Associations between diagnostic activity and measures of patient experience in primary care: A cross-sectional ecological study of English general practices

29 Sep 2017

Background: Lower use of endoscopies and urgent referrals for suspected cancer has been linked to poorer outcomes for cancer patients; it is important to examine potential predictors of variable use. Aim: To examine associations between general practice measures of patient experience and practice use of endoscopies or urgent referrals for suspected cancer. Design and Setting: Cross-sectional ecological analysis; English general practices Method: We used data from the General Practice Patient Survey and the Cancer Services Public Health Profiles. After adjustment for practice population characteristics, we examined practice-level associations between use of endoscopy and urgent referrals for suspected cancer, and ability to book an appointment (used as proxy for ease of access), ability to see a preferred doctor (used as proxy for relational continuity), and doctor/nurse communication skills. Results: Taking into account practice scores for ability to book an appointment, practices rated higher for our proxy measure of relational continuity used urgent referrals and endoscopies less often (e.g. 30% lower urgent referral and a 15% lower gastroscopy rates between practices in the 90th/10th centiles, respectively). In contrast, practices rated higher for doctor communication skills used urgent referrals and endoscopies more often (e.g. 26% higher urgent referral and 17% higher gastroscopy rates, between practices in the 90th/10th centiles, respectively). Cancer patients of practices rated higher for doctor communication skills were less likely to be diagnosed as emergencies (1.7% lower between practices in the 90th compared to the 10th centile). Conclusions: Practices where patients rate doctor communication highly were more likely to investigate and refer patients urgently but, in contrast, practices where patients could see their preferred doctor more readily were less likely to do so. We discuss the possible implications of these findings for clinical practice.