, there is still a higher demand for non-covered rather than covered CAM. The greater number of visits to non-covered CAM could have been associated with the greater variety of non-covered CAM available and longer course of treatment for some modalities. Unlike previous studies, none of our non-covered CAM users complained about their negative expe-riences with conventional/orthodox medicine. They did state that there were some problems that health care pro-viders within insurance system could not treat, showing their concerns about the efﬁcacies of these non-covered CAM services. The phenomena that people combined care from non-covered and covered CAM providers shows that people in Taiwan not only recognize the limitations of ser-vices provided within the insurance system but also tend to use health services selectively and adaptively (Lew-Ting, 2005). Since Taiwan’s Medical Care Act does not regard non-covered CAM services as invasive, non-covered CAM is not regulated by Taiwan’s Department of Health. Because their efﬁcacy is questionable and they are not regulated, there is a need for evaluation and control.

Although it is hard to compare our findings with previ-ous studies due to different definition of CAMs, study design, and health system setting, our findings on the socio-demographic determinants of non-covered CAM use such as age, gender, and education were consistent with previous representative population studies in the United Kingdom (Thomas & Coleman, 2004), United States (Eisenberg et al., 1993), and Australia (MacLennan, Myers, & Taylor, 2006). However, as for covered CAM, among these variables, only age was associated with its use decision although the pattern of relationship is slightly different from that for non-covered CAM. Our study showed that women tended to use and visit more to non-covered CAM while this situation was not the case for covered CAM. This might be partly because women were more willing to adopt or try different kinds of care outside of the insur-ance system for treating their health problems, and partly because it is easier for women than men to access to

non-covered CAM information through their social net-works. In addition, educational background was not associ-ated with covered CAM use, suggesting that access to covered CAM did not vary across people with different education levels. Unlike some previous studies, income was not associated with use or frequency of use whether the modalities were covered or not. There are two possible explanations. First, under the NHI, people, especially for low income groups, may have more available financial sources, which, in turn, lead to better access to non-covered CAM. The spillover effect of NHI allows people to consume resources across its boundary. Second, it is also likely that people are willing to pay as much as they can to cure their diseases outside the NHI because they may have experi-ences with the limitations of conventional medicine. The perceived inefficacy and deficiency of services under the NHI pushes people make alternative choices.

In addition to the socio-demographic factors mentioned above, health belief or behavioral characteristics of insur-ance covered and non-covered CAM users also varied. Given that there were various types of non-covered CAM and less diversified therapeutic modalities of covered CAM, variables as proxies for health beliefs such as religion or having unhealthy behaviors were found to be more asso-ciated with non-covered CAM use than covered CAM use. Those who practiced traditional religions or Buddhism were more likely to use non-covered CAM; this might be due to the nature or philosophy of some non-covered CAM. Supernatural powers and sacred deity inherent in these modalities seem more acceptable to these people. Our study also found that a greater number of unhealthy behaviors led to choose and use more non-covered CAM. This could be found among people who were younger and healthier and may not in need of health care services provided by the NHI system. They might seek for non-covered CAM as a type of preventive medicine to ward of illness and protect their health. Further investigation is needed to confirm this hypothesis. Due to lack of sufficient sample size for each type of non-covered CAM, we could not conduct multivariate analysis to identify factors associated with use of specific type of CAM. Efforts such as collecting more complete data set on various types of CAMs should be encouraged as suggested byAstin (2000). Our findings confirmed the hypothesis that CAM use can be seen as being more compatible with the patients’ spiri-tual/religious philosophy, or beliefs regarding the nature and meaning of health and illness. For two sets of CAMs, more frequent visits of CAM use by Fukianese and Hakka ethnic groups has illustrated a deep cultural and belief in-fluences on the continuity of CAM use. Those who practiced Buddhism were more likely than those with other western religions to visit more to covered and non-covered CAM; this also highlights the important role of cultural and social factors. As for health needs, our study indicates that more health dimensions found to be associated with use fre-quency than with use decision for both two sets of CAMs and the dimensions differed for two decisions except bodily pain. Our study suggests that for both two sets of CAM, they were used as means to treat illness in order to restore health rather than to prevent diseases or maintain health, especially in the case of pain relief. That those who had

(6)

better social functioning tended to use more non-covered CAM might be due to the inﬂuences by others or informa-tion from their social groups/network members (Honda & Jacobson, 2005).

