Clinical practice guidelines are intended to improve the consistency and quality of medical care through improving the incorporation and prompt translation of evidence from research to the bedside.1,2 In recent years, as the evidence base has grown more complex, guidelines have proliferated,1,3,4 including those specifically pertaining to cancer management.5
Scholars have devoted considerable attention to defining the optimal processes for guideline development and have investigated the impact of guidelines, with many concluding that their use is associated with improved processes of care, clinical outcomes, and patient satisfaction.6-10 Research has also considered physicians’ attitudes, which may differ according to specialty and practice setting.11,12
Although studies have shown that physicians generally find clinical practice guidelines helpful, they have also revealed several perceived limitations of guidelines that may affect adherence.5,13-15 In a comprehensive analysis of barriers to physician adherence to guidelines, Cabana et al16 identified 7 general categories of barriers that have been explored in previous studies, including lack of familiarity, lack of awareness, lack of agreement with guidelines, lack of outcome expectancy, lack of self-efficacy (physicians’ perceptions that they cannot perform the recommended behavior), lack of motivation, and external barriers.
Little recent information exists regarding physicians’ use of cancer management guidelines5 or about the barriers to guideline use in cancer care specifically. To address these gaps in the existing literature, the authors conducted a nationwide survey of medical oncologists and surgeons who had seen at least one new patient with breast cancer in the past year. They chose to study these providers because breast cancer management is inherently multidisciplinary and based on a complex evidence base that lends itself particularly well to guideline development. The authors sought to address 3 areas of question: 1) what are physician attitudes toward cancer management guidelines in general; 2) how familiar with existing guidelines are physicians who treat patients with breast cancer, how do they use these guidelines, and how often do they believe their recommendations are guideline-compliant; and 3) do attitudes, familiarity, use, or compliance with guidelines differ by specialty or other physician characteristics?
Methods
This study was approved by the University of Michigan Institutional Review Board.
Population Sampling and Data Collection
The study sample was drawn from the AMA Physician Masterfile, a relatively comprehensive list of US physicians, assembled based on medical school and residency enrollment and licensing records. A randomly selected sample of physicians was obtained whose specialty was medical oncology or hematology/oncology, or surgery or surgical oncology, along with their contact information. Before survey mailing, these physicians were telephoned to verify addresses and eliminate subjects whose practices clearly excluded breast cancer.
Between January and June 2012, a letter, the questionnaire, and $50 cash were mailed to 750 surgeons (including surgical oncologists) and 750 medical oncologists (including hematologist/oncologists). A modified Dillman approach17 was used, including reminders to nonrespondents, to maximize response rate.
Respondents were asked to indicate whether they had seen at least one new patient with breast cancer in the preceding year and to return the survey without completing it further if they had not. Deidentified survey data were entered into a REDCap database and exported to the SAS System, version 9.2 (Cary, NC) for analysis.
Measures
A 15-page questionnaire was developed after literature review and consideration of Cabana et al’s theoretical model.16 An iterative design process was used that first expanded the list of potential questions and then refined the final set based on face validity, relation to the core concepts, clarity, and readability. Standard techniques of content validation were applied, including systematic review of questions by experts in content matter and survey design, and cognitive pretesting.
General attitudes toward guidelines were evaluated with a battery of 8 items developed by previous scholars in this area.13 Responses were elicited on a 4-point scale ranging from strongly agree to strongly disagree.
Two items measured general communication behaviors when complying with and deviating from guidelines. The first asked, “When you make a treatment recommendation that is consistent with practice guidelines do you routinely tell the patient that the recommendation is in agreement with guidelines?” Respondents chose “yes” or “no.” The second asked, “When you make a treatment recommendation that is inconsistent with practice guidelines, do you routinely tell the patient that the recommendation is not in agreement with guidelines.” Response choices were “yes,” “no,” and “I don’t make guideline-inconsistent recommendations.”
Use of specific cancer management guidelines was evaluated with an item inquiring, “Which of the following guidelines generally influence your cancer management decisions.” Respondents were to check all that applied from response choices of “NCCN (National Comprehensive Cancer Network),” “ASCO (American Society of Clinical Oncology),” “Other (please specify),” and “None.”
