This article shows that where good evidence exists, decision analysis is a feasible way of incorporating patients' values and preferences into clinical decisions. The fact that only about half of the patients who were approached participated should not be viewed critically: decision analysis will not suit all patients.
Patients' choices of treatment frequently disagreed with both consensus guidelines and guidelines based on an assessment of absolute risk. Overall, the proportion of people who preferred warfarin treatment was lower than the proportion for whom such treatment is recommended by either of the guidelines. Patients' preferences did not, however, all point in the same direction. Many people preferred warfarin treatment, even though this was not recommended by either of the guidelines. Given good information, the participants were able to weigh the benefits and drawbacks of the intervention and make a personal choice.
The study shows that when patients are actively involved in clinical decision making, their preferences may strongly influence treatment decisions. Successfully involving patients in clinical decisions requires good information. The most reliable source of information about the effects of interventions comes from sufficiently large, well-conducted randomized controlled trials.1
By definition, randomized controlled trials measure the effects of randomly assigned interventions. Randomization is the key process by which bias and confounding are minimized. Can the importance of patients' preferences be reconciled with the benefits of randomization? This issue has been discussed in depth elsewhere.2,3,4 The best study design that has been proposed to tackle this dilemma uses a 2-stage approach.5 During the first stage, participants are randomly allocated to 2 groups: a “random” group and a “preference” group. In the second stage, participants in the random group are randomly allocated a second time to the 2 interventions being compared in the trial. Participants in the preference group are free to choose between the 2 interventions being assessed. This design has the unique advantage of being able to measure the influence of patients' preferences on the estimate of the treatment effect. Clearly there will be times when patients' (or clinicians') preferences for 1 treatment or another are sufficiently strong to preclude randomization.
The study by Protheroe et al shows that shared decision making can be achieved when high-quality relevant research evidence about clinical questions is available to patients and clinicians. Good clinical practice can then be informed by the evidence; it may not always follow the evidence.
Competing interests: None declared
This commentary was published in BMJ 2000;320:1384
References
- 1.Kleijnen J, Gotzsche P, Kunz RH, Oxman AD, Chalmers I. So what's so special about randomisation? In: Maynard A, Chalmers I, eds. Non-random Reflections on Health Services Research: on the 25th Anniversary of Archie Cochrane's Effectiveness and Efficiency. London: BMJ Publishing; 1997: 93-106.
- 2.McPherson K. The Cochrane lecture: the best and the enemy of the good: randomised controlled trials, uncertainty, and assessing the role of patient choice in medical decision making. J Epidemiol Community Health 1994;48: 6-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.McPherson K, Britton AR, Wennberg JE. Are randomized controlled trials controlled? patient preferences and unblind trials. J R Soc Med 1997;90: 652-656. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Torgerson DJ, Sibbald B. Understanding controlled trials: what is a patient preference trial? BMJ 1998;316: 360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Rücker G. A two-stage trial design for testing treatment, self-selection and treatment preference effects. Stat Med 1989;8: 477-485. [DOI] [PubMed] [Google Scholar]