Let's talk about pain catastrophizing measures: an item content analysis

Key points

1. 0194 participants rated how relevant each questionnaire item was to pain catastrophizing versus five other pain-related constructs.
2. 02None of the six pain catastrophizing measures distinctively captured catastrophizing, defined as viewing or presenting pain as considerably worse than it actually is.
3. 03Items from worrying, vigilance, severity, distress, and disability scales did map distinctively onto their own constructs, unlike the catastrophizing items.
4. 04The catastrophizing scales overlapped most with worrying about pain and pain-related distress.
5. 05Authors propose a person-centered approach and renaming the tools as 'pain-related worrying' rather than building yet another catastrophizing questionnaire.

How it was conducted

Design

Cross-sectional online discriminant content validity study

Participants

94 adults recruited via Prolific, English first language, age 18 or older

Items rated

58 items from six pain catastrophizing measures plus 34 contrast items from worrying, vigilance, severity, distress, and disability scales (87 items in three randomized sets)

Constructs compared

Pain catastrophizing, worrying about pain, pain vigilance, pain severity, pain-related distress, pain-related disability

Rating task

Yes/no relevance plus 0 to 100% confidence per construct per item, scored from -10 to +10

Analysis

Bayesian hierarchical models (JAGS in R), 95% credibility intervals, 5% two-sided significance

What they found

• Of 138 participants collected, 44 were excluded (1 attention-check failure, 36 unreliable scoring, 2 non-native English, 4 under 15 minutes, 1 requested exclusion), leaving n=94 (mean age 36, SD 12; 63% male).
• CSQ scored highest on the catastrophizing construct and AEQ second highest, not significantly lower than CSQ (l=0.13, 95% CI [-1.25 to 1.52]); CCSI, PRSS, PCS, and PCL scored significantly lower.
• The PCL scored significantly below 0 on catastrophizing, meaning its items were not endorsed as measuring catastrophizing at all.
• AEQ scored higher on catastrophizing than disability (l=10.41), severity (l=8.64), and vigilance (l=6.65), but lower than worrying (l=-1.68) with no difference from distress (l=0.58).
• CSQ scored higher on catastrophizing than disability (l=6.25), severity (l=2.29), and vigilance (l=8.23), but lower than distress (l=-2.93) with no difference from worrying (l=0.30).
• PCS scored higher than disability (l=5.12) and vigilance (l=3.77) but lower than distress (l=-4.74) and worrying (l=-3.90).
• The highest-scoring item, 'When I'm in pain, it's terrible and I think it's never going to get any better', rated catastrophizing 7.41 [5.36 to 9.44] but also distress 7.21 and worry 6.21, showing it did not distinctively measure catastrophizing.

Limitations

• Raters were online non-expert participants rather than clinicians or pain experts, and no cognitive interviewing was done.
• Construct definitions were adapted from the Oxford Living Dictionaries, and the worrying and distress constructs themselves could not be clearly distinguished by raters.
• Only content validity was examined, not construct or criterion validity, and not all catastrophizing instruments were included.
• The sample of 94 raters was modest and skewed toward people with low current pain (60% in the lowest pain grade).

Why it matters

For patients

A high score on a 'pain catastrophizing' questionnaire likely reflects worry or distress about pain rather than an exaggeration of it, so the label should not be taken as a judgment of your thinking.

For clinicians

Treat existing pain catastrophizing scores as indicators of pain-related worry or distress, and interpret the 'catastrophizing' label cautiously when communicating with patients.

For readers

This study questions whether one of pain research's most-used constructs can be measured by self-report at all, suggesting the field rethink both its tools and its terminology.