Identifying the biologic closest to the ideal to treat chronic plaque psoriasis in different clinical scenarios: using a pilot multi-attribute decision model as a decision-support aid

Objective: Multi-attribute decision-making (MADM) models evaluate competing solutions for complex problems to identify the closest fit to the ideal solution. MADM models may assist dermatologists when selecting between biologics for plaque psoriasis. Here, is described the development of a pilot model to identify the preferred biologic from the dermatologist's perspective. Research design and methods: A group of European dermatologists were surveyed to identify treatment attributes they consider when prescribing a biologic. The relative importance of each was determined by allocation of 100 importance points in the context of seven case vignettes, reflecting the breadth of disease encountered in dermatological practice. Biologic performance was rated anonymously on a scale of 1-10, scores entered into a MADM matrix, and TOPSIS (Technique for Ordered Preference by Similarity to the Ideal Solution) analysis applied to identify the biologic closest to the hypothetical ideal. Results: Long-term efficacy and safety were the most important attributes considered by dermatologists when selecting a biologic. For one case vignette (chronic stable psoriasis), TOPSIS scores showed that etanercept was closest to the ideal for 63% of respondents, with adalimumab closest to the ideal for 32% of respondents. Differences among the biologics were highly significant (p<0.0001). For severe unstable psoriasis, infliximab and adalimumab were preferred. Limitation: This study was conducted with a group of dermatologists attending a Wyeth-sponsored advisory board meeting. Conclusions: Based on responses from this expert group, etanercept was the preferred choice for stable chronic plaque psoriasis for the majority, with infliximab preferred for more severe disease. However, there are several limitations to this pilot model, most notably the non-random selection of the expert group. Further development of the model encompassing a random survey of dermatologists and inclusion of other treatment alternatives and the latest clinical data, will add to the clinical utility of the tool.