Related Experiment Videos
The prognosis of medical coma in Ibadan: results of multivariate analysis
1Department of Paediatrics, University College Hospital, Ibadan, Nigeria.
Insights
Predicting medical coma outcomes in children is possible using multivariate analysis of early clinical data. This method accurately classified 92% of cases, aiding in prognosis for pediatric patients.
Area of Science:
- