What is the association between patient experience questionnaires and patient complaints? An analysis of routine data.

Problem

Recorded written complaints against the NHS are on the rise; on average, over 4000 complaints are made every week in England. Of these, 43% concern primary care (comprising general practice and dentistry). Whilst complaints are still relatively rare compared to the number of patient contacts in primary care, they represent a failure of the service in the eyes of the patients involved. Receiving a written complaint can also have a devastating impact on the primary care staff concerned. By contrast, patient experience reported through mechanisms such as the national GP Patient Survey (GPPS) tends to be very positive. Whilst there is evidence that some patients may be reluctant to report poor quality care on patient experience surveys, the large representative sample of the GPPS means that reliable practice-level assessments of experience are obtained. What is unknown is to what degree GPPS scores may predict the rate of complaints made against general practices.

Approach

Data were obtained from NHS Digital on written complaints made in each general practice in the year 2015/16. These data were linked to the GPPS data at practice level. Case-mix adjusted practice scores for four key areas of patient experience, as well as overall satisfaction, were created using previously established methodology. Mixed effects Poisson regression was used to examine the association between the rate of complaints and GPPS scores. We used separate models to look at complaints flagged as either medical, practice administration, other, communication/attitude or premises. Univariable analyses were first used to compare complaints in one area to one patient experience measure, and subsequently multivariable models adjusted for all patient experience measures together.

Findings

In univariable analysis higher scores on all aspects of patient experience were associated with fewer complaints in all areas. For example, practices at the 97.5th percentile of practice scores for overall satisfaction had a 52% lower rate of medical complaints than those at the 2.5th percentile (RR 0.48, 95%CI 0.44, 0.52, p<0.0001). When adjusting for all patient experience items, different aspects of patient experience predicted different types of complaint. For example, practices with high scores on the ability to see a preferred doctor and GP communication had lower rates of medical complaints, whereas for complaints about practice administration, ability to see a preferred doctor and the helpfulness of receptionists were the strongest predictors.

Consequences

Patient experience, as measured by the GPPS, is strongly associated with the rate of recorded written complaints made against general practices. This provides robust evidence for the construct validity of patient experience measured with a survey. Furthermore, our findings suggest that by taking action in practices known to have poor patient experience scores it may be possible to reduce the volume of complaints made against primary care.