Acceptance and Use of Complementary and Alternative Medicine Among Medical Specialists: A Systematic Review and Meta-analysis

Otolaryngology (43%; 95%CI: 30-57%), Anesthesiology (42%; 95%CI: 37-47%), Internal Medicine (38%; 95%CI: 36-41%), Physical Medicine and Rehabilitation (32%; 95%CI: 24-41%), and Surgery (25%; 95%CI: 22-29%). Based on the studies, meta-regression showed no statistically signicant difference across geographic regions, economic levels of the country, or sampling methods.


Abstract Background
Complementary and Alternative Medicine (CAM) has gained popularity among the general population but its acceptance and use among medical specialists have been inconclusive.

Methods
We conducted a systematic literature search in PubMed and Scopus databases for the acceptance and use of CAM among medical specialists.
Each article was assessed by two screeners. Only survey studies relevant to the acceptance and use of CAM among medical specialists were reviewed. The pooled prevalence estimates were calculated using random-effects meta-analyses.

Conclusion
Acceptance and use of CAM were moderate and varied across medical specialists.

Systematic review registration
This systematic review has been registered in PROSPERO (CRD42019125628) and the protocol can be accessed at http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42019125628.

Background
Medical specialist is a healthcare professional who has undertaken specialized medical studies to diagnose, treat and prevent illness, disease, injury, and other physical and mental impairments in humans, using specialized testing, diagnostic, medical, surgical, physical and psychiatric techniques, through application of the principles and procedures of modern medicine [1].
Complementary and Alternative Medicine (CAM) is de ned as medicine or treatment which is not considered as conventional (standard) medicine. National Center for Complementary and Integrative Health (NCCIH) categorized most types of complementary medicines under two categories: (1) natural products, and (2) mind-body practices [2]. Natural products include herbs, vitamins, minerals, and probiotics whereas mind-body practices include yoga, chiropractic, massage, acupuncture, yoga, meditation, and massage therapy. Types of CAM may vary across studies, but they overlap in most senses.
CAM is used by people throughout the world. A study showed that the prevalence estimate of CAM usage from 32 countries from all regions of the world to be 26.4%, ranging from 25.9% to 26.9%. For example, in 2013, the prevalence use of CAM in Australia, the USA, United Kingdom and China were 34.7%, 21.0%, 23.6%, and 53.3%, respectively. The prevalence estimate of CAM satisfaction was as high as 71.9%, ranging from 71.0% to 72.7% [3]. Although patients are highly satis ed with CAM treatment, however, professional health care providers who are medical doctors do not offer CAM because it is not part of the standard medical care services. A study showed that less than 20% of the medical doctors were very comfortable in answering questions about CAM [4] so patients who do not have the option to use CAM instead of standard medical care will be lost to follow-up.
In the eld of oncology, the 5-year survival rate of breast cancer patients who refused standard treatment was 43.2% (95% CI 32.0-54. 4) whereas for those who underwent the standard treatment, it was 81.9% (95% CI 76.9-86.9) [5]. When CAM was used, the 5-year survival rate was worse. The 5-year survival rate of cancer patients who used CAM versus those who used standard treatment were strati ed by cancer type were as follows: [6]  The 28-day mortality of patient with Sepsis and Acute gastrointestinal injury (AGI) who received CAM bundle with conventional therapy was statistically signi cantly lower than those who received only conventional therapy (21.2% vs 32.5%, p-value = 0.038) [7].
There are few studies that have investigated the acceptance and use of CAM. Aside from that, previous studies could not compare the use of CAM across medical specialties. Also, many studies could not determine the effect of specialist, use of CAM by region and economic level of the country, survey method, and sampling method.
This systematic review and meta-analysis aimed to identify studies that have accepted and used CAM among medical specialists. The selected studies must quantify the number of acceptance or usage of CAM by medical specialist.

Protocol and registration
This systematic review has been registered in PROSPERO (CRD42019125628) and the protocol can be accessed at http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42019125628.

Literature search
This systematic review and meta-analysis were conducted and reported according to the PRISMA statement guidelines. A systematic literature search was performed by two independent authors (PP and KP) using PubMed and Scopus databases. The search was limited to observational studies of human subjects and the English language. The medical specialist's perspective related to CAM studies were focused. For the Scopus database, the following combinations were applied: (ALL("Traditional Medicine") OR ALL("Alternative Medicine") OR ALL("Complementary Medicine") OR ALL("Acupuncture Therapy") OR ALL("Holistic Health") OR ALL("Homeopathy") OR ALL("Spiritual Therapies") OR ALL("Faith Healing") OR ALL("Yoga") OR ALL("Witchcraft") OR ALL("Shamanism") OR ALL("Meditation") OR ALL("Aromatherapy") OR ALL("Medical Herbalism") OR ALL("Mind-Body Therapies") OR ALL("Laughter Therapy") OR ALL("Hypnosis") OR ALL("Tai Ji") OR ALL("Tai Chi") OR ALL("Relaxation Therapy") OR ALL("Mental Healing") OR ALL("Meditation")) AND (ALL("Health care provider") OR ALL("Health care providers") OR ALL("Health personnel")) AND PUBYEAR AFT 2001 AND PUBYEAR BEF 2018 AND DOCTYPE(ar)