The remaining outcomes measures focused on familiarity with and use of the NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) specifically, given prior work suggesting that these were the most commonly used cancer management guidelines.5 Respondents were first asked if they knew that NCCN provides guidelines for the management of patients with cancer before taking the survey. Those who did were then asked several questions exploring their use of the NCCN Guidelines for Breast Cancer.18 These included how much of the guidelines they had read (none, part, or all), how frequently they looked at them (dichotomized for analysis between more than yearly vs once a year or less), and how they accessed them (online, CD, print, or not applicable). A battery of items assessed attitudes toward the NCCN Guidelines specifically. Respondents were asked in what proportion of breast cancer cases were their treatment recommendations consistent with NCCN Guidelines and in what proportion did they refer patients to the patient versions of the NCCN Guidelines.
Demographic and practice characteristics evaluated in the survey included age, experience (years since completing residency), gender, race/ethnicity, practice characteristics (whether the practice teaches residents or fellows; whether the respondent practices in an NCI-designated comprehensive cancer center; urban, suburban, or rural setting), work hours, number of new patients with breast cancer seen in the past year, percent of patients with newly diagnosed breast cancer discussed in a tumor board, and percent of patients insured by Medicaid (to measure the socioeconomic status of the patient population).
Analysis
The authors first generated descriptive statistics regarding the general demographic and practice characteristics of the respondent population, comparing responding surgeons with medical oncologists using the chi-square or 2-sample t test statistics for categorical and continuous data, respectively. They then generated descriptive statistics regarding attitudes, familiarity, use, and compliance and compared responses by specialty, again using the chi-square or 2-sample t test statistics for categorical and continuous data, respectively. For initial description and comparison of attitudes, the 4-point response scales (agree vs disagree) were dichotomized for clarity of presentation.
The authors also constructed a scale evaluating general attitudes toward guidelines, based on the 8 items previously developed by Tunis et al.13 Each item was weighted equally and was scored from 1 to 4, with higher scores representing more positive attitudes toward guidelines. Multiple variable models of this scaled attitude measure were then constructed using linear regression, with theoretically selected independent variables of physician gender, race, years since residency, specialty, teaching status, cancer center status, practice location, and percent of Medicaid patients.
Multiple variable models were also constructed to evaluate correlates of guideline adherence, using a logistic regression framework. The binary dependent variable was defined as making recommendations that were consistent with NCCN Guidelines in almost all patients with breast cancer versus a lesser share of those patients. Independent variables included physician gender, race, years since residency, specialty, teaching status, cancer center status, practice location, percent of Medicaid patients, and number of new patients with breast cancer.
Full models that include all variables and parsimonious models constructed using iterative backward selection methods are reported. Test statistics with P values of 0.05 or less were considered statistically meaningful.
Results
Figure 1 details the response to this survey. A total of 896 responses were received from the 1481 practicing physicians for whom valid mailing addresses were available (60%). Of the respondents, 130 did not see any new patients with breast cancer in the past year. The remaining 766 respondents constituted the analytical sample.
Evolution of the analytical sample.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 12, 2; 10.6004/jnccn.2014.0021
Evolution of the analytical sample.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 12, 2; 10.6004/jnccn.2014.0021
Evolution of the analytical sample.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 12, 2; 10.6004/jnccn.2014.0021
Demographic and Practice Characteristics of Analytical Sample
Table 1 details the characteristics of the sample by specialty. Mean participant age was 52 years and mean time in practice since completing residency was 19 years, with small differences by specialty. Most respondents were male, and this proportion was higher among surgeons (85%) than medical oncologists (73%). A higher proportion of surgeons (85%) than medical oncologists (73%) were white, and a higher proportion of medical oncologists were Asian (26% vs 10%). Most respondents were in nonteaching practices (75%), and few worked in a comprehensive cancer center (13% of medical oncologists and 18% of surgeons). Medical oncologists in this sample had a higher volume of patients with newly diagnosed breast cancer, with 30% reporting more than 50 in the previous year, compared with 16% of surgeons. Across providers, the mean percentage of patients with newly diagnosed breast cancer discussed in a multidisciplinary tumor board was 45%.
As shown in Table 2, general attitudes toward guidelines were positive. Nearly all survey participants agreed that guidelines were good educational tools, convenient, and intended to improve the quality of care, and 88% felt that guidelines were unbiased. However, a nontrivial minority reported feeling that guidelines were oversimplified (24%), too rigid (20%), and a challenge to physician autonomy (20%). Opinion was split regarding whether guidelines were intended to decrease costs, with 51% of physicians believing that they were.
