Development and External Validation of Individualized Prediction Models for Pain Intensity Outcomes in Patients With Neck Pain, Low Back Pain, or Both in Primary Care Settings.
Lucinda Archer, Kym I E Snell, Siobhán Stynes and 6 others
PMID 37756617WHAT IT FOUND
Prediction models using routine questionnaire data failed to accurately forecast pain intensity at the point of consultation.
They systematically overpredicted pain at 6 months and underpredicted it at 2 months. These tools are not accurate enough for clinical use yet.
Key findings
01Models using data collected at the point of consultation performed poorly, with calibration slopes between 0.44 and 0.65 for pain outcomes.
02The 6-month pain intensity model systematically overpredicted scores by an average of 0.93 points on a 0-10 scale when using consultation data.
03Discrimination was low, with C statistics around 0.65 for binary pain outcomes when using point-of-consultation data.
STILL TO COME
How it was doneWhat they found
Read the rest of this summary
You get three full summaries a month, free, and we do not ask for a card. Search, the TL;DRs and your library stay unlimited either way.
What it does not show
The models were limited to self-report items from the Keele STarT MSK Tool and did not include clinical examination findings or imaging. The external validation sample size was smaller than required for precise estimation, leading to uncertainty in performance estimates. Neck pain and low back pain patients were combined due to low numbers of neck pain patients, meaning results are weighted toward low back pain presentations. The models only predicted pain intensity, not disability or other functional outcomes.
Declared interests
Funded by European Horizon 2020, NIHR, Versus Arthritis, and other UK health research bodies. The funders played no role in the design, conduct, or reporting of the study.
The easy way to misread this
Do not assume that risk stratification tools like the Keele STarT MSK Tool can accurately predict an individual patient's specific future pain score. While useful for grouping patients by risk, these models failed to accurately forecast exact pain intensity levels when used during the consultation.