Context Only a patient and his or her family can judge many of the most important aspects of the doctor–patient interaction. This study evaluates the feasibility and reliability of children and their families assessing the quality of paediatricians' interactions using a rating instrument developed specifically for this purpose.
Methods A reliability analysis using generalisability theory on the ratings from 352 doctor–patient interactions across different speciality clinics.
Results Ratings were normally distributed. They were highest for ‘overall’ performance, and lowest for giving time to discuss the families' agenda. An appropriate sample of adults' ratings provided a reliable score (G = 0.7 with 15 raters), but children's ratings were too idiosyncratic to be reproducible (G = 0.36 with 15 raters).
Conclusions and further work Accompanying adults can provide reliable ratings of doctors' interactions with children. Because an adult is usually present at the consultation their ratings provide a highly feasible and authentic approach. Sampling doctors' interactions from different clinics and with patients of both genders provides a universal picture of performance. The method is ideal to measure performance for in-training assessment or revalidation. Further work is in progress to evaluate the educational impact of feeding ratings back to the doctors being assessed, and their use in a range of clinical contexts.