Pediatricians' Attitudes, Beliefs, and Practices Regarding Clinical Practice Guidelines: A National Survey

COMMENT

Most general pediatricians are familiar with CPGs and use them at least in part. We found, however, that no specific CPGs, with the exception of those for asthma, are used by >27% of general pediatricians. This discrepancy in part may be attributable to lack of familiarity with CPGs for certain commonly-encountered conditions. Fever and gastroenteritis, for example, are frequent diagnoses in pediatric practices, but only 40% of pediatricians said they were familiar with fever CPGs, and 34% with gastroenteritis CPGs. Practical solutions for enhancing familiarity with CPGs were offered by study participants; most overwhelmingly cited mailings and publication in journals, and a few suggested newer alternatives such as the Internet.

Pediatricians may avoid using most pediatric CPGs because few have been subjected to the rigorous clinical evaluation necessary to demonstrate improved quality and reduced costs. In contrast, recent evaluations of adult CPGs have shown that they can reduce the length of hospital stays for gastrointestinal tract hemorrhage,² and decrease radiography, waiting times, and costs for acute ankle injuries.^{3 ,8} Clinical testing of CPGs is important because, even for CPGs based on extensive scientific evidence, use in actual practice may not improve patient care. Problems with external validity, misapplication, co-morbidity, and lack of acceptance may thwart the clinical effectiveness of seemingly reasonable CPGs.⁹ It has been argued that CPGs—like any new medical intervention—should be carefully evaluated for risks, benefits, feasibility, and cost.⁹ Several authors have cogently stated that CPGs should not only be outcome-based but outcome-justified.^{5 ,9 ,10} Pediatricians in our study echoed this sentiment, reporting that they would be more likely to follow CPGs if they were evidence-based and had been shown to improve outcomes. Similarly, a study of Australian general practitioners revealed that the most important factor in deciding to follow a guideline was whether it was based on evidence.¹¹

The diversity of CPGs used for a given condition was surprising. In the case of asthma, for example, pediatricians reported using 10 CPGs, including those from national organizations, regional quality improvement organizations, local institutions, publications, and clinicians' own personal CPGs. This was unexpected, given the wide dissemination of both the original and revised versions of the National Heart, Lung, and Blood Institute's guidelines for the diagnosis and management of asthma.^{12 ,13} It is also noteworthy that none of the respondents reported using the AAP's published practice parameter for the management of acute asthma exacerbations in children, even though all respondents are members of the AAP.¹⁴ The reasons for pediatricians' use of multiple alternatives to such widely-publicized CPGs are not clear. Consistent with what study participants identified as general CPG problems and reasons for nonuse, the aforementioned national asthma CPGs may be perceived by pediatricians as not adequately subjected to clinical testing, not sufficiently allowing for clinical judgment, or differing little from actual clinical practice. Additional research might elucidate why clinicians choose to use such a diversity of alternative CPGs instead of well-publicized national CPGs for a given condition.

Four published studies have examined clinician attitudes and beliefs regarding CPGS,^{4 ,11 ,15 ,16} but none focused on the prevalence of CPG use, or practitioners' reasons for use/nonuse. Surveys of Australian general practitioners¹¹ and American internists¹⁶ confirm our finding that many clinicians view CPGs as good educational tools. Consistent with our data, objections to CPGs being overly rigid and “cookbook medicine” were frequent among American clinicians^{4 ,16} but less common among Australian general practitioners.¹¹ Two studies^{11 ,15}also supported our finding that clinicians are quite concerned about the use and abuse of CPGs in medicolegal contexts. As in our study, Australian general practitioners identified scientific evidence of efficacy as one of the most important factors in choosing to follow CPGs.¹¹

Several limitations of this study should be noted. Although the survey response rate was comparable to other national postal surveys, nonresponse bias may have influenced the results. A skewed sample consisting of respondents who view CPGs either more or less favorably than the national average may have resulted. To ensure confidentiality for participants, we did not mark surveys with identifiers. Although this may have yielded more frank responses, it precluded comparisons between respondents and nonrespondents regarding personal and practice characteristics. Participant comments, however, suggest that the extremes of favorable and unfavorable views toward CPGs were captured and that a reasonable national sample of general pediatricians was obtained. A second potential limitation is that self-reported use of CPGs may not accurately reflect actual use. For example, as many as 8% of internists said that they were familiar with a fictitious CPG for computed tomography of the head,¹⁶ and 49% of Australian general practitioners reported that a fictitious CPG for the diagnosis and management of pelvic inflammatory disease was minimally to extremely useful.¹¹ for self-reported use of CPGs by pediatricians in our study, therefore, might overestimate the proportion that actually use CPGs. Although this is the first study that we are aware of to report partial use of CPGs by clinicians, another limitation is that we were unable to explore what was meant by use of a CPG “in part.” CPGs can contain information on diagnosis, therapy, follow-up, and prognosis, and it would be of interest to identify which component is most often selected by those who use CPGs only partially. A final potential limitation is the exclusion of subspecialist responses, leading to a sample that may not reflect the attitudes and practices of pediatricians certified by the American Board of Pediatrics. The study goal, however, was to examine the attitudes and practices of generalist pediatricians; future research might investigate whether there are important differences in the adoption of CPGs between generalists and subspecialists.

In this study, we did not define what constitutes a CPG. We surveyed pediatricians about well-known CPGs, as well as other statements sometimes referred to as guidelines, such as the health supervision guidelines of the AAP.¹⁷ The term guideline has been avoided by the AAP, which labels guides to practice as practice parameters. Several methodological problems in CPG development have been identified,⁵ including what are appropriate sources, whether consensus statements should be considered CPGs, and what represents adequate scientific evidence for inclusion in a CPG. Clinician confusion and lack of CPG acceptance may be an inevitable response to inconsistent terminology and the absence of an accepted definition for CPGs.

CPG use is significantly more likely among general pediatricians in HMOs and those who are non-white, but significantly less likely among those seeing greater numbers of patients. Because many HMOs routinely encourage and internally publicize the use of CPGs in daily practice, it is not surprising that HMO pediatricians are more likely to use CPGs. It is not clear why non-white ethnicity is independently associated with greater odds of CPG use. Pediatricians' responses to the question: “What would make you most likely to follow practice guidelines?” (Table 5) may provide insight on why those with the highest weekly patient visits are least likely to use CPGs. For the busy clinician, it would seem that CPG features of particular importance would include simplicity, feasibility, demonstrated efficacy, logic, appropriateness, and accessibility. Lower use among busier practitioners may, therefore, reflect the dearth of scientific evidence on the efficacy of most pediatric CPGs, and perceived complexity and difficulty in using certain more detailed CPGs.

Responses of this national sample of general pediatricians have implications for improving CPG quality and acceptance. Participants repeatedly emphasized that a CPG is most likely to be used when it is simple, feasible, allows for clinical judgment, does not impede autonomy, and has been shown to be effective. Most agreed that CPGs should not be used for medical liability or disciplinary actions, and should not be motivated by desires to control costs. The data suggest that American pediatricians believe that CPGs can improve patient care, but only when the CPGs are flexible, rigorously tested, not used punitively, and are motivated by desires to improve quality, not reduce costs.

ACKNOWLEDGMENTS

This work was supported by Grant 030878 from the Robert Wood Johnson Minority Medical Faculty Development Program, an Institutional Research Training Grant of the Health Resources Services Administration, and the David and Lucile Packard Foundation.