Published Online: February 19, 2021
Abstract
Objective
To compare the accuracy of the whiplash specific risk-screening tool (WhipPredict) with a modified generic short-form Orebro Musculoskeletal Pain Screening Questionnaire (SF-ÖMPSQ-M) for predicting poor recovery.
Design
Inception cohort study.
Methods
202 participants with acute whiplash completed WhipPredict and SF-ÖMPSQ-M at baseline. Poor recovery was measured at 6 and 12 months: Neck Disability Index (> 10%), Numeric Pain Rating Scale (>3/10), global perceived recovery (< +4) and return to pre-injury work (RTW). Accuracy statistics for predicting poor recovery were calculated. A STARD Checklist was completed.
Results
71% (143) of participants completed the 6 month and 56% (113) the 12 month assessment. At baseline, agreement between WhipPredict and SF-ÖMPSQ-M was fair (Prevalence-adjusted and bias-adjusted kappa= 0.26; 95% 0.12 – 0.39). Depending on the outcome, 20-72% of people had not recovered at 6 months and 12-71% at 12 months. At 6 and 12 months, WhipPredict showed high sensitivity (88-94%) and low specificity (all <31%) on all 4 separate outcomes. The SF-ÖMPSQ-M showed moderate sensitivity (54-75%) for NDI, NRS and global perceived recovery, high sensitivity for RTW (81-90%) and moderate specificity (57-77%) for all 4 outcomes.
Conclusion
Both tools showed acceptable accuracy in predicting poor recovery. WhipPredict is recommended if the goal is to correctly identify patients who will not recover but will falsely classify some who will recover well. SF-ÖMPSQ-M will result in fewer patients falsely categorised as being at risk of poor recovery and may result in some people being undertreated. J Orthop Sports Phys Ther, Epub 19 Feb 2021. doi:10.2519/jospt.2021.9987