Selection of studies
The titles and abstracts of the primary studies identi ed in the electronic search were screened by the same two authors. Duplicated studies were excluded. For the meta-analysis, the following inclusion criteria were set: (1) medical specialist's perspective, (2) prevalence of acceptance or usage of CAM, (3) observational study design, and (4) published between 2002 to 2017. The following exclusion criterion was set: (1) Not relevant to the practice. We contacted the authors for studies that had incomplete and unclear information. If the authors did not respond within 14 days, we proceeded to analyze the data we had. Any disagreement was resolved through discussion and the nal determination was made by the rst author (PP).

Data extractionand management
Two authors worked independently to review and extract the following variables: (1) general information, including the name of the studies, authors, and publication year, (2) characteristics of the studies, including the design of the studies, sampling method, country, and setting, (3) characteristics of the participants, including sample size, response, and type of specialty, and (4) outcomes, including the prevalence of acceptance, and usage of CAM. All relevant text, tables, and gures were examined for data extraction. Discrepancies between the two reviewers were resolved by the rst author (PP).

Study quality/Risk of bias
We used the tool developed by Hoy et al [8] to evaluate the study quality/risk of bias of the studies included in the analysis. The tool has 11 items: (1) national representativeness, (2) target population representativeness, (3) random selection or census undertaken, (4) minimal nonresponse bias, (5) data collection direct from the subject, (6) de nition of the case used, (7) valid and reliable instrument, (8) same mode of data collection for all subjects, (9) length of shortest prevalence period, (10) appropriate numerator and denominator used, and (11) summary assessment. Items 1 to 4 assessed the external validity, items 5 to 10 assessed the internal validity, and items 11 evaluated the overall study quality/risk of bias. Each item was assigned a score of 1 (High quality/Low risk) or 0 (Low quality/High risk), and the scores were summed to generate an overall quality score that ranged from 0 to 10. According to the overall score, we classi ed the studies as having a high quality/low risk of bias (>6), moderate quality/risk of bias (4 to 6), and low quality/high risk of bias (<4). Two authors (PP and KP) independently assessed the study quality/risk of bias and any disagreement was resolved by discussion and consensus.

Con ict of interest
We assessed the con ict of interest of the authors' declarations in the studies.

Statistical analysis
Unadjusted prevalence estimates of acceptance and usage of CAM were calculated based on the information of crude numerators and denominators provided by the studies and medical specialty [9]. Pooled prevalence was estimated from the prevalence as reported by the eligible studies. Forest plots were generated displaying the prevalence with a 95% CI for each study. The overall random-effects pooled estimate with its 95% CI were reported. To examine the magnitude of the variation between the studies, we quanti ed the heterogeneity by using I 2 and its 95% CI.
To assess the level of heterogeneity as de ned in Chapter 9 of the Cochrane Handbook for Systematic Reviews of Interventions, the following I 2 cut-offs for 0% to 40% represented that the heterogeneity may not be important, 30% to 60% may represent moderate heterogeneity, 50% to 90% may represent substantial heterogeneity, 75% to 100% represented that there was a considerable heterogeneity. For the X 2 test, statistical heterogeneity of the included trials was assessed with a p-value of less than 0.05 (statistically signi cant). The random-effects meta-analysis by DerSimonian and Laird method was used, and statistical heterogeneity was encountered. The meta-analysis was performed using Stata/MP software version 15 (StataCorp 2017, College Station, TX).

Additional Analysis
Meta-regression was performed to investigate the pooled prevalence differences between various regions (African region, region of the Americas, Eastern Mediterranean region, European region, Southeast Asia region, Western Paci c region, and mixed region) [10], economic levels of the country (low-income, lower-middle-income, upper-middle-income, high-income, and mixed-income) [11], and the sampling method (random and convenience sampling).

Selection of the studies
The literature search yielded 5,628 articles. After 794 duplicates were removed, 4,831 titles and abstracts were screened, and 4,719 irrelevant articles were removed. Of 115 articles selected for full-text screening, 62 were excluded for the following reasons: two were not relevant to this study's objective, 17 had the wrong target population, 22 did not have the study design required for this review, two study was not published in English, 19 did not have full-text available, and 28 did not provide the prevalence. Finally, a total of 25 articles, published between 2002 and 2017, ful lled the selection criteria and were included in this meta-analysis (Fig 1).
Meta-regression showed that there were no signi cant differences in the pooled prevalence of CAM acceptance by region, economic levels of the country, and the sampling method ( Table 3).
Assessment of study quality/risk of bias/con ict of interest A total of 24 (96%) studies were categorized as high quality/low risk of bias, whereas one (4%) was categorized as moderate quality/moderate risk of bias. No study met the criteria of low quality/high risk of bias (Fig 6). Only ve studies (20%) declared that there were con icts of interest.

