Abstract
The PSI, an assessment of severity of illness with 34 physiologic variables and 75 variable ranges, is becoming important in clinical research, cost containment, and quality assurance. This analysis simplifies the PSI to enable broader use. METHODS. 2642 patients from 9 PICUs were used. The sample was split into estimation (n = 1415, 116 deaths) and validation (n = 1227, 105 deaths) groups. A series of statistical techniques developed the PRISM score. The PRISM score performance was tested using goodness-of-fit tests and ROC analysis. RESULTS. Using the estimation group, the number of admission day, variables was reduced to 14 (all non-invasive) with 23 ranges. Performance evaluation in the validation set indicated the observed deaths were accurately predicted in all 6 PICUs in the validation group. (X2(5): 0.88, 1.05, 2.91, 4.95, 5.61, 9.44) as well as the total validation sample (X2(5) = 1.33). Overall, 105 deaths were observed and 108.9 were predicted. The PRISM score also performed extremely well in the all pt classification groups including: operative pts (N = 403, X2(5) = 5.26); nonoperative pts (N = 824, X2(5) = 3.32); cardiovascular pts (N = 229, X2(5) = 2.47); respiratory pts (N = 359, X2(5) = 7.71). The performance of the PRISM score as assessed by ROC analysis was excellent (area index = .92). CONCLUSION. The PRISM score is a greatly simplified version of the PSI score with equivalent performance. It is institutionally independent and not influenced by diagnosis. Supported by MCH Grant MCJ-11-527.
Article PDF
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Pollack, M., Ruttimann, U. & Getson, P. THE PEDIATRIC RISK OF MORTALITY (PRISM) SCORE: A SIMPLIFICATION OF THE PHYSIOLOGIC STABILITY INDEX (PSI). Pediatr Res 21 (Suppl 4), 206 (1987). https://doi.org/10.1203/00006450-198704010-00238
Issue Date:
DOI: https://doi.org/10.1203/00006450-198704010-00238