Resolving the Decision Aid Paradox

Abstract

Formostmedical treatment decisions, there ismore than one reasonable choice.Althoughbacterialmeningitisneedsantibiotics,decisionsabouttreatmentsforconditionssuchasclinically localized prostate cancer havemany options, including observation,radiationtherapy,andsurgery.Inturn,eachofthesebroad treatmentcategorieshasvariants:watchfulwaiting, active surveillance,brachytherapy,externalbeamradiotherapy,opensurgery,and laparoscopic robotic-assistedsurgery.Evidenceabout comparative effectiveness of these treatments is available but limited, and the preferences of informed patients vary.For suchpreference-sensitivedecisions, geographicpracticevariation is the rule, reflectingphysicianmore thanpatient preferences.1 For example, in a report2 fromtheDartmouthAtlas ofHealth Care, population-based rates of radical prostatectomyforprostate canceramongmen in306UShospital referral regions ranged almost 10-fold from2007 to 2012, froma lowof 0.5 toahighof4.7per 1000maleMedicarebeneficiariesaged65 to75years. Shareddecisionmaking, inwhichpatientsandphysicianscollaboratetomakedecisionsaboutpreference-sensitive treatments, has been proposed as an approach to reduce unwantedpracticevariation, butmaintaindesirablevariation, by tailoring decisions to patients’ clinical characteristics and informed preferences. In this issue,Wangandcolleagues3 reportonanational survey of US radiation oncologists and urologists, the specialists who largely treat menwith localized prostate cancer, regarding their attitudes towardanduseofdecisionaids in theirpractices. Patient decision aids are tools to make shared decision making more practical in clinical practice but are not shared decision making themselves.4 In a recent Cochrane systematic review,5 across 115 randomized trials, use of decision aids was associatedwith increased knowledge,more accurate risk perceptions, greater value concordancewithdecisions, lower decisional conflict, and less passivity in decisionmaking. Despite these positive effects on the quality of medical decisions, decision aids are underused. This report helps explain some of the barriers to their widespread use. The fact that roughly a third of respondents in both disciplines said theywereusingdecisionaids is encouraging, but, of course, one cannot be sure what materials they were actuallyusingandwhether theywouldreallycountasdecisionaids. Approximately84%of respondents thoughtdecisionaidswere at least somewhatuseful, andapproximately 78%wereat least moderately confident that they improve treatment decisions.Moreover,most respondents inboth specialties thought decision aidswere applicable to their patients, did not lead to less effective choices, could be processed by average patients, and could estimate risks of recurrence at least as well as the physicians themselves. Reassuringly, decision aid users among these specialists were even more positive in responding to thesequestions thannonusers, althoughonecannot be surewhether decision aid use is the chicken or the egg in this relationship. The results of the Cochrane review and these physicians’ generallypositive ratingsofdecisionaidscontrastedto therelatively low use rates, which presents a paradox: if these tools are so good,why aren’t theyusedmore often andwhat can be done to increase their use? One key issue is that decision aids donot fit easily into theworkflowof clinical care.Decisionaids are best initially deployed outside a physician visit so patients can get up to speed about their condition and the treatment options, particularly when the condition and the options are complex, as is the case for localized prostate cancer. Family members can be invited into the deliberations if deRelated article page 792 Prostate Cancer DecisionMaking Original Investigation Research