Treatment decision aids are unlikely to cut healthcare costs.

The growing emphasis on patient centered care is increasing the demand on physicians’ time and effort to engage patients and their families in decision making regarding treatment. At the same time, the clinical encounter is becoming more challenging because evaluative testing strategies and decisions about disease management are increasingly complex. In this context, growing interest in decision aids is motivated by expectations that they can increase the efficiency and effectiveness of decision making regarding treatment and improve patients’ experiences. Indeed, there is compelling evidence that decision aids can increase a patient’s knowledge, satisfaction with the clinical encounter, and engagement with clinicians. More recently, decision aids have been advocated as a tool for dealing with the growing concerns about overtreatment and cost inflation. The argument is anchored on the assumption that a better informed and more engaged patient would be less likely to choose a management plan as extensive as that recommended by their clinician. The linked paper by Walsh and colleagues (doi:10.1136/bmj.g188) reviews the suggestion that decision aids could reduce costs, which has permeated policy statements and commentaries on both sides of the Atlantic. Walsh and colleagues underscore the gaps in the literature examining the question of whether decision aids can save money. Only a handful of articles have dealt with these issues over the past decade, and they vary markedly with respect to design, provider setting, decision context, structure of the intervention (such as educational booklets, video discs, phone consultation, or coaching), and approach to outcome measurement. None sufficiently measured mechanistic factors that could have informed how the interventions actually worked. This is a critical weakness because it is difficult to determine whether these interventions influence patient decisions or clinician directed decisions about treatment. The difference has important implications for strategies to influence the outcomes of clinical encounters including costs. Finally, the authors note that in some studies there was lack of attention to concomitant initiatives to lower costs, which might have further confounded the findings. Thus, not surprisingly, they conclude that there is insufficient evidence that decision aids save money. Ironically, Walsh and colleagues cite a recent Cochrane meta-analysis to support the argument that decision aids can reduce overtreatment. All of the aforementioned criticisms, however, apply to most of the studies that were used in that analysis as well. Perhaps the most important problem with the argument that decision aids can reduce overtreatment and lower costs is inadequate attention paid to placing research on patient communication into a broader framework of cost, quality, and value in medical care. The factors that contribute to growth in healthcare spending are well known. The dominant cause is the increasing intensity of technology and services per capita—largely controlled by clinicians. Thus, proposed strategies to improve quality and value in healthcare are mainly directed at the clinician and system level, including payment reform. Through this prism, patient decision support tools are best viewed as cost effective rather than cost saving technologies. The difference between framing technologies as cost effective versus cost saving is often confused, but never trivial. Cost effective strategies improve outcomes at acceptable costs; cost saving strategies reduce the budget. Few effective medical technologies reduce the budget. The argument that decision aids could be one of those cost saving strategies is weak. In the context of the complexity of patient-provider communication in clinical encounters there is no compelling evidence that decision aids change patient behavior. Thus, it is difficult to argue that they do so in any one direction, let alone in the direction of less costly treatments. While preferences for treatments are no doubt sensitive to the patient’s own out of pocket costs, there is no reason to believe that patients in the examination room are sensitive to system or payer costs. Furthermore, there is no evidence that patients are less likely to choose a management plan less extensive than that recommended by their doctors. On the contrary, there is evidence that patients often have unrealistic expectations regarding the perceived benefit of treatments. 10 Thus, the growing acceptance that more patient engagement in decision making about treatment could reduce overtreatment and yield