Is the English Cancer Patient Experience Survey representative? A comparative analysis with the National Lung Cancer Audit

Objectives: Healthcare systems increasingly recognise the importance of service users' perspectives for improving care organisation and delivery. The English Cancer Patient Experience Survey (CPES) is carried out annually, however, its representativeness within cancer types is unknown. We have explored if the CPES results are representative of people with lung cancer. Materials and methods: We linked cancer registry data across multiple sources to assess how CPES represents sociodemographic and clinical characteristics of the National Lung Cancer Audit population, accounting for post-sampling mortality bias. Multivariable logistic regression was used to compare people included and not included in CPES. Results: Of 240,375 people diagnosed (2009-2015), 15,967 (7 %) were included in CPES. Gender and ethnicity were reasonably represented, as were sociodemographic and clinical groupings, although more received anti-cancer treatment (96 % of CPES respondents vs. 56 % of patients nationally; adjusted odds ratio = 10.3, 95 confidence interval 9.4-11.2 for any anti-cancer treatment) with chemotherapy most over-represented, followed by surgery and then radiotherapy. CPES under-represented older, more socioeconomically deprived, and certain clinical groups, including those with worse performance status, multiple comorbidities, and diagnosis via emergency presentation. Conclusion: CPES includes patients across the sociodemographic and clinical spectrum indicating its value for research and service planning. Unbalanced representation of incident lung cancer cases is a limitation that must be considered in context of using CPES findings to implement service changes. Although half the national lung cancer population who received no anti-cancer treatment do not have their experiences represented, the strength of this dataset is in providing detailed comparisons of patient experiences across different treatment groups.