OUP user menu

Opinions of Japanese rheumatology physicians regarding clinical practice guidelines

Takahiro Higashi, Takeo Nakayama, Shunichi Fukuhara, Hisashi Yamanaka, Tsuneyo Mimori, Junnosuke Ryu, Kazuo Yonenobu, Norikazu Murata, Hiroaki Matsuno, Hajime Ishikawa, Takahiro Ochi
DOI: http://dx.doi.org/10.1093/intqhc/mzp060 78-85 First published online: 15 January 2010


Objective. To examine the views of rheumatology physicians concerning clinical practice guidelines in Japan, and changes to them following the dissemination of new guidelines for rheumatoid arthritis (RA) in 2004.

Design. Two cross-sectional questionnaire surveys, the first conducted before publication of new evidence-based RA clinical practice guidelines and the second conducted after implementation.

Setting. Rheumatology-focused practices in Japan.

Participants. A random sample of physicians registered with the Japan Rheumatism Foundation who satisfied the registration criteria with regard to experience with RA care.

Results. The percentage of guideline users increased from 48 to 60% following publication of the new RA guidelines in 2004 (P < 0.01). The majority agreed that clinical practice guidelines support decision-making in practice, although the proportion of supportive responses decreased slightly in the second survey, from 83 to 77% (P < 0.01) for decision-making, while concern about restricting physician autonomy increased from 18 to 22% (P = 0.01). While only 39% of physicians felt that clinical practice guidelines would contribute to malpractice litigation, the proportion of physicians who were concerned that clinical practice guidelines would be used to bring legal action against providers was larger than that who expected they would defend providers (58 vs 30%, P < 0.001).

Conclusions. Clinical practice guidelines are well accepted among Japanese rheumatology physicians, albeit that the proportion decreased slightly after the introduction of new guidelines. One reason for this may be concern about the use of the guidelines in malpractice litigation. To facilitate implementation, trends in physician support for the guidelines should be closely monitored.

  • surveys
  • general methodology
  • guidelines
  • appropriate health care
  • rheumatoid arthritis


Clinical practice guidelines are intended to improve healthcare quality by providing both patients and practitioners with knowledge regarding current standards of care [1]. They typically summarize available evidence and provide specific recommendations in a wide range of clinical circumstances. Successful implementation of clinical practice guidelines enables practices to remain up-to-date and assure physicians that their practices meet current standards.

While intended to aid in practice, however, clinical practice guidelines have also been shown to provoke mixed reactions among practicing physicians [2, 3]. Common criticisms include statements that guidelines oversimplify practice or are ‘too cookbook.’ Some physicians are concerned that the guidelines may be used to bring malpractice action against physicians, while others feel that they threaten physician autonomy and the patient–physician relationship [4].

In 1999, the Japanese government began funding research groups connected with professional societies to support the development of clinical practice guidelines in various clinical areas [2, 3, 5, 6]. The Japan Rheumatism Foundation, an affiliate of the Japanese College of Rheumatology, took this opportunity to revise its RA guidelines, both then and presently the only RA guidelines in Japan developed by a professional RA association. The then-current guidelines, published in 1997, were based on an informal expert consensus rather than a systematic review of evidence. The 2004 revised version was the first RA guidelines in Japan to be formally drafted after a systematic review of the literature and evaluation of evidence [7, 8], and included information on new laboratory tests (e.g. antibodies to citrullinated protein antigen) and recommendations for disease-modifying anti-rheumatic drugs (DMARD) use early in the course of treatment.

Thanks to the Government's efforts [5], the development of clinical practice guidelines proliferated in many areas of medicine around this time. To ensure the implementation of guidelines, however, and the high-quality, evidence-based healthcare system they are designed to promote, it is important to understand how clinical practice guidelines are perceived and accepted in medical community. To examine the views and experiences of practicing physicians, and changes following the publication of new RA clinical practice guidelines, we surveyed practicing physicians registered with the Foundation before and after publication of the guidelines.


Sample subjects

We surveyed a sample of physicians randomly selected from among registered members of the Japan Rheumatism Foundation, an organization which supports rheumatology research and the education of rheumatology professionals and patients. Registered physicians included specialists in the treatment of rheumatic diseases such as rheumatologists and orthopedic surgeons, as well as general internists, among others. To be registered with the Foundation, physicians must have 5 or more years’ experience in rheumatology practice and receive continuing medical education provided by the Foundation. Although the registered physicians are not necessarily rheumatology specialists with formal rheumatology training, they nevertheless provide care to rheumatology patients, and are hereafter referred to as ‘rheumatology physicians.’

