Determining minimally important differences for the PDQ-39 Parkinson's disease questionnaire

Objective: to determine minimally important differences for dimensions of t he PDQ-39, a 39-item Parkinson's disease questionnaire. A minimally importa nt difference is defined as the smallest change between two scores that is subjectively meaningful to patients. Data on minimally important difference s are essential for the calculation of sample sizes in trials and surveys. Methods: we conducted a postal survey of randomly selected members of 13 lo cal branches of the Parkinson's Disease Society, asking them to complete th e PDQ-39 on two occasions, 6 months apart. On the first occasion respondent s received the PDQ-39, demographic questions and a request to provide their name and address if they were willing to take part in the follow-up survey . After 6 months, we sent those who had agreed another copy of the question naire and also asked them to indicate how much change they had experienced since baseline in overall health and in each of the eight domains of the qu estionnaire. Results: we calculated minimally important difference for each dimension an d the index score for those reporting minor change since baseline. The mini mally important difference varied across dimensions. Conclusions: these results indicate the minimum magnitude of change that sh ould be sought when designing studies to evaluate change over time in Parki nson's disease. Since minimally important differences differ across dimensi ons, those designing studies in which sample size Calculations are based on the PDQ-39 as an outcome measure should select the dimension which is the primary variable of interest.