Discussion
This study is the rst of its kind to compare the acceptance and usage of CAM across various medical specialties. Nearly three-quarters of the specialties accepted CAM whereas nearly a third were using CAM.
The synthesis of all prevalence estimates of acceptance and usage was 52% and 45%, respectively. The highest prevalence of acceptance was in Family Medicine, followed by Psychiatry and Neurology, Neurological Surgery, Obstetrics and Gynecology, Pediatrics, Anesthesiology, Physical Medicine and Rehabilitation, Internal Medicine, and Surgery. The highest prevalence of usage was in Obstetrics and Gynecology, followed by Family Medicine, Psychiatry and Neurology, Pediatrics, Otolaryngology, Anesthesiology, Internal Medicine, Physical Medicine and Rehabilitation, and Surgery. These ndings were useful in terms of improving care plan, decision-making processes, and communication in terms of CAM between the doctors and the patients.
All of the medical specialties mentioned above had a higher prevalence of acceptance than the prevalence of CAM use, except for Obstetrics and Gynecology because the gynecologic oncologists have used CAM to treat a large number of breast cancer patients [12]. There was a small difference in the prevalence (<5%) between the acceptance and the usage in Family Medicine (4%), Obstetrics and Gynecology (4%), Internal Medicine (3%), and Surgery (1%).
A highest difference of prevalence of CAM acceptance and usage was in the eld of Physical Medicine and Rehabilitation (19%). This difference may be due to the reduction in the use of acupuncture in the academic hospitals [13] as well as personal use. Nearly two-thirds of the rehabilitation physicians advised against the use of CAM as a therapeutic option [14]. The lowest prevalence of acceptance and usage of CAM was observed in Surgery. This relatively low prevalence compared to other medical specialties may be due to the belief that CAM products were ineffective. Many surgeons lacked information regarding CAM usage.
From the meta-analysis, it showed that the acceptance of CAM was neutral in European region, and region of the Americas. The World Health Organization reported that the prevalence of CAM usage in the European region, region of the Americas, and Western Paci c region in 2018 was 89%, 80%, and 95%, respectively [10], while this review found that the corresponding prevalence was 54%, 59%, and 37%, respectively. The lower prevalence may be from the dominating studies that were conducted before 2010 whereas CAM has used more often after 2010.
The variation of prevalence of CAM used was investigated in relation to the economic level of the countries. There was a higher prevalence of CAM use in the upper-middle-income economies than the high-income economies which may be due to cultural, historical in uences, and implementation of CAM in the national health system as seen in Brazil [15] and Mexico [16].
Our study has some limitations that should be considered when interpreting the ndings. All studies included did not cover some medical specialties that might have different acceptance and usage of CAM. Therefore, the prevalence of acceptance and usage of CAM in these populations need to be con rmed by further studies. The prevalence of acceptance in some specialties like Neurological Surgery, Obstetrics and Gynecology, Otolaryngology, Pediatrics, and Psychiatry and Neurology was reported by a single study, thus limiting the generality of such ndings.
High heterogeneity of acceptance and usage of CAM between medical specialty referred to the variation in professional characteristic and practice, measurement methods, and study questionnaire. Most of the studies were from high-income economic countries. There were no studies from low-middle, and low-income economic countries which is of concern.
We found that no studies compared the relevant demographic characteristics between the responders and non-responders that would increase non-response bias when estimating the prevalence of CAM use. Although most of the studies demonstrated low risk of bias, over 88% of the studies did not use a validated instrument. Finally, the con ict of interest was not declared in more than 80% of the studies which may result in unintentional bias in the collection, analysis, and interpretation of the data. This can consequently lead to claims that the CAM used was bene cial because the researcher and/or entity may have a nancial or management interest in the CAM used.

Declarations
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analysed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interest.

Funding
This study received nancial support from the Ratchadapiseksompotch Fund, Faculty of Medicine, Chulalongkorn University (Grant Number RA62/059), and Department of Thai Traditional and Alternative Medicine, Ministry of Public Health. The sponsors have no involvement in the systematic search, abstract screening, data extraction, or manuscript preparation.

Authors' contributions
Phutrakool P conceptualized and designed the study, collected the data, analyzed and interpreted the data, drafted the article, and nalized the manuscript for submission. Pongpirul K conceptualized and designed the study, collected the data, analyzed and interpreted the data, drafted the article, and nalized the manuscript for submission.