Desirability Scores for Assessing Multiple Outcomes in Systemic Rheumatic Diseases

Inflammatory rheumatic diseases like Rheumatoid Arthritis, Psoriatic Arthritis, and Systemic Sclerosis, generally are of the syndrome-type. For the assessment of these diseases, multiple outcomes are used. While a reason to combine outcomes may be to increase precision, the way in which outcomes are combined predominantly is an issue of validity. Multiple outcomes can be transformed to the same scale by desirability functions. We demonstrate how desirability scores are developed using expert opinion, and how they can be combined using the Lp-norm, using the example of Psoriatic Arthritis. The performance of the combinations of desirability scores is demonstrated in a randomised placebo-controlled trial of a tumour necrosis factor inhibitor. The principles that we used do generalize to other chronic inflammatory rheumatic diseases, such as Rheumatoid Arthritis and Systemic Sclerosis. They all have in common that multiple domains are involved and that an overall assessment on all these domains is required. Combination of outcomes may lead to better precision or to better validity.