Variability of "optimal" cut points for mild, moderate, and severe pain: neglected problems when comparing groups

Area of Science:

• Pain Medicine
• Biostatistics
• Pediatric Pain Research

Background:

• Defining pain intensity cut points (mild, moderate, severe) based on functional interference is common.
• The Serlin et al. (1995) statistical method for deriving these cut points is widely adopted.
• Interpreting differing cut points between chronic pain populations as meaningful differences is a frequent practice.

Purpose of the Study:

• To explore the variability of optimally defined pain intensity cut points.
• To assess this variability in a large chronic pain sample and in homogeneous subsamples.
• To question the interpretation of inter-group differences in cut points without considering variability.

Main Methods:

• Collected maximal pain intensity (0-10 numeric rating scale, NRS) and pain-related disability data.
• Determined optimal cut points for the entire sample of 2249 children with chronic pain.
• Quantified cut point variability using bootstrap techniques in the full sample and in subsamples (constant pain, chronic daily headache, musculoskeletal pain).

Main Results:

• Optimal cut points for mild, moderate, and severe pain in the total sample were 4 and 8 on the 0-10 NRS.
• High variability was observed; optimal cut points were identified only 40% of the time within the whole sample.
• Similar high variability in cut points was found across homogeneous pediatric pain subsamples.

Conclusions:

• Optimal cut points are significantly influenced by random sample fluctuations.
• Observed differences in optimal cut points between study groups may be attributable to chance variation.
• Future research comparing groups must incorporate measures of optimal cut point variability.