Abstract
INTRODUCTION: The Numerical Rating Scale (NRS) is frequently used to assess patient-reported breathlessness in both a research and clinical context. A subgroup of patients report average breathlessness as worse than their worst breathlessness in the last 24 hours (paradoxical average). The Peak/End rule describes how the most extreme and current breathlessness influence reported average. This study seeks to highlight the existence of a subpopulation who give 'paradoxical averages using the NRS, to characterise this group and to investigate the explanatory relevance of the 'Peak/End' rule. METHODS: Data were collected within mixed method face-to-face interviews for three studies: the Living with Breathlessness Study and the two subprotocols of the Breathlessness Intervention Service phase III randomised controlled trial. Key variables from the three datasets were pooled (n=561), and cases where participants reported a paradoxical average (n=45) were identified. These were compared with non-cases and interview transcripts interrogated. NRS ratings of average breathlessness were assessed for fit to Peak/End rule. RESULTS: Patients in the paradoxical average group had higher Chronic Respiratory Questionnaire physical domain scores on average p=0.042). Peak/End rule analysis showed high positive correlation (Spearman's rho=0.756, p<0.001). CONCLUSIONS: The NRS requires further standardisation with reporting of question order and construction of scale used to enable informed interpretation. The application of the Peak/End rule demonstrates fallibility of NRS-Average as a construct as it is affected by current breathlessness. Measurement of breathlessness is important for both clinical management and research, but standardisation and transparency are required for meaningful results.