Introduction Informal carers are a valuable source of information on patients’ symptom experiences for clinicians, and carer assessment determines decisions regarding symptom management by carers themselves. However, previous studies have reported that proxies overestimate symptom burden, particularly subjective psychological and emotional symptoms, but proxy studies in COPD are few. We sought to assess congruency between patient and carer assessment of symptom burden, and to identify factors associated with incongruence.

Methods Well-characterised patients with advanced COPD and their carers (n = 117 patient-carer dyads) independently rated patients’ breathlessness, fatigue, constipation, diarrhoea, anxiety and depression on a 4-point scale, and average breathlessness in prior 24 hours using a Numerical Rating Scale (NRS). McNemar’s and Wilcoxon signed rank test were performed to identify differences between patients and carers in proportions reporting presence and reporting of severity of symptoms respectively. Intraclass correlation (ICC) was used to assess agreement on symptom scores in dyads.

Results Patient mean age was 71.4 (SD 8.7) and 62% were male; carer mean age was 64.2 (SD 14.5) and 27% were male. 87% of patients lived with their carer and 84% of carers were spouses. There were no significant differences between patients and carers in total proportions reporting presence or assessment of severity, of any symptom. ICCs (Table 1) showed patient-carer agreement was only fair to moderate. Higher agreement was found for physical symptoms (constipation, diarrhoea) than psychological (anxiety, depression) or those with emotional valence (breathlessness, fatigue). Carers more frequently underestimated than overestimated symptoms, with the exception of physical symptoms.

Conclusions Patient-carer agreement on symptom burden was generally low, and differed depending on symptom type. Poorer agreement for emotional symptoms and symptom underestimation by carers in this prospective population-based study may reflect patient concealment within dyads or the differences of a cohort recruited through primary care, compared to previous proxy studies in secondary care. These findings may also be due to longer disease trajectories in COPD, compared to previous studies in cancer, leading to carer compassion-fatigue or response-shift. Our findings have implications for the interpretation of proxy data in COPD, and suggest the need for carer education and support in symptom assessment.