Development of a generic fidelity measure for rehabilitation intervention research for children with physical...

To increase research rigour and create a plausible way to assess clinical effectiveness, it is necessary to measure the degree to which interventions are delivered as intended (fidelity). Generic fidelity measures enable evaluation of more than one intervention through observation of unique and general characteristics relevant across interventions. This study describes the first generic fidelity measure in paediatrics.

Items were constructed from multiple sources to create a general attributes domain and two paediatric cerebral palsy (CP) intervention-specific domains. After a pre-testing procedure, raters were trained and videos of clinical interventions were rated to test the measure's psychometric properties.

The Paediatric Rehabilitation Observational measure of Fidelity (PROF) consisted of 30 items. Six raters were trained on the PROF and rated 25 videos. Internal consistency (α) and interrater reliability (IRR) for the frequency scale showed the following results: context therapy: α=0.71, IRR=0.75; child therapy: α=0.85, IRR=0.87; and general attributes; α=0.78, IRR=0.82. Quality scale scores across domains demonstrated internal consistency greater than 0.80 and interrater reliability of less than 0.40. Pearson's correlations (r=-0.71, p<0.001) and analyses of variance (p=0.01) validated that each intervention domain was an independent construct.

The PROF is reliable and valid for evaluating interventions used for children with CP. Future studies may use the measure's framework, general attributes domain, and procedures to test the psychometric properties of other interventions.