Overall, factors such as age, residential area, typical healthcare-seeking behavior and health status were found to be not only significantly associated with use but also with frequency of use for both non-covered and covered CAM, although the pattern of relationship was slightly dif-ferent in some cases. Middle-age persons tended to choose non-covered CAM while those aged 20–24 and 50–64 tended to choose covered CAM. However, the tendency to be frequent users across age groups was slightly different from the tendency to be users for non-covered CAM while it was found to be similar for covered CAM. Residential area was associated with availability and accessibility to CAM use, which cannot be controlled in our study due to lack of data, however, high frequent use of covered CAM in cen-tral area might be related to higher density of Chinese med-icine doctor (Lee, Chou, Chen, & Chang, 2004). Typical healthcare-seeking behavior implied the usual care pro-vider people chose and reflected the freedom of care seek-ing under Taiwan’s NHI (Cheng, 2003). More frequent use of both two sets of CAM were only associated with mixing the traditional Chinese medicine with other types of vices but not with mixing Western medicine and other ser-vices. Health needs remained statistically significantly associated with use and frequency of use for both two sets of CAMs. Of all health needs, body pain remained the only ailment to be highly associated with use and fre-quency of use for both covered and non-covered CAMs. According to a recent Taiwan study, musculoskeletal and neurologic disorders were two major diseases commonly treated with acupuncture covered by the NHI (Chen, Kung, Chen, & Hwang, 2006). Therefore, further investiga-tion might be needed to understand how people use ser-vices within and outside NHI system to treat pain-related conditions.

This study has three limitations. First, since types of non-covered CAM being used were self-reported, the prev-alence might be underestimated. Most participants only thought of some local or folk therapies or modalities that have existed in Taiwan for a long time but not other types of modalities such as aromatherapy, homeopathy, or com-plementary nutrition. However, this might also indicate that those ‘‘foreign CAMs’’ or others to be considered as CAM in other countries might not be so popular or not be perceived by the respondents as CAM in year 2001. Second, due to lack of measures in the survey data, we were not be able to directly examine the inﬂuences of health beliefs and attitudes on CAM use. To capture these effects, factors such as education, religion or ethnicity were used as prox-ies instead. Third, we lost some participants who did not agree that survey and NHI data be linked. However, we do not believe that declining to give consent affected our results much although one study has shown that the elderly, the illiterate, those with lower income, and subur-ban area residents were more likely to decline consent (Huang, Shih, Chang, & Chou, 2007).

In conclusion, our study indicates that covered and non-covered CAM users were characterized by different

socio-demographic and behavior factors, depending on the contents of services provided. Notably, it has evi-denced unmet health needs of people even though most of them are covered by the NHI. This is what policy makers should pay greater attention to while planning an efﬁcient and quality health care system in a society like Taiwan. Finally, in terms of measurements of CAM use, our evidence suggests that future research must delin-eate between use and frequency of use to better under-stand the underlying factors contributing to initiation and continuity of CAM use.

References

Astin, J. A. (2000). The characteristics of CAM users: a complex picture. In M. Kelner, B. Wellman, B. Pescosolido, & M. Saks (Eds.), Complemen-tary and alternative medicine: Challenge and change (pp. 101–114). Amsterdam: Harwood Academic.

Chen, F.-P., Kung, Y.-Y., Chen, T.-J., & Hwang, S.-J. (2006). Demographics and patterns of acupuncture use in the Chinese population: the Taiwan experience. The Journal of Alternative and Complementary Medicine, 12(4), 379–387.

Cheng, T.-M. (2003). Taiwan’s new national health insurance program: genesis and experience so far. Health Affairs, 22(3), 61–76. Chi, C. (1994). Integrating traditional medicine into modern health care

systems: examining the role of Chinese medicine in Taiwan. Social Science & Medicine, 39(3), 307–321.

