Current clinical practice is characterized by substantial variation in delivery of health care for the same conditions.1 In turn, clinical variation is considered one of the major drivers of ever-increasing health care costs1 contributing to the estimated 30% of inappropriate or wasteful health care.2 Perhaps as a natural response to this unsatisfactory situation, a widespread and influential school of thought has emerged contending that greater uniformity of clinical practice is desirable.1,3 Advocates maintain that by achieving uniformity in care, practice variation can be decreased, in turn leading to large cost reductions. The suggested mechanism to achieve uniformity in part involves clinician adherence to practice guidelines, which is seen as synonymous with evidence-based practice.3 In this Viewpoint, we explain that this position is based on a misunderstanding of trustworthy guidelines4 and that striving for uniformity of practice as an end is misguided. The first limitation in the drive for uniformity is a failure to appreciate the need for guidelines that achieve a high standard of trustworthiness.4 Flawed guidelines, particularly those that offer strong recommendations when only weak recommendations are warranted, are unlikely to facilitate optimal practice. Advocates for uniformity should highlight the distinction between flawed vs trustworthy guidelines and ensure that only the latter are seen as a guide for clinical practice.4,5 Uniformity of practice means that patients with a similar presentation of underlying condition, severity, and circumstances will receive the same treatment— that is, they will all receive or not receive particular diagnostic tests, medications, or surgical procedures. There are circumstances in which such uniformity is desirable. In particular, when clinicians are confident in the estimates of the effects of health interventions on all outcomes of importance to patients (ie, high-quality evidence is available), the desirable consequences of the one management strategy clearly outweigh its undesirable consequences, and the relative importance patients place on the key outcomes (ie, their values and preferences) are similar across patients. That the recommended intervention is not overly costly, is highly feasible, is acceptable to clinicians and patients, and reduces health disparities further strengthens the case for uniformity of clinical practice. There are many, indeed the majority of, situations in which clinicians and patients make management decisions of importance, but these conditions are not met. For example, the American Board of Internal Medicine’s Choosing Wisely campaign has identified 135 health interventions in which evidence suggests equivalence but not superiority to other options in terms of benefits, harms, and costs.6 Studies directly addressing the relevant questions may not have been undertaken, or if they have, they may be small, poorly designed or implemented, show inconsistent results, be limited by publication bias, or have enrolled idiosyncratic populations of questionable applicability. In these cases, the confidence in the estimate of effects will often be low or very low. In addition, if values and preferences differ widely across patients (which is often if not uniformly the case), the right decision for one patient may be the wrong decision for another. For example, Montori et al7 illustrated how recent guidelines by the American College of Cardiology and the American Heart Association for the use of statins for primary prevention of heart disease do not mandate uniform practice—some patients informed about cardiovascular disease risk reduction will choose the recommended course of action and use statins, but others will not. Organizations that produce guidelines should distinguish between situations in which confidence in effect estimates is high and the balance between desirable and undesirable consequences is clear and when these conditions do not exist. In the former situation the guideline will include strong, definitive recommendations that often warrant uniformity of practice. In the latter, they will issue weak (conditional or contingent) recommendations. The recommended course of action will be right for many patients, but not right for all.5 In this scenario, uniformity of practice will not be appropriate. For example, administration of at least 1 antiplatelet agent and a statin for patients who have experienced a myocardial infarction; a β-blocker and an angiotensin-converting enzyme inhibitor for patients with systolic heart failure; corticosteroids for patients experiencing asthma exacerbation; and curative chemotherapy for patients with large cell lymphoma all warrant strong recommendations. These situations therefore warrant uniform practice, other than in exceptional situations such as drug allergy or severe comorbidities. In each case, clinicians should have high confidence in estimates of effect and in the desirable consequences of the intervention being much greater than the undesirable consequences. In contrast, use of anticoagulants for low-risk patients with atrial fibrillation; indefinite anticoagulation for patients with unprovoked venous thromboembolism; use of corticosteroids for patients with idiopathic pulmonary fibrosis; and chemotherapy for patients with advanced non–small cell lung cancer warrant weak recommendations. In these instances, in which clinicians have either low confidence in VIEWPOINT
[1]
D. Berwick,et al.
Eliminating waste in US health care.
,
2012,
JAMA.
[2]
James S. Goodwin,et al.
Tracking Medicine: A Researcher's Quest to Understand Health Care By John E. Wennberg
,
2011
.
[3]
Victor M Montori,et al.
Patient-centered and practical application of new high cholesterol guidelines to prevent cardiovascular disease.
,
2014,
JAMA.
[4]
Problems of stopping trials early
,
2012,
BMJ : British Medical Journal.
[5]
G. Guyatt,et al.
GRADE: an emerging consensus on rating quality of evidence and strength of recommendations
,
2008,
BMJ : British Medical Journal.
[6]
M E Hammond,et al.
The challenge of variation in medical practice.
,
2009,
Archives of pathology & laboratory medicine.
[7]
Martin Emmert,et al.
Effects of pay for performance in health care: a systematic review of systematic reviews.
,
2013,
Health policy.
[8]
S. Pearson,et al.
Evidentiary rationales for the Choosing Wisely Top 5 lists.
,
2014,
JAMA.