95 Quality assessment of patient reported outcome measures for patients with multiple ear complaints 5 is multidimensional and has a CFA for the impact questions only, they describe an EFA for the rest of the questionnaire of which the design model is formative. The only other study that elaborates on the EFA is the PAN-QOL18. The COMQ-1225 performed a principle component analysis (PCA), EFA and PCA are both variable reduction techniques but differ in their purpose. The structural validity of the COMQ-12 using the PCA was scored ‘very good’. Four studies scored ‘acceptable’, three scored ‘doubtful’ and two ‘inadequate’. Internal consistency was more frequently performed than the structural validity. Five questionnaires scored a ‘Very good’ with an Cronbach’s Alfa > 0.7. Six questionnaires scored ‘Acceptable’ and one ‘Inadequate’. Five questionnaires have their questions translated in one or more languages. The method of translation is often described and acceptable. The studies discuss the validity of the questionnaire in a different population and good or valid, but none of the ‘translations’ has actually done interviews with patients to see whether adaptations needed to be made on the original questionnaire based on relevance, comprehensiveness or comprehensibility. None of the studies developed their questionnaire using an Item Response Theory (IRT) model. Remaining properties Measurement error is part of reliability and important order to interpret the data correctly. Repeated measurements may display variation arising from several sources leading to measurement. A Standard Error of Measurement (SEM), Smallest Detectable Change (SDC) or Limits of Agreement (LoA) was not calculated in any of the studies. Although not mentioning a SEM, SDC or LoA four studies had an ‘Acceptable’ measurement error. Eight studies had ‘Inadequate’ measurement error reporting. Test-retest is the second part of reliability. Test-retest period varies greatly between the questionnaires ranging from 1-2 days up until 6 months. Eight out of twelve studies mentioned their Cronbach’s alpha. Cronbach’s alpha varied between 0,761 and 0,91. Interpretability and feasibility Distribution of scores and subscores and change scores of all relevant (sub)group were often mentioned in the articles addressing the PROM. The other items of interpretability were often lacking. In specific a minimal important change or -difference was often not calculated or mentioned. Considering these limitations interpretability was scored ‘doubtful’ in five questionnaires, ‘adequate’ in six
RkJQdWJsaXNoZXIy MjY0ODMw