On multivariable analysis, attitudes toward guidelines were associated with specialty (P=.03); medical oncologists reported slightly more positive attitudes toward guidelines than surgeons, with an effect size of 0.6 on the 24-point scale. Attitudes toward guidelines were not associated with physician gender (P=.48), race (P=.22), experience (P=.99), work hours (P=.59), practicing in a teaching context (P=.87) or comprehensive cancer center (P=.65), location (P=.97), or percent of patients with Medicaid (P=.41).
Medical Oncologist and Surgeon General Attitudes Toward Guidelinesa (N=766)
Most respondents reported that they discussed with patients whether their recommendations were guideline-concordant. When making a guideline-consistent recommendation, 63% reported discussing with patients the fact that their recommendation was consistent with guidelines, with no difference according to specialty (P=.34). When asked what they do when they make guideline-inconsistent recommendations, 25% reported that they never make guideline-inconsistent recommendations, 57% reported that they routinely discuss the inconsistency with the patient, and 17% reported that they do not discuss it with patients. In this response, a significant difference was seen by specialty (P=.001), with 22% of medical oncologists and 28% of surgeons reporting never making guideline-inconsistent recommendations, 56% of medical oncologists and 60% of surgeons reporting that they discuss the inconsistency, and 22% of medical oncologists and 12% of surgeons failing to discuss the inconsistency.
As shown in Figure 2, nearly all medical oncologists (96%) and a significantly lower proportion of surgeons (70%; P<.001) reported that the NCCN Guidelines generally influence their cancer management decisions. A substantial proportion of each group, but again more often medical oncologists (65% vs 45% of surgeons; P<.001), reported being influenced by ASCO guidelines, and 19% of each group described being influenced by other guidelines (P=.96). Very few medical oncologists (1%) and a higher proportion of surgeons (9%; P<.001) reported that no guidelines influenced their cancer management decisions.
Physician responses regarding which guidelines generally influenced their cancer management decisions.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 12, 2; 10.6004/jnccn.2014.0021
Physician responses regarding which guidelines generally influenced their cancer management decisions.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 12, 2; 10.6004/jnccn.2014.0021
Physician responses regarding which guidelines generally influenced their cancer management decisions.
Citation: Journal of the National Comprehensive Cancer Network J Natl Compr Canc Netw 12, 2; 10.6004/jnccn.2014.0021
Medical oncologists and surgeons differed in their familiarity with and use of the NCCN Guidelines specifically. All responding medical oncologists and a lower proportion of responding surgeons (74%; P<.001) were aware that NCCN provides guidelines for cancer management before participating in this survey. Although most (91%) medical oncologists reported looking at NCCN Guidelines more than once a year, fewer of the surgeons who were aware of the guidelines did so (59%; P<.001). Overall, 48% of medical oncologists and 24% of surgeons reported having read all of the NCCN Guidelines for Breast Cancer (P<.001). Medical oncologists were more likely to use the online version of the guidelines than surgeons (95% vs 80%; P<.001), and less likely to use the print versions (20% vs 39%; P<.001).
Among those aware of the NCCN Guidelines, attitudes were positive, as detailed in Table 3. Of this group, 96% reported often agreeing with NCCN recommendations, and only a minority agreed that NCCN Guidelines were outdated (6%) or too restrictive (10%). Still, 30% of respondents reported that they rely on their own interpretation of data over the NCCN Guidelines, and 23% believed that NCCN Guidelines are too complex. In contrast to the substantial differences in familiarity and use observed between medical oncologists and surgeons, the only difference by specialty in attitudes toward these guidelines was in regard to whether the NCCN Guidelines were too restrictive to allow individualized care, which was endorsed by 12% of medical oncologists and 7% of surgeons (P=.04).
Most respondents aware of the NCCN Guidelines (75%) reported that almost all of their breast cancer treatment recommendations were consistent with these guidelines, with no difference by specialty (P=.49). Still, most (77%) also reported that they refer one-fourth or fewer of their patients with breast cancer to the NCCN Guidelines for Patients. As shown in Table 4, high adherence to the NCCN Guidelines for Breast Cancer was not associated with specialty but was significantly associated with male sex, white race, lesser experience, and practicing outside a comprehensive cancer center.
