Three little words: an empirical test of the optimum scoring method for the RCP 3 questions

Abstract

Background:The Royal College of Physicians three questions are widely used in clinical practice and research for assessing the impact of asthma on individuals. These simple questions assess core experiences of asthma assessing impact on sleep, daily symptom experiences and interference with normal functioning. The 3 questions are usually scored categorically by eliciting a ‘Yes’ vs. ‘No’ response to each question. Responses are scored 0 or 1 giving a scale score of 0–3 (RCP 0–3). This has the advantage of simplicity and ease of completion. However, the lack of differentiation in the response frame may compromise responsiveness to change. Small but clinically significant changes, such as an improvement from 3 to 1 night of disturbed sleep in a week, might not register if the patient ticks ‘yes’ to indicate that their sleep is still disturbed by asthma. For this reason an alternative response frame has been developed where the patient is asked to indicate the number of times over the past week they have experienced each of the three asthma ‘impacts'. In this method responses are scored 0–7 for each item giving a scale range of 0–21 (RCP 0–21). The aim of the study was to compare the RCP (0–3) against the RCP (0–21) in terms of patient acceptability, sensitivity and concurrent validity. Method: In a cross-sectional survey of community-managed patients at Step 2 or 3 of the asthma guidelines, participants were asked to complete a study questionnaire incorporating the RCP (0–3), RCP (0–21), Juniper's Asthma Control Questionnaire (ACQ) and a Medication Adherence Report Scale assessing adherence to inhaled corticosteroids (ICS). The acceptability of the RCP (90–21) was assessed according to its performance on 3 criteria:1. Equivalent patient acceptability: no significant differences in number of incomplete responses to RCP questions.2. Equivalent validity as a measure of asthma control: correlations between the RCP (0–21) and ACQ should be of a similar order to correlations between the RCP (0–3) and ACQ.3. Sensitivity to treatment: as a preliminary test we hypothesised that, compared with the RCP (0–3), the RCP (0–21) would be more strongly correlated with reported adherence to ICS.Results:The study questionnaire was completed by 78 community-managed patients at Step 2 or 3 of the asthma guidelines. The was no difference in response rates between RCP (0–3) and RCP (0–21), suggesting equivalent patient acceptability. The RCP (0–21) demonstrated high concurrent validity. Both RCP (0–21) and RCP (0–3) were highly correlated with the ACQ but the RCP (0–21) attaining a numerically higher correlation and stronger relationship with ACQ (RCP (0–21) r= 0.87: p< 0.001; RCP (0–3) r = 0.73; p < 0.001). A statistically significant correlation was found between the RCP (0–21) and reported adherence to ICS (r= 0.24; p< 0.05) wheras reported adherence to ICS was not correlated with the RCP (0–3), suggesting that the RCP (0–21) may be more sensitive to the effects of treatment.Conclusion:These data provide preliminary evidence in support of the RCP(0–21) scoring system. Using a 0–7 day response frame, in place of the 0–1 score appears to be equally acceptable to patients and equally valid. The RCP (0–21) may confer additional advantages in sensitivity and responsiveness to change. The findings justify further work to evaluate the RCP (0–21).