Eisenberg, D. M., Kessler, R. C., Foster, C., Norlock, F. E., Calkins, D. R., & Delbanco, T. L. (1993). Unconventional medicine in the United States – prevalence, costs, and patterns of use. New England Journal of Medicine, 328(4), 246–252.

Ernst, E. (2000). Prevalence of use of complementary/alternative medicine: a systematic review. Bulletin of the World Health Organiza-tion, 78(2), 252–257.

Garrow, D., & Egede, L. E. (2006). National patterns and correlates of complementary and alternative medicine use in adults with Diabetes. The Journal of Alternative and Complementary Medicine, 12(9), 895–902.

Grzywacz, J. G., Suerken, C. K., Quandt, S. A., Bell, R. A., Lang, W., & Arcury, T. A. (2006). Older adults’ use of complementary and alterna-tive medicine for mental health: ﬁndings from the 2002 national health interview survey. The Journal of Alternative and Complementary Medicine, 12(5), 467–473.

Harris, P., & Rees, R. (2000). The prevalence of complementary and alternative medicine use among the general population: a systematic review of the literature. Complementary Therapies in Medicine, 8, 88–96.

Honda, K., & Jacobson, J. S. (2005). Use of complementary and alternative medicine among United States adults: the inﬂuences of personality, coping strategies, and social support. Preventive Medicine, 40(1), 46–53.

Huang, N., Shih, S.-F., Chang, H.-Y., & Chou, Y.-J. (2007). Record linkage research and informed consent: who consents? BMC Health Services Research, 7, 18.

Kleinman, A. (1980). Patients and healers in the context of culture: An explo-ration of the borderland between Anthropology, Medicine, and Psychia-try. Berkeley: University of California Press.

Lee, C.-H., Chou, Y.-J., Chen, L.-S., & Chang, H.-J. (2004). Utilization of am-bulatory Chinese medical services under the National Health Insur-ance in Taiwan. Taiwan Journal of Public Health, 23(2), 100–107. Lew-Ting, C.-Y. (2003). Who uses non-biomedical, complementary and

alternative health care? Sociodemographic undifferentiation and the effects of health needs. Taiwan Journal of Public Health, 22(3), 155–166.

Lew-Ting, C.-Y. (2005). Antibiomedicine belief and integrative health seeking in Taiwan. Social Science & Medicine, 60(9), 2111–2116. MacLennan, A. H., Myers, S. P., & Taylor, A. W. (2006). The continuing use

of complementary and alternative medicine in South Australia: costs and beliefs in 2004. Medical Journal of Australia, 184(1), 27–31. Maskarinec, G., Shumay, D. M., Kakai, H., & Gotay, C. C. (2000). Ethnic

dif-ferences in complementary and alternative medicine use among can-cer patients. Journal of Alternative and Complementary Medicine, 6(6), 531–538.

Mullahy, J. (1986). Speciﬁcation and testing in some modiﬁed count data models. Journal of Econometrics, 33(3), 341–365.

(7)

Ni, H., Simile, C., & Hardy, A. M. (2002). Utilization of complementary and alternative medicine by United States adults. Results from the 1999 National Health Interview Survey. Medical Care, 40(4), 353–358. Shih, Y.-T., Chang, H.-Y., Liu, J.-P., Lin, H.-S., Chang, M.-C., Chang, F.-C., et al.

(2003). The design, contents, operation and the characteristics of the respondents of the 2001 National Health Interview Survey in Taiwan. Taiwan Journal of Public Health, 22(6), 419–430.

Thomas, K., & Coleman, P. (2004). Use of complementary or alternative medicine in a general population in Great Britain. Results from the National Omnibus Survey. Journal of Public Health, 26(2), 152–157. Verhoef, M. J., Balneaves, L. G., Boon, H. S., & Vroegindewey, A. (2005).

Reasons for and characteristics associated with complementary and alternative medicine use among adult cancer patients: a systematic review. Integrative Cancer Therapies, 4(4), 274–286.