Effective Clinical Practice
Effective Clinical Practice, September/October 2001
Christel Mottur-Pilson, Phd, Vincenza Snow, MD, Kyle Bartlett, PhD
For author affiliations, current addresses, and contributions, see end of text.
Context. Substantial effort has been devoted to improving physician compliance with evidence-based guidelines.
Objective. To explore physicians' reasons for not following so-called "best practices" in caring for patients with type 2 diabetes.
Design. Descriptive study of self-assessed compliance with five measures of performance.
Participants. Eighty-five internists who volunteered to participate in a practice-based reasearch network created to improve clinical practice.
Data Collection. Physicians reviewed their own charts of patients with type 2 diabetes mellitus (1755 patient encounters) to assess compliance and offered open-ended comments concerning their reasons for not complying with "best practices."
Results. The physician volunteers reported not complying with the annual foot examination in 13% of encounters. A similar level of noncompliance was reported for the annual lipid profile (15%) and retinal examination (17%). Among the five measures examined, noncompliance was most common for screening urinalysis (26%) and screening microalbuminuria (46%). The physicians' open-ended comments suggested that physician oversight, patient nonadherence, and systems issues were common reasons for noncompliance. However, noncompliance also resulted from a conscious decision by the physician, as indicated by comments about patient age and comorbid illness or, with nephropathy screening, established renal disease or current therapy with angiotensin-converting enzyme inhibitors.
Conclusions. Even among a self-selected group of physicians, noncompliance with best practices in diabetes is common. Although physician forgetfulness and external factors are frequently offered as reasons for noncompliance, it may also result from a conscious decision, as physicians may disagree about what constitutes "best practices."
Physicians are under increasing pressure to comply with evidence-based guidelines and to achieve high "scores" for performance measures. Nevertheless, compliance with so-called "best practices" is far from ideal. (1-3) The barriers to guideline compliance have been categorized into three domains: physician knowledge (lack of awareness, lack of familiarity, or oversight), physician attitudes (lack of agreement; lack of self-efficacy—that is, the belief that a physician can perform guideline recommendations; lack of outcome expectancy—that is, skepticism that complying with the guideline would help patients; or the inertia of previous practice), and external barriers. (4) In most previous studies on physician compliance, the physician participants answered closed-ended survey questions, and this limited the scope of their responses to those hypothesized by the investigators. (4) Other methods used in the literature include focus groups, (5, 6) interview, (7) retrospective review of administrative data, (8, 9) and hypothetical clinical scenarios. (10, 11) But these previous studies focused on general barriers rather than on patient-specific (and encounter-specific) reasons for not following guidelines.
The two studies that did attempt to elicit clinicians' reasons for noncompliance in specific patient encounters involved acute care/hospital settings. In the first study, a retrospective chart review determined why low-risk patients admitted for chest pain were not discharged according to a local practice guideline. (12) The second report (13) prospectively evaluated compliance with an emergency department guideline for identifying patients at low risk for pneumonia. In both studies, important comorbid conditions, older age, and systems issues were among the most common reasons cited for noncompliance.
In this paper, we report on a group of physicians who volunteered to try to improve their own outpatient practice in patients with type 2 diabetes. Physicians generally reviewed their own charts and, in selected cases, offered their own explanations of why they had not followed the best practice. To our knowledge, no study analogous to ours has been reported in the ambulatory care setting.
The 85 internists who volunteered for this study are members of a practice-based research network called "QNet," which is short for Quality Network. The American College of Physicians-American Society of Internal Medicine (ACP-ASIM) created QNet to raise physician awareness of best practices and to provide a glimpse into the constraints and reality of everyday office-based care. To gauge physician compliance, we focused on five common conditions with substantial scientific evidence for disease management: diabetes, cholesterol screening, warfarin therapy, asthma, and congestive heart failure.
To recruit physician participants, we sent 7000 invitations to ACP-ASIM members. These 7000 names were selected from the former ASIM Practice Guidelines Network databases (n = 1000), a randomly generated list from ACP-ASIM membership rosters (n = 3000), and ACP-ASIM's Community Based Teaching Project's database (n = 3000). Of the 7000 physicians invited, 800 expressed an interest in participating, and 137 completed the project. Each participating internist chose one module among the five offered. Table 1 reports data from the 85 internists who selected the diabetes module.