Survey instruments

Survey participants received packets of the survey instruments by postal mail in March 2003, before the 2004 publication of the new evidence-based clinical practice guidelines for rheumatoid arthritis (RA) care, and again in March 2006 after their publication [7]. The second survey was mailed out only to those physicians who responded to the first, and included a copy of the new clinical practice guidelines to compensate respondents for their time and effort.

Questions that were common to both surveys inquired into overall attitudes and opinions regarding clinical practice guidelines and experience with RA guidelines, including the use in the respondent's personal practice and unprompted mention by patients. The second survey explored a wider range of perspectives by adding questions concerning the influence of clinical practice guidelines on practice and decision-making; expected usage of the guidelines in malpractice lawsuits; opinion regarding important characteristics of useful clinical practice guidelines and the expected effect of providing guidelines to patients and families. Attitudes toward guidelines and expected usage in malpractice lawsuits were evaluated using a five-level Likert scale of how strongly respondents agreed with the statement listed in the survey, namely strongly agree, somewhat agree, uncertain, somewhat disagree and strongly disagree. Important characteristics of useful clinical practice guidelines and the expected effect of providing guidelines to patients and families were asked by yes/no questions. The individual items are provided in the tables of survey results (Tables 25).

Statistical analysis

Categorical variable distribution was compared using a chi-square test, while continuous variables were compared using Student's t-test. Because the assumption of normal distribution was not satisfied for either age or years in practice, statistical differences were tested using the Kruskal–Wallis test. With regard to general opinions of clinical practice guidelines, responses were collapsed into two (positive and negative) or three categories (positive, uncertain and negative) for the purposes of presentation. Changes in the response to questions common to the first and second surveys were compared using statistical tests, while responses to questions listed only in the second survey were simply described. All analyses were performed using STATA version 10.1 (Stata Corp., College Station, TX, USA). The study protocol was approved by the Institutional Review Board of Kyoto University School of Medicine and Public Health.


Respondent characteristics

Of the 2000 physicians initially approached, 1117 responded to the first survey, of whom 682 also responded to the second. Table 1 lists the characteristics of survey respondents. Mean respondent age was 50 years, and most were male (95%). Participants specialized in various fields, with about two-thirds being orthopedic surgeons. The only statistically significant difference observed between those responding to both surveys versus those responding only to the first was that the non-responders tended to treat fewer RA patients in their practice (P = 0.04).

View this table:
Table 1

Characteristics of respondents and non-respondents to the 2006 survey among respondents to the first survey (n = 1117)

Respondents (n = 682)Non-respondents (n = 435)P valuea
Age in years, mean (SD)b50 (10)50 (10)0.83
Male, %b95950.86
Years in practice, mean (SD)b25 (10)26 (11)0.98
East Japan residents, %b41370.22
Practice area, %
 No answer10
Specialty, %
 Orthopedic Surgeon70670.96
 Internal Medicine2726
 No answer37
Practice size, % (n of continuous RA patients, %)
 No answer11
  • aChi-square test for categorical variables and Kruskal–Wallis test for continuous variables (due to violation of normality assumptions) for persons who provided answers. bSeven, 5, 10 and 3 respondents did not provide age, gender, years in practice or practice prefecture, respectively.

Experience with clinical practice guidelines

Table 2 describes respondents’ experience with clinical practice guidelines. The proportion of rheumatology physicians who reported at least occasionally using the Japanese RA guidelines (always/three or more times per month or occasionally/about once a month) increased from 48 to 60% following publication of the new clinical practice guidelines. The percentage who used foreign rheumatology guidelines (use the original/familiar through Japanese review articles) did not change (33–32%), but use through reviews and sources other than the original rose from 18 to 21% (overall P < 0.01). The proportion of respondents who felt that the Japanese RA guidelines were very or sometimes useful increased from 62 to 69% (P < 0.01).

