Outcome after treatment of neurological decompression illness is predicted by a published clinical scoring system

After recompression therapy, some cases of neurological decompression illness (DCI) have a significant residual deficit. Boussuges et al. report a scoring system to predict sequelae using weighted historical and clinical indices at presentation which we applied to the British Hyperbaric Association (BHA) database of UK diving accidents, held at the Institute of Naval Medicine (INM).

A database search identified 234 cases of neurological DCI from the 2 yr, 1995-6. Of these, 217 case records contained sufficient data to apply the scoring system. Outcome was classified as severe sequelae (i.e., causing a functionally important deficit) or as mild/no sequelae.

The median score in UK cases with severe sequelae was 13 (95% C.I. 11.5 to 14.5) and in cases with mild/no sequelae, 6 (95% C.I. 5.5 to 6.5). Significantly more cases with scores > 7 had severe sequelae than cases with scores < or = 7 (chi2, p < 0.0001). The sensitivity of a score > 7 (for severe sequelae) was 94% and the specificity was 65%. The positive predictive value of a score > 7 (for severe sequelae) was 18% and the negative predictive value of a score < or = 7 was 99%.

When applied to the INM/BHA database a score of > 7 by this scoring system has a higher sensitivity and lower specificity for severe sequelae than reported by Boussuges et al. It has a higher predictive accuracy for successful outcome of treatment (99% vs. 89%) but a much lower predictive value for severe sequelae (18% vs. 86%). Convergence between this and other published scoring systems may allow derivation of a generic scoring system that could then be evaluated prospectively in multiple centers.