Integrating Evidence and Individual Preferences Using a Web-Based Multi-Criteria Decision Analytic Tool An Application to Prostate Cancer Screening

Background: Annalisa (c) (AL) is a web-based decision-support template grounded in multi-criteria decision analysis (MCDA). It uses a simple expected value algorithm to calculate a score for each option by taking into account the individual's preferences for different criteria (as importance weights) and the evidence of the performance of each option on each criterion. Given the uncertainty surrounding the trade offs between benefits and harms for prostate cancer screening, this topic was chosen as the vehicle to introduce this new decision-support template. Objective: The aim of the study was to introduce a new decision-support template, AL, and to develop and pilot a decision-support tool for prostate cancer screening using it. Methods: A decision-support tool for prostate cancer screening (ALProst) was implemented in the AL template. ALProst incorporated evidence on both the benefits and the potential harms of prostate cancer screening (the 'attributes') from published randomized controlled trials (RCTs). Individual weights for each attribute were elicited during interviews. By combining the individual's preferences and the evidence, the best option for the user was identified on the basis of quantified scores. A convenience sample of computer-proficient primary-care physicians (general practitioners [GPs] in Australia) from the Sydney Metropolitan area (Australia) were invited to complete a face-to-face interview involving the decision-support tool. Preference for undergoing prostate-specific antigen testing for prostate cancer, both personally and for their patients, was sought prior to seeing the tool. After gaining hands-on experience with using the tool, GPs were asked to comment on the merits of the template and the tool. Preference for presenting the benefits of prostate cancer screening as the relative or absolute risk reduction in prostate cancer-specific mortality was also sought. Results: Of 60 GPs approached, ten (six men and four women) completed an interview (16.7% response rate). Most GPs agreed/strongly agreed with positive statements about the ease with which they could use AL (seven GPs), and understand the information in, and format of, AL (nine and eight, respectively). Eight agreed/strongly agreed that ALProst would be a useful tool for discussing prostate cancer screening with their patients. GPs were also asked to nominate difficult clinical decisions that they, and their patients, have had to make; responses included cancer screening (including prostate cancer); treating patients with multiple illnesses/diseases; managing multiple cardiovascular disease risk factors; and managing patients who are receiving multiple medications. The common element was the need to consider multiple factors in making these complex decisions. Conclusions: AL is distinguishable from most other decision-support templates available today by its underlying conceptual framework, MCDA, and its power to combine individual preferences with evidence to derive the best option for the user quantitatively. It therefore becomes potentially useful for all decisions at all levels in the healthcare system. Moreover, it will provide a universal graphic 'language' that can overcome the burden to patients of encountering a plethora of widely varying decision aids for different conditions during their lifetime.