The Thinking that Drives Low Value Care

In this blog Jason Soon, Senior Policy Officer (Health Economics), Policy & Advocacy, The Royal Australasian College of Physicians discusses the recent MJA article on "Countering cognitive biases in minimising low value care"

Low value healthcare is healthcare that offers little or no benefit, may cause patient harm, is not aligned with patient preferences or is only partially beneficial at a disproportionately high cost. Such low value care (LVC) has spawned numerous evidence-based and clinician led initiatives in Australia and around the world which attempt to raise the awareness of clinicians of when such care occurs so they can take steps to reduce its incidence. Our recent article in the MJA argues that while making clinicians aware of instances of commonly occurring LVC and the evidence base for why particular practices constitute LVC are laudable objectives, they are necessary but not sufficient for reducing the incidence of LVC. The appeal to clinicians as fully rational decision makers obviously has its place and is the basis for traditional knowledge translation tools such as clinical decision support, audits and feedback, guidelines and quality incentives. However these approaches by themselves will ultimately bump up against limits with some evidence suggesting that they optimise decisions in only up to 20% of instances.


This is because the tendency to deliver LVC is not just a conscious intellectual error on the part of clinicians but is partly driven by intuitive decision making processes of which they are only partially aware. These decision making processes, while useful in a variety of contexts, are obviously not perfect and can be manifested through a number of cognitive biases which we  identify in our article such as:

  • Commission bias: doctors try to avoid experiencing a sense of regret at not administering an intervention that could have benefited at least a few recipients, leading them to over-administer treatments.


  • Attribution bias: relying on anecdotal and selective observations of positive outcomes to a treatment leading to undue confidence in its effectiveness.


  • Impact bias: patients and clinicians may tend to overestimate the benefits and underestimate the harms of treatments. This may tend be exacerbated by framing effects (e.g. presenting benefits in relative measure terms rather than using absolute measures tends to put treatments in a more favourable light).


  • Availability bias: emotionally strong case studies with either good or bad outcomes that come easily to mind can unduly inflate estimates of the likelihood of the same scenario being repeated.


  • Extrapolation bias: because a treatment works in one type of patient, doctors think it will work in another.


  • Ambiguity bias: even when the evidence that defines a treatment as being of low value is well known and accepted, treatments are still performed to reassure the patient or their peers.


  • Endowment effects and default bias: When patients and clinicians place a greater value than they may otherwise on a longstanding form of care that is about to be withdrawn.


Our article goes on to suggest a number of ways of mitigating the influence of these cognitive biases such as:

  • Cognitive huddles and autopsies: getting together in a collegiate environment to discuss case studies of LVC including identifying where errors in decision making may have crept in to lead to the example of LVC.


  • Narratives of patient harm: Providing sobering case narratives of significant patient harm resulting from ill-advised actions. In particular, this approach can serve to counter commission bias.


  • Defining acceptable levels of risk of adverse outcomes if an intervention is withheld: this can help reorient thinking back towards ensuring there is a substantive rationale for action.


  • Providing alternatives to the LVC that would otherwise be provided as a means of channeling clinician’s need to ‘do something’. These alternatives may be relatively low intensity but ultimately less harmful e.g. ‘watchful waiting’.


  • Nudge strategies and default options: these are attempts to exploit the very same cognitive biases discussed previously in order to channel them more constructively e.g. setting defaults and reminders in electronic health records so that clinicians end up defaulting towards higher value alternatives to LVC.


  • Shared decision making: By empowering patients for instance through decision aids and other means of familiarising them with the various options available to manage their condition, shared decision making between patients and their clinicians can be facilitated. As informed patients are unlikely to agree to consent to LVC, this provides an additional pressure on clinicians to reduce its incidence. For instance there is evidence that the use of decision aids, which present individualised estimates of absolute benefit and harm, reduces the need for elective procedures by 21%.


In conclusion, we believe that cognitive biases create predispositions to LVC that need to be addressed and which may otherwise limit the impact of more traditional knowledge translation tools such as clinical guidelines. At the same time our article acknowledges that the debiasing strategies that we suggest, although they have strong face validity, are still relatively new so more research is needed in this area, including through randomised clinical trials, to figure out which debiasing strategy designs work best in reducing LVC.

The thinking that drives low value care