Discussion
In a large national sample of medical oncologists and surgeons who treat breast cancer, results generally showed favorable attitudes toward physician-directed guidelines for cancer management. Of the many potential barriers to guideline adherence identified in previous studies, lack of familiarity with existing guidelines seems to be the most prevalent issue in this population, particularly among surgeons.
The general attitudes reported by the participants are largely consistent with those observed in previous studies of physicians’ attitudes towards practice guidelines.13,14 Few other studies have focused on attitudes toward, familiarity with, use of, and compliance with cancer management guidelines specifically. In 2000, ASCO’s Health Services Research Committee conducted a formal survey of ASCO membership that identified a high level of readership and support for the content of the first 4 ASCO clinical practice guidelines.19 Additionally, results from a more recent informal survey study of ASCO members suggested that oncologists are more likely to use NCCN Guidelines than ASCO guidelines.5
This work builds on these efforts and extends the understanding of the use of cancer management guidelines beyond that of select specialty society members, allowing the estimation of impact on a broader population of physicians who care for patients with cancer. Results showed that NCCN and ASCO are the most frequently used cancer management guidelines, and these guidelines seem to have differing levels of penetration overall, with physicians most likely to report being influenced by the NCCN Guidelines. Like others who have considered this issue,5,19 the authors suspect that this relates to differences in timing of guideline development, and differences in their scope, focus, and methods of development.
Medical Oncologist and Surgeon Attitudes Toward NCCN Guidelines Among Those Aware of Their Existencea (n=661)
These study results reveal important differences in the familiarity with and use of existing guidelines by medical oncologists and surgeons. Medical oncologists were significantly more likely than surgeons to report that their cancer management decisions were influenced by either the NCCN or ASCO guidelines. Additionally, medical oncologists and surgeons differed in their awareness of the existence of the NCCN Guidelines, how much they had read, and how frequently they looked at them. Previous studies have shown that subspecialists tend to prefer guidelines issued from their subspecialty groups more than general guidelines, which may explain some of the differences observed.16 However, the lack of awareness of even the existence of the NCCN Guidelines by a substantial minority of respondents, all of whom treat patients with cancer, is striking, particularly because the NCCN Guideline development process intentionally includes representatives of the surgical community.
The authors did not find systematic differences in attitudes toward guidelines by practice setting, gender, or experience, and only a modest difference by specialty. They did, however, find associations between gender, race, experience, and practice setting with adherence to the NCCN Guidelines for Breast Cancer. The finding that non-white physicians were less likely to make guideline-consistent recommendations would be particularly important if non-white physicians were more likely to serve non-white patients; this merits further research and understanding. The finding that experienced physicians and those practicing at comprehensive cancer centers were less likely to make guideline-consistent treatment recommendations might relate to the expanded availability of clinical trial and novel therapies at cancer centers, and a possible greater willingness of experienced practitioners and experts in a field to deviate from guidelines or recognize that they are doing so, although this was not directly measured in the survey.
Multiple Variable Model to Explain Reporting That “Almost All” of the Physician’s Breast Cancer Treatment Recommendations Are Consistent With NCCN Guidelines
This study has several strengths, including a high response rate from a national sample of oncologists and surgeons who were selected solely based on having seen at least one new breast cancer case in the preceding year, and detailed measures adapted from prior work in the field that allow for rich description of practicing clinicians’ attitudes and behaviors related to cancer management guidelines. It also has limitations. The investigators did not directly observe practice and were only able to measure guideline adherence through physician self-report, which may be subject to bias relating to misinformation, recall, or social desirability.20 They did not measure reasons for guideline nonadherence or ethnic composition of the patients treated by the physicians surveyed, and therefore further research is necessary to build on these findings. Nonresponse bias is also possible, although the high response rate somewhat alleviates this concern.
These findings have important implications. Among those aware of the NCCN Guidelines, attitudes were generally favorable, and adherence, in the context of breast cancer management, was high. However, guidelines seem to have greater visibility among the medical oncology audience than among surgeons, and patient versions of the NCCN Guidelines seem to be particularly infrequently recommended by providers. This is particularly striking given numerous observational studies that have suggested that not all patients with breast cancer receive appropriate care, and disparities may affect certain vulnerable subgroups.21-23 Because surgical providers and patients may make initial breast cancer management decisions before interacting with medical oncologists, greater efforts to familiarize these groups with existing cancer management guidelines may be particularly useful in promoting the delivery of consistently high-quality care to all patients.
