Barriers to Choosing Wisely® in Primary Care: It’s Not Just About “the Money”

In this issue of the Journal of General Internal Medicine, Zikmund-Fisher and colleagues report on results from two national surveys of primary care clinicians exploring the barriers to implementing 12 specific Choosing Wisely® (CW) recommendations.¹ The investigators describe survey findings from 34 % of 2000 internal medicine, family medicine, and geriatrics physicians contacted through a nationally representative random sample and 48 % of 2500 representative primary care clinicians (physicians, nurse practitioners, and physician assistants) practicing in Department of Veterans Affairs (VA) facilities. Despite the differences in the two surveyed clinician populations, and in their practice settings, the different respondents noted remarkably consistent concerns regarding the feasibility of following these CW recommendations in their setting.

Perhaps the least surprising finding is the four CW recommendations respondents thought would be easiest to follow. Clinicians know well the difficulty convincing patients to adhere to recommendations for colon cancer screening. Thus, it is not surprising that responding clinicians did not anticipate difficulty deferring “colorectal cancer screening…if the patient has a negative colonoscopy” for 10 years or 5 years for repeat colorectal cancer screening when a prior colonoscopy found low-risk lesions. Similarly clinicians involved in the preoperative evaluation understand well the potential inconvenience to patients and surgery colleagues that can result from unneeded cardiovascular testing for patients undergoing low-risk surgery and thus may be pleased to follow this CW recommendation. While screening for carotid artery stenosis (CAS) in asymptomatic adult patients has at times been advocated, this screening test has not been readily reimbursed by private payers nor commonly performed in primary care settings. Furthermore, the US Preventive Services Task Force 2014 renewal of their 2007 position against screening generated little editorial outcry; indeed, the Society of Vascular Surgery included this among their CW recommendations. Therefore, it is not surprising that primary care clinicians might find this recommendation “both comparatively easy to follow and easy for patients to accept.”

Clinician views regarding the recommendations they find more “difficult to follow” are more interesting. The recommendation to avoid “DEXA screening among younger patients with no risk factors” proved to be an issue for 23 % of the non- VA physician respondents, but only 9 % of VA-based clinicians. Unfortunately, the data presented are insufficient to probe the specific cause for this differing view. It is of course plausible that the much lower prevalence of female patients eligible for DEXA screening who are cared for at VA facilities makes this recommendation less salient. It might also be the case that the very different financial incentives around this test have affected its availability and use. DEXA became an additional source of FFS revenue for many private physician offices but would likely require a referral to radiology for many VA primary care clinicians. Indeed this is the one CW recommendation studied here that could result in a reduction in billable professional services by the decision-makers studied (since the colonoscopies and imaging studies averted are not usually performed by primary care clinicians). In work on barriers to specialist physicians following CW recommendations, an impact on the physician’s “productivity measures” derived from billable services was noted as a factor in the ease of following some recommendations.² In Zikmund-Fisher’s study, 35 % of non-VA survey respondents noted their primary clinical practice compensation was from billings; another 38 % received salary plus bonus, but the specific incentives are not described.

The other two “difficult to follow” recommendations may arise from different origins. Many physicians have been dismayed by the use of quality metrics in their practices,³ and some policy researchers have expressed concern regarding the impact of such measures on health professional decision-making. Thus, it is plausible that the concern regarding the hemoglobin A1c recommendation derives from past frustration with such quality measures and difficulty with either physician or local administrator understanding of their nuances. The CW recommendation regarding imaging for suspected pulmonary embolism reflects neither common financial incentives nor performance measures in primary care, however. This may therefore derive from a more basic challenge to rational clinical decision-making—regret⁴—and the clinician’s desire to avert life-threatening illness through use of a highly accurate, relatively low-risk test.