View this table:
Table 2

Experience and opinions regarding RA guidelines (n = 682)

Survey 2003Survey 2006P valuea
Use of Japanese RA guidelines, %b
 Always or three or more times per month1021<0.01
 Occasionally or about once a month3839
 Aware of the guidelines but never used1416
 Not aware of the guidelines52
 No response21
How useful are Japanese RA guidelines?, %
 Very useful1419<0.01
 Sometimes useful4850
 Seldom or never useful55
 Don't use1810
 No response20
Use of guidelines from other countries, %
 Use the original1511<0.01
 Familiar through Japanese review articles1821
 Wish to use, but never used3023
 Irrelevant to practice in Japan811
 Never thought about usefulness2430
 No response54
How useful are RA guidelines published in other countries?, %b
 Very useful13120.97
 Sometimes useful4244
 Not useful22
 No response139
Do patients mention guidelines?, %c
 Once in a while7
 No response3
  • aChi-square test for persons who provided answers. bAsked only to those who used clinical practice guidelines from other countries. cQuestion added in the second survey.

General attitude toward clinical practice guidelines

General attitudes toward clinical practice guidelines are presented in Table 3. Most respondents stated support for clinical practice guidelines, with 83 and 77% in the 2003 and 2006 surveys agreeing that they support decision-making in practice, and 91 and 89% agreeing that they are useful for medical education, respectively. A minority of respondents in both surveys agreed with negative sentiments that clinical practice guidelines restrict physician autonomy and oversimplify clinical medicine (32 and 35% in each survey). In questions appearing only in the 2006 survey, guidelines were generally believed to aid in sharing information with patients (56%), while a larger proportion felt that they should be available to the general public (46 agreed versus 15% disagreed, P < 0.01). Although general attitudes to clinical practice guidelines did not notably change before and after publication of the new clinical practice guidelines, the prevalence of positive opinions decreased slightly (e.g. support decision-making, 83–77%, P < 0.01) while that of negative opinions increased slightly (e.g. restrict physician autonomy, 18–22%, P < 0.01).

View this table:
Table 3

Physician opinions of clinical practice guidelines (n = 682)

Survey year20032006P valueb
Strongly agree/agree, %aUncertain, %aDisagree/strongly disagree, %aStrongly agree/agree, %aUncertain, %aDisagree/strongly disagree, %a
General attitudes toward practice guidelines
 Guidelines support decision-making in practice8314277184<0.01
 Guidelines support medical education9162891490.19
 Guidelines restrict physician autonomy1831502236420.01
 Guidelines aim for cost containment224631144738<0.01
 Guidelines oversimplify clinical medicine3235323535290.48
 Guidelines should be promoted in various clinical areas60328513611<0.01
 Guidelines facilitate sharing information with patients563211
 Guideline contents should be available to the general public463815
 Experts should know the guideline content9451
 Guideline adherence is an important indicator of quality513712
 Justification should be documented when not following recommendations373725
Perception toward guidelines in the context of malpractice litigation
 Publication of guidelines will lead to an increase in malpractice lawsuits394216
 Guidelines will be used to bring legal action against medical providers58319
 Guidelines will be used to defend medical providers against legal action304721
  • aNon-responses were less than 3%, and not included in the percentage. bComparing distribution pre- and post-publication of the new RA guidelines.

The 2006 survey contained several questions about guidelines and quality of care. Although 94% of respondents agreed that experts should know the contents of clinical practice guidelines, only 51% supported the idea that actual adherence to guidelines is an important indicator of quality. Further, only 37% agreed that a reason should be documented in situations in which guideline recommendations are not followed, although this percentage was larger than the 25% who actively disagreed with the necessity of documentation.

The lower part of Table 3 highlights the physician views from the 2006 survey regarding the influence of clinical practice guidelines on malpractice lawsuits. Thirty-nine percent of respondents felt that clinical practice guideline publication might lead to an increase in malpractice lawsuits. Further, more respondents (58%) felt that clinical practice guidelines would be used to bring legal action against physicians than reported that guidelines would be used to defend them against such actions (30%).

Clinical practice guidelines and patients

The 2006 survey contained questions about patients and guidelines. The results showed that patients rarely mention clinical practice guidelines. Most respondents (about 90%) reported that patients seldom or never mentioned clinical practice guidelines (Table 2). More than half felt positively about providing guidelines to patients and families, considering that they would help them (Table 4). A minority were concerned about patient ability to understand the guidelines.

