Premium
The PAC ‐ SYM questionnaire for chronic constipation: defining the minimal important difference
Author(s) -
Yiannakou Y.,
Tack J.,
Piessevaux H.,
Dubois D.,
Quigley E. M. M.,
Ke M. Y.,
Da Silva S.,
Joseph A.,
Kerstens R.
Publication year - 2017
Publication title -
alimentary pharmacology and therapeutics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 3.308
H-Index - 177
eISSN - 1365-2036
pISSN - 0269-2813
DOI - 10.1111/apt.14349
Subject(s) - medicine , constipation , receiver operating characteristic , minimal clinically important difference , functional constipation , mean difference , significant difference , chronic constipation , clinical trial , randomized controlled trial , confidence interval
Summary Background The Patient Assessment of Constipation‐Symptoms ( PAC ‐ SYM ) questionnaire is frequently used in clinical trials of constipation. However, the threshold for reduction in total PAC ‐ SYM score used to define a clinical response on this 0‐4 point scale has not undergone formal appraisal, and its relationship with clinical benefit as perceived by patients has not been defined. Aim To determine the minimal important difference in PAC ‐ SYM score, and the optimum cut‐off value for defining responders. Methods The minimal important difference was estimated using data from six international phase 3/4, double‐blind, randomised controlled trials of prucalopride in patients with chronic constipation ( NCT 01147926, NCT 01424228, NCT 01116206, NCT 00485940, NCT 00483886, NCT 00488137), with anchor‐ and distribution‐based approaches. Five appropriate patient‐reported outcomes were selected as anchors. In addition, receiver operating characteristics ( ROC ) curve analyses were used to investigate responder discrimination for each anchor. Results Data from 2884 patients were included. Minimal important difference estimates ranged from –0.52 to –0.63 across the five anchors. Estimates were not affected by study location but were consistently lower for rectal symptoms than for abdominal and stool symptoms. Distribution‐based estimates were considerably lower than anchor‐based estimates. ROC curve analyses showed optimum cut‐off scores for discriminating responders to be similar to anchor‐based minimal important difference estimates. Conclusions Anchor‐based methods gave consistent results for the minimal important difference, at approximately –0.6, and this value was close to the ROC ‐determined optimal cut‐off scores for responder discrimination. This value could be considered in clinical practice. A slightly more conservative threshold (eg –0.75) could be used in clinical trials to reduce the placebo response rate.