Psychometric properties of the Patient Reported Outcomes, Burdens and Experiences (PROBE) questionnaire

Objective To assess the psychometric properties of the Patient Reported Outcomes, Burdens and Experiences (PROBE) questionnaire. Methods This study was a cross-sectional, multinational study. Participants were enrolled if they were more than 10 years old and people with haemophilia A or B or people without a bleeding disorder. Participants were invited through non-governmental patient organisations in 21 countries between 01/27/2016 and 02/23/2017. The following psychometric properties: missing data, floor and ceiling effects, exploratory factor analysis and internal consistency reliability were examined. A PROBE Score was derived and assessed for its convergent and known groups validity. Results The study analysed the data on 916 participants with median age of 37.0 (IQR 27.0 to 48.0) years, 74.8% male. In the domain assessing patient-reported outcomes (PROs), more than 15% of participants presented a ceiling effect for all items but two, and a floor effect for one item. Factor analysis identified three factors explaining the majority of the variance. Cronbach's alpha coefficient indicated good internal consistency reliability (0.84). PROBE items showed moderate to strong correlations with corresponding EuroQol five dimension 5-level instrument (EQ-5D-5L) domains. The PROBE Score has a strong correlation (r=0.67) with EQ-5D-5L utility index score. The PROBE Score has a known groups validity among various groups. Conclusions The results of this study suggest that PROBE is a valid questionnaire for evaluating PROs in people with haemophilia as well as control population. The known-group property of PROBE will allow its use in future clinical trials, longitudinal studies, health technology assessment studies, routine clinical care or registries. Additional studies are needed to test responsiveness and sensitivity to change. Trial registration number NCT02439710; Results.