OBJECTIVE: To compare patient-reported anxiety, depression and quality-of-life (QoL) outcomes, with data registered in patients' primary care electronic health record (EHR). DESIGN: Cross-sectional study. SETTING: Primary care in the UK. PARTICIPANTS: A convenience sample of 608 women registered in the Clinical Practice Research Datalink GOLD primary care database (data from a previous study on 356 breast cancer survivors (8.1 years postdiagnosis) and 252 women with no prior cancer). OUTCOME MEASURES: Patient-reported data on anxiety, depression and QoL, collected through postal questionnaires, and compared with coded information in EHR up to 2 years prior. RESULTS: Abnormal anxiety symptoms were reported by 118 of 599 women who answered the relevant questions (21%); 59/118 (50%) had general practitioner (GP)-recorded anxiolytic/antidepressant use, and 2 (1.6%) had anxiety coded in the EHR. 26/601 women (11%) reported depression symptoms, of whom 17 (65.4%) had GP-recorded antidepressant use and none had depression coded. 65 of 123 women reporting distress on the pain QoL domain (52.8%) had a corresponding record in the EHR <3 months before and 92 (74.8%) <24 months before. No patients reporting fatigue (n=157), sexual health problems (156), social avoidance (82) or cognitive problems (93) had corresponding codes in the EHR. There were no meaningful differences in the concordance results between breast cancer survivors and women with no history of cancer. CONCLUSION: Many patients reporting mental health and QoL problems had no record of this in coded primary care data. This finding suggests that coded data does not fully reflect the burden of disease. Further research is needed to understand whether or not GPs are aware of patient distress in cases where codes have not been recorded.