The authors reported that for 5 of the 12 CW recommendations studied, a substantial fraction of respondents anticipated patients would find these difficult to accept. Two of these recommendations also avoid advanced imaging—for either initial evaluation of uncomplicated low back pain or for evaluating simple syncope. Like the aforementioned testing for PE, these also employ advanced imaging to rule out serious disease, but the clinicians’ reported concern relates more to the patient than to clinician expectations. Various studies have shown the complex influence of the practice environment on clinician judgements⁵. Additional research may be required to better understand the practice setting factors that contribute to these seeming differences in clinician perception of patient expectations for testing (and the educational strategies or point of care interventions to support implementing these CW recommendations).

The other three CW examples were relevant to medications often prescribed in primary care: use of antibiotics for sinusitis, use of sedative-hypnotics for insomnia, agitation, or delirium in older adults, and antimicrobials for asymptomatic bacteriuria in older adults. All of these might reflect some concern with achieving patient or family member satisfaction during the limited time for clinical decision-making. Other studies have found that clinicians express concern over effects on patient satisfaction scores when implementing such CW recommendations.² Although certainly direct-to-consumer advertising (DTCA) of pharmaceuticals has become a factor in some primary care decisions,⁶ these clinical scenarios are unlikely to be the focus of DTCA. It may be however that although many primary care patients (and family members) willingly accept such evidence-based recommendations, the clinicians’ memory of these encounters as more time-consuming or otherwise less satisfying influences their perceptions and practices.

Thus, it is not surprising that “patient requests for tests and treatments” and “lack of time for shared decision-making with patients” were frequently listed as major barriers to implementing the CW recommendations under study. Additional understanding of the practice settings and clinician patient relationship may be required to identify the relevant tools and techniques to address these concerns. Patients considering CW recommendations emphasize the importance of a trusting relationship with the clinician,⁷ and past studies of primary care demonstrate greater patient adherence to evidence-based recommendations when clinicians have a continuity relationship with patients.⁸ However, such interpersonal continuity can vary considerably across primary care settings and encounters, which may alter the resources required to support patient acceptance of clinician recommendations in circumstances like these. A practice environment characterized by direct patient access to heavily marketed specialized centers may also be perceived by primary care clinicians to create patient expectations for certain services (like early advanced imaging for back pain). Interestingly, when interviewed about their own barriers to CW recommendations, various specialists cite primary care referrals for unneeded testing and treatment as a factor.²

Regarding policy solutions to promote CW, fee for service (FFS) payment has been widely discussed as a key driver of “overutilization” of clinician services. Accordingly, the recently passed Medicare Access and CHIP Reauthorization Act (MACRA) sets forth an ambitious agenda to shift clinicians to “alternative payment models.”⁹ However, as noted above, physicians practicing in very different payment environments report similar problems implementing most CW recommendations. Indeed, for only one of the 12 examples studied here, DEXA screening, is it even plausible that FFS incentives might be a factor for use in primary care practices. Thus, a simple shift away from FFS is unlikely to offer the chief solution to implementing these recommendations. From these findings it is also not apparent that shifting physicians from smaller practices to larger health care systems (like VA facilities) necessarily promotes CW. Payment reform can offer mechanisms to promote improved resources for better patient communication and shared decision-making, however. For example, MACRA will provide FFS incentives to reward practices that implement a variety of clinical practice improvement initiatives including “use evidence-based decision aids to support shared decision-making” and “participation in a QCDR (data registry) that promotes use of processes and tools that engage patients for adherence to treatment plan.”¹⁰ Furthermore, some Medicare alternative payment models are now promoting the introduction of resources for shared decision-making into the local practice setting.¹¹

As the authors conclude, “cross-recommendation variations in provider attitudes…imply that implementation efforts will need to be tailored to the specific barriers of implementing each CW recommendation.” Their findings suggest that these barriers may vary by practice setting as well. The changing environment for clinician payment and health care delivery creates opportunities to better understand the diverse factors that promote, or interfere with, more evidence-based care. Achieving such understanding can help medical educators and practice leaders develop the robust tool set needed to help patients and clinicians “choose more wisely.”

Compliance with Ethical Standards

Conflict of Interest

The author declares he has no conflicts of interest.