Table 3.
Identification of patients for ‘fast-track’ referrals to spinal specialist services. Three scenarios investigated
| Sensitivity (95% CIa) | Specificity (95% CI) | Positive predictive value (95% CI) | Negative predictive value (95% CI) | % of total sample referred | |
|---|---|---|---|---|---|
| Use of only clinical characteristics (total of 4b): | |||||
| - Patients with 3 out of 4 clinical characteristics | 68% (55–80) | 60% (55–65) | 21% (15–27) | 93% (88–96) | 43% (186/429) |
| - Patients with all 4 clinical characteristics | 32% (20–45) | 89% (85–92) | 30% (19–44) | 89% (86–92) | 14% (59/429) |
| Use of clinical characteristics combined with STarT Back prognostic risk score (medium, high)c | |||||
| - patients at ‘high’ risk on STarT Back, with 3 or more clinical characteristics | 51% (37–64) | 73% (68–78) | 22% (16–31) | 91% (87–94) | 30% (129/429) |
| OR | |||||
| - patients at ‘medium’ risk on STarT Back, with all 4 clinical characteristics | |||||
aCI; Confidence Intervals
bCurrent leg pain > 6 (NRS 0–10), pain below knee, interference with ability to do work/home activities > 6 (NRS 0–10) or ‘yes’ response on the RMDQ item, sensory deficit in painful leg with pin/prick testing
cPatients with a low risk STarT Back score were excluded from this analysis, as from the ATLAS cohort, only one patient at low risk of poor outcome was referred to spinal specialist services