View this table:
Table 4

Agreement with statements about effects of guidelines on patients (2006 survey, n = 682)

StatementAgree, %
Guidelines will help the process of informed choice53
Guidelines will help provide knowledge to patients and families55
Additional materials besides guidelines are necessary for full understanding39
Guidelines will confuse patients and families24

Opinions regarding important characteristics of clinical practice guidelines

Table 5 details characteristics that respondents felt were important to include in clinical practice guidelines, based on questions in the 2006 survey only. Eighty percent felt that clearly stating the reasons for guideline recommendations was important, while a somewhat smaller percentage (61%) felt that discussing the validity of clinical evidence was important. Fewer respondents placed import on aspects of the development process, including the specification of aims (48%) and presentation of the results of external review (42%). Only 20% of respondents felt it was important to state conflicts of interest.

View this table:
Table 5

Physician opinions regarding important characteristics of guidelines (2006 survey, n = 682)

FeatureAgree, %
Reasons for recommendations are clearly stated81
Discussion of validity of evidence in the literature61
Guideline aims are clearly specified48
External review results are presented42
Developed by trustworthy organization or persons38
Uniform format for each recommendation29
Conflict of interests of the persons involved are specified20

The 2006 survey also inquired into awareness of the formal evaluation instrument developed by the Appraisal of Guidelines for Research and Evaluation (AGREE) collaboration [9]. Only 10% of the respondents were aware of the contents of this instrument, while 32% reported that they had at least heard of the name. More than half, however, had never heard of the instrument (55%).


This study found that the new RA clinical practice guidelines were generally accepted by rheumatology physicians in Japan, with about 60% reporting that they used them and a majority reporting that they were useful. The proportions of both increased after publication. A majority also felt that clinical practice guidelines in general support clinical practice and are useful in medical education, and that experts should know the contents of guidelines. Most also supported the development of clinical practice guidelines in clinical areas outside of RA. Overall, the survey revealed that the number of physicians who at least occasionally use the Japanese RA guidelines has increased, providing further evidence to indicate support for guideline use.

Our results showed that Japanese rheumatology physicians have a comparable or greater support for clinical practice guidelines than the average level reported in a systematic review by Farquhar et al. [4] of surveys conducted in 1990–2000. They found that the percentages of respondents who regarded clinical practice guidelines as helpful sources of advice and as good educational tools were 75 and 71%, respectively, versus 77 and 89% in our 2006 survey. Further, the negative views represented by the proportion of respondents who felt that clinical practice guidelines restrict physician autonomy or oversimplify clinical medicine were comparable or smaller than the average in Farquhar's review (22 and 35%, respectively, in our 2006 survey, versus 34% overall in the reviewed surveys). The large number of positive opinions in the present study may be attributed to underlying trust with the RA guidelines, which were issued by the Japan Rheumatism Foundation, to which the all respondents belonged. In addition, the studies included in Farquhar's review were all conducted in Western countries and none were focused on rheumatology physicians. Future research may need to address how these differences in settings and samples lead to differences in attitudes toward guidelines. Another explanation may be that the study was conducted during a period in which clinical practice guidelines proliferated in many clinical areas in Japan, fueled by a government initiative [5]. Guideline support has hopefully contributed to support for the periodic revisions which are generally necessary every 3 to 5 years [10].

The minority of non-support for the guidelines is not attributable to respondent preference for western over Japanese guidelines. Since both the Japanese and international guidelines base their recommendations on a similar set of evidence found on extensive literature review [7, 8], there was no major discrepancy between them in direction. Nevertheless, the Japanese guidelines clearly reflect particular characteristics of the Japanese healthcare system and practice environment. An example is that Japanese national insurance does not cover methotrexate as a first-line drug for RA, but rather only as second-line or later drug. The Japanese guidelines acknowledge this restriction in the section for DMARDs. They also mention domestically developed DMARDs such as bucillamine, actarit, mizolibine and lobenzarit, and acknowledge that evidential support of efficacy from large-scale randomized controlled trials is lacking.

