Evaluating acute pain intensity relief: challenges when using an 11-point numerical rating scale.

Abstract

Percentage of pain intensity difference (PercentPID) is a recognized way of evaluating pain relief with an 11-point numerical rating scale (NRS) but is not without flaws. A new metric, the slope of relative pain intensity difference (SlopePID), which consists in dividing PercentPID by the time between 2 pain measurements, is proposed. This study aims to validate SlopePID with 3 measures of subjective pain relief: a 5-category relief scale (not, a little, moderate, very, complete), a 2-category relief question ("I'm relieved," "I'm not relieved"), and a single-item question, "Wanting other medication to treat pain?" (Yes/No). This prospective cohort study included 361 patients in the emergency department who had an initial acute pain NRS > 3 and a pain intensity assessment within 90 minutes after analgesic administration. Mean age was 50.2 years (SD = 19.3) and 59% were women. Area under the curves of receiver operating characteristic curves analyses revealed similar discriminative power for PercentPID (0.83; 95% confidence interval [CI], 0.79-0.88) and SlopePID (0.82; 95% CI, 0.77-0.86). Considering the "very" category from the 5-category relief scale as a substantial relief, the average cutoff for substantial relief was a decrease of 64% (95% CI, 59-69) for PercentPID and of 49% per hour (95% CI, 44-54) for SlopePID. However, when a cutoff criterion of 50% was used as a measure of pain relief for an individual patient, PercentPID underestimated pain-relieved patients by 12.1% (P < 0.05) compared with the SlopePID measurement, when pain intensity at baseline was an odd number compared with an even number (32.9% vs 45.0%, respectively). SlopePID should be used instead of PercentPID as a metric to evaluate acute pain relief on a 0 to 10 NRS.