Investigating the meaning of 'good' or 'very good' patient evaluations of care in English general practice: a mixed methods study.

Objective: To examine concordance between responses to patient experience survey items evaluating doctors' interpersonal skills, and subsequent patient interview accounts of their experiences of care.

Design: Mixed methods study integrating data from patient questionnaires completed immediately after a video-recorded face-to-face consultation with a general practitioner (GP) and subsequent interviews with the same patients which included playback of the recording.

Setting: 12 general practices in rural, urban and inner city locations in six areas in England.

Participants: 50 patients (66% female, aged 19-96 years) consulting face-to-face with 32 participating GPs.

Main Outcome Measures: Positive responses to interpersonal skills items in a postconsultation questionnaire ('good' and 'very good') were compared with experiences reported during subsequent video elicitation interview (categorised as positive, negative or neutral by independent clinical raters) when reviewing that aspect of care.

Results: We extracted 230 textual statements from 50 interview transcripts which related to the evaluation of GPs' interpersonal skills. Raters classified 70.9% (n=163) of these statements as positive, 19.6% (n=45) neutral and 9.6% (n=22) negative. Comments made by individual patients during interviews did not always express the same sentiment as their responses to the questionnaire. Where questionnaire responses indicated that interpersonal skills were 'very good', 84.6% of interview statements concerning that item were classified as positive. However, where patients rated interpersonal skills as 'good', only 41.9% of interview statements were classified as positive, and 18.9% as negative.

Conclusions: Positive responses on patient experience questionnaires can mask important negative experiences which patients describe in subsequent interviews. The interpretation of absolute patient experience scores in feedback and public reporting should be done with caution, and clinicians should not be complacent following receipt of 'good' feedback. Relative scores are more easily interpretable when used to compare the performance of providers.