Of some concern is the fact that the number of rheumatology physicians who expressed positive views towards clinical practice guidelines decreased over time, while those expressing negative views increased. Between the first and second surveys, for example, the percentage of respondents who felt that guidelines support decision-making in practice decreased from 83 to 77%, while the percentage who felt that they restrict physician autonomy increased from 18 to 22%. One explanation may be the dramatic change in rheumatology practice occurring between publication of the revised guidelines and the second survey with the advent of biological agents such as infliximab. This may have made many physicians feel that the guideline was already outdated at the time of the second survey. Another explanation is that these small changes simply reflect over-expectations before the arrival of the new guidelines, or regression to the mean if those who had more positive attitudes than the average practicing physician were more likely to have responded to the first survey. Future research should closely monitor clinician opinions regarding clinical practice guidelines.

Reports of patient use or mention of clinical practice guidelines were rare, possibly due to the unavailability of patient-use Japanese RA guidelines at the time of the survey. The first patient-use guidelines for RA were published in June 2006, immediately after the survey was completed, and the number of patients who read the guidelines is therefore expected to increase. Physicians generally expect that clinical practice guidelines will exert a beneficial influence on patients and families. Physician experience with patient mention of guidelines during clinical visits bears further investigation.

Although only a minority of respondents felt that clinical practice guidelines restrict physician decision-making, a substantial proportion expressed concern regarding malpractice lawsuits. Further, many physicians felt that clinical practice guidelines might be more often used as a basis for bringing legal action for malpractice than as a defense against such action, despite the explicit assertion within the new Japanese RA guidelines that they should not be used to set standards for malpractice lawsuits [7, 8]. This concern is compounded by a somewhat outdated study regarding medical malpractice insurance claims from the USA, which reported that clinical practice guidelines are more often used for inculpatory rather than exculpatory purposes [11]. Moreover, while Japan entertains relatively few malpractice lawsuits, with an annual filing of 944 in 2007, this number was nevertheless rising rapidly until 2005, after which it decreased slightly [12, 13]. Since legal pressure may be more likely to promote defensive practice and increase malpractice insurance premiums than improve quality of care [14], implementation of mechanisms to prevent the abuse of clinical practice guidelines may be necessary.

Overall, most clinicians appeared to be unfamiliar with the AGREE criteria [9], which focus on the methodological appraisal of clinical practice guideline quality. Further, some aspects of the criteria set did not seem important to many clinicians. Clinicians naturally placed emphasis on the logic which underpinned recommendations, such as adequate search for evidence and a clear statement of reasoning. Only a minority of physicians, however, regarded statements of conflict of interest and external review as important, both of which are included in the AGREE criteria. This may be because the use of clinical practice guidelines is not particularly common in Japan, and thus many clinicians have never experienced obviously biased recommendations due to these factors, or been forced to choose among multiple clinical practice guidelines for the same area. Further, given that guidelines are meant to facilitate decision-making rather than dictate clinical practice, decision-makers must judge for themselves the validity of any guideline-directed recommendation. A conflict of interest statement itself does not necessarily guarantee or provide evidence of the validity of an individual recommendation. Given the variation in perceived importance across factors, the exploration of weighting systems in AGREE instruments may be of interest, perhaps based on user opinion.

The present study should be interpreted within several limitations. First, the response rate was not high. Physicians who support clinical practice guidelines may have been more likely to respond than those who do not, resulting in overstatement of support for guidelines. Although we have no information on non-respondents to the first survey, respondents to the second did not substantially differ from those responding to the first only. Second, our survey sample was rheumatology physicians, who may differ from physicians in other specialties. The generalizability of the results should be interpreted within this sampling framework. Finally, positive views of clinical practice guidelines do not necessarily lead to the use of such guidelines, and self-reporting of use may not be accurate. A more accurate understanding of guideline use and adherence would require a review of medical records or insurance claims, or the direct observation of clinical encounters.

In conclusion, our survey revealed that clinical practice guidelines are generally accepted by Japanese RA physicians. This overall positive view is supportive of the implementation of the clinical practice guidelines in practice. While placing little attention on some procedural aspects of guideline development, such as conflict of interest and external review, clinicians do seem to consider the points necessary to evaluate recommendations on a case-by-case basis. Some decline in general support for guidelines was observed, possibly due to concerns about their use in malpractice litigation. Close monitoring of trends in guideline support should facilitate implementation in actual practice.


This study was supported in part by a grant-in-aid for health technology assessment from the Ministry of Health, Labor and Welfare, Japan.


View Abstract