This work was supported by a Young Investigator Award from the National Comprehensive Cancer Network to Dr. Jagsi. The authors have disclosed that they have no financial interests, arrangements, affiliations, or commercial interests with the manufacturers of any products discussed in this article or their competitors.
References
- 1.↑
Institute of Medicine. Clinical practice guidelines we can trust. Available at: http://www.iom.edu/Reports/2011/Clinical-Practice-Guidelines-We-Can-Trust.aspx. Accessed December 20, 2014.
- 2.↑
Eccles MP, Grimshaw JM, Shelkelle P et al.. Developing clinical practice guidelines: target audiences, identifying topics for guidelines, guideline group composition and functioning and conflicts of interest. Implement Sci 2012;7:60.
- 3.↑
Genuis SJ. The proliferation of clinical practice guidelines: professional development or medicine-by-numbers? J Am Board Fam Pract 2005;18:419–442.
- 4.↑
Hurwitz B. Clinical guidelines: proliferation and medicolegal significance. Qual Health Care 1994;3:37–44.
- 5.↑
Dillmon MD, Goldberg JM, Ramalingam SS et al.. Clinical practice guidelines for cancer care: utilization and expectations of the practicing oncologist. J Oncol Pract 2012;8:350–353.
- 6.↑
Grimshaw JM, Russell IT. Effect of clinical guidelines on medical practice: a systematic review of rigorous evaluations. Lancet 1992;342:1317–1322.
- 7.
Worrall G, Chaulk P, Freake D. The effects of clinical practice guidelines on patient outcomes in primary care: a systematic review. CMAJ 1997;156:1705–1712.
- 8.
Grol R. Personal paper. Beliefs and evidence in changing clinical practice. BMJ 1997;315:418–421.
- 9.
Reker DM, Duncan PW, Horner RD et al.. Postacute stroke guideline compliance is associated with greater patient satisfaction. Arch Phys Med Rehabil 2002;83:764–770.
- 10.↑
Pieper C, Haag S, Gesenhues S et al.. Guideline adherence and patient satisfaction in the treatment of inflammatory bowel disorders—an evaluation study. BMC Health Serv Res 2009;9:17.
- 11.↑
Brand C, Landgren F, Hutchinson A et al.. Clinical practice guidelines: barriers to durability after early implementation. Intern Med J 2005;35:162–169.
- 12.↑
Carlsen B, Bringedal B. Attitudes to clinical guidelines—do GPs differ from other medical doctors? BMJ Qual Saf 2011;20:158–162.
- 13.↑
Tunis SR, Hayward RSA, Wilson MC et al.. Internists’ attitudes about clinical practice guidelines. Ann Intern Med 1994;120:956–963.
- 14.↑
Ward MM, Vaughn TE, Uden-Holman T et al.. Physician knowledge, attitudes and practices regarding a widely implemented guideline. J Eval Clin Pract 2002;8:155–162.
- 15.↑
Lugtenberg M, Burgers JS, Besters CF et al.. Perceived barriers to guideline adherence: a survey among general practitioners. BMS Fam Pract 2011;21:98.
- 16.↑
Cabana MD, Rand CS, Powe NR et al.. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA 1999;282:1458–1465.
- 17.↑
Dillman DA, Smyth JD, Christian LM. Internet, mail, and mixed-mode surveys: the tailored design method, 3rd ed. Hoboken, NJ: Wiley & Sons; 2009.
- 19.↑
Bennett CL, Somerfield MR, Pfister DG et al.. Perspectives on the value of American Society of Clinical Oncology clinical guidelines as reported by oncologists and health maintenance organizations. J Clin Oncol 2003;21:937–941.
- 20.↑
Steinman MA, Fischer MA, Shlipak MG et al.. Clinical awareness of adherence to hypertension guidelines. Am J Med 2004;117:747–754.
- 21.↑
Martinez SR, Beal SH, Chen SL et al.. Disparities in the use of radiation therapy in patients with local-regionally advanced breast cancer. Int J Radiat Oncol Biol Phys 2010;78:787–792.
- 22.
Freedman RA, He Y, Winer EP, Keating NL. Trends in racial and age disparities in definitive local therapy of early-stage breast cancer. J Clin Oncol 2009;27:713–719.
- 23.↑
Du XL, Gor BJ. Racial disparities and trends in radiation therapy after breast-conserving surgery for early-stage breast cancer in women, 1992 to 2002. Ethn Dis 2007;17:122–128.