The internists were asked to prospectively fill out a questionnaire at the point of service for 25 patients with type 2 diabetes mellitus. Physicians received personalized reports depicting their results in comparison with those of their peers.
Best Practices in Diabetes Survey
The survey form included questions on whether the physician had met six performance measures for type 2 diabetes. Five questions addressed whether the patient had received one of the following tests in the previous 12 months: dilated eye examination, lipid profile, comprehensive foot examination, urinalysis testing, or microalbuminuria testing.
The final question, regarding glycosylated hemoglobin, asked, "Within the past 12 months, in which of the following quarters was this patient seen at least once? (make sure you identify both the quarter and the year). In addition, indicate quarters when this patient had at least one glycosylated hemoglobin test and the highest value for the quarter." Because this question resulted in ambiguous responses (i.e., physicians did not clearly distinguish between year and quarter visits, resulting in ambiguous assignments of glycosylated hemoglobin values), we discarded data for this sixth performance measure.
Finally, in cases where physicians did not follow "best practices," they had the option to provide an open-ended reason.
Analysis of Comments
The participants reported data on 1755 diabetic patient encounters. Since each questionnaire contained five usable performance measures, physicians had the opportunity to measure compliance with 8775 best practices. For each best practice, we determined the total number of self-reports of noncompliance. Among the 2073 instances when best practices were not followed, 505 (24%) were accompanied by comments, 335 of which explained why one of the five best practices was not achieved. Review of the physicians' reasons for noncompliance patterns showed that the reasons fell into four distinct categories.
Conscious Decision. In some instances, the comments suggested that the internist made a conscious decision not to follow a "best practice." A common rationale was that the patient had more pressing medical problems. Another was that the best practice did not apply to the patient either because of advanced age or established disease (e.g., blindness or renal failure).
Patient Nonadherence. In other instances, physician comments indicated that the patient declined to follow the indicated practice (e.g., failure to modify lifestyle, unwillingness to visit an ophthalmologist, or missing an appointment).
Systems Issues. There were other instances in which the physician comment indicated some discontinuity of care not clearly attributable to either the physician or patient (e.g., lack of communication between providers caring for the same patient, seasonal change in patient residence, or inadequate insurance coverage).
Oversight. The oversight category comprises instances in which physicians acknowledged forgetting to apply the best practice.
Analysis of the physicians' comments was performed by using a qualitative software package (NVivo Nudist QSR International, Victoria, Australia).
The self-assessed performances of the 85 internists revealed an overall noncompliance rate of 24% (2073 of 8775 instances). Figure 1 shows noncompliance for each of the five diabetes care measures. Noncompliance ranged from 46% of all encounters for annual screening microalbuminuria to 13% for annual foot examination.
We grouped physician reasons for noncompliance into the four categories listed in Table 2. This table also provides selected examples of comments in each category. Figure 2 shows how the pattern of explanation according to these four categories varied for specific performance measures.
For annual screening microalbuminuria and urinalysis, most comments suggested that the physician had made a conscious decision not to comply with the best practice. In many cases, physicians reported that patients already had gross proteinuria or end-stage renal disease. Moreover, in more than one third of the comments on patient encounters involving failure to perform annual microalbuminuria screening, physicians reported that the patient was already receiving treatment with angiotensin-converting enzyme (ACE) inhibitors and that test results would not change their management decisions.
For the annual dilated eye examination, patient nonadherence and systems issues were the most common reasons given for noncompliance. Patients would not make or keep appointments with the ophthalmologists; ophthalmologists in turn would not send their examination results to the primary care provider. The most frequently offered reason for not performing the annual lipid profile and foot examination was physician oversight.
Because the physicians provided explanatory remarks, we gained insight into clinicians' patient-specific reasons for providing care that diverged from best practices. Although the reason most often discussed in the literature has been physician oversight, we found that a substantial amount of noncompliance was attributable to other factors, such as systems issues and patient nonadherence. Furthermore, our research clearly indicates that physicians occasionally made a conscious decision not to comply with a best practice. This raises questions about the appropriateness of the term best practice.
In our study, this was particularly true for nephropathy screening. Many physicians raised valid questions about the utility of performing microalbuminuria or urinalysis in patients with end-stage renal disease. Some physicians also judged annual screening for microalbuminuria to be inappropriate when their patients were already receiving ACE inhibitor therapy because patient management would remain the same regardless of the test result. Conscious decisions not to comply with screening were also reported in the care of patients with important comorbid conditions or limited life expectancy.
Our study has several limitations. It involved only members of the QNet, a group of volunteer internists interested in quality improvement; therefore, the results are not generalizable to all internists or physicians. Because the data reflect only the self-reports of volunteers, our noncompliance estimates are probably lower than the actual figures in general practice. Furthermore, because the physicians chose when to provide comments, we believe that the physician participants were more likely to write a comment if they believed noncompliance was justified; therefore, we suspect that we have underestimated the role of physician oversight as an explanation. Despite these caveats, our study offers a brief account of the various limitations physicians encounter in their day-to-day practice.
In summary, QNet participants have provided some insights into why physicians do not always comply with best practices. Our data suggest that failure to follow guidelines is not necessarily explained by "bad doctors" or forgetfulness; rather, noncompliance may reflect valid questions about the usefulness and applicability of a best practice to an individual patient. As Vijan pointed out, "Not all patients are the same. Treating them as such not only minimizes autonomy but is also a recipe for inefficiency." (14) The difficult task for the future will be to determine how such factors as comorbid illness, age, and patient wishes can be incorporated into performance measures to more accurately reflect the intricacies of quality care in clinical practice.
|Take Home Points
1. McLaughlin TJ, Soumerai SB, Willison D, et al. Adherence to national guidelines for drug treatment of suspected acute myocardial infarction: evidence for undertreatment in women and the elderly. Arch Intern Med. 1996; 156:799-805.
2. European Secondary Prevention Study Group. Translation of clinical trials into practice: a European population-based study of the use of thrombolysis for acute myocardial infarction. Lancet. 1996; 347:1203-7.
3. Antman EM, Lau J, Kupelnick B. A comparison of results of meta-analyses of randomized control trials and recommendations of clinical experts: treatments for myocardial infarction. JAMA. 1992; 268:240-8.
4. Cabana MD, Rand CS, Powe NR, et al. Why don't physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999; 282:1458-65.
5. Jackson L, Yuan L. Family physicians managing tuberculosis. Qualitative study of overcoming barriers. Can Fam Physician. 1997; 43:649-55.
6. Cabana MD, Ebel BE, Cooper-Patrick L, et al. Barriers pediatricians face when using asthma practice guidelines. Arch Pediatr Adolesc Med. 2000; 154:685-93.
7. McIlvain HE, Crabtree BF, Gilbert C, et al. Current trends in tobacco prevention and cessation in Nebraska physicians' offices. J Fam Pract. 1997; 44:193-202.
8. Mukamel DB, Bresnick GH, Wang Q, et al. Barriers to compliance with screening guidelines for diabetic retinopathy. Ophthalmic Epidemiol. 1999; 6:61-72.
9. Freeborn DK, Shye D, Mullooly JP, et al. Primary care physicians' use of lumbar spine imaging tests: effects of guidelines and practice pattern feedback. J Gen Intern Med. 1997;12: 619-25.
10. James PA, Cowan TM, Graham RP. Patient-centered clinical decisions and their impact on physician adherence to clinical guidelines. J Fam Pract. 1998; 46:311-8.
11. Katz DA. Barriers between guidelines and improved patient care: an analysis of AHCPR's unstable angina clinical practice guideline. Health Serv Res. 1999; 34:377-89.
12. Ellrodt AG, Conner L, Reidinger M, et al. Measuring and improving physician compliance with clinical practice guidelines: a controlled interventional trial. Ann Intern Med. 1995;122:277-82.
13. Halm EA, Atlas SJ, Borowsky LH, et al. Understanding physician adherence with a pneumonia practice guideline. Arch Intern Med. 2000;160:98-104.
14. Vijan S. Are we overvaluing performance measures? Eff Clin Pract. 2000; 3:247-9.
The authors thank William E. Golden, MD, for his clinical acumen in helping us select appropriate module content. We also thank Steve Lascher, DVM, MPH, for his careful structure and supervision of the quantitative data elements of QNet; Jolene Chou for her untiring attention to data entry and analysis; and David Hanna for the structure of the qualitative data.
This study was made possible by an unrestricted educational grant from Aventis Pharmaceuticals.
Christel Mottur-Pilson, PhD, Director, Scientific Policy, American College of Physicians-American Society of Internal Medicine, 190 North Independence Mall West, Philadelphia, PA 19106; telephone: 215-351-2840; fax: 215-351-2594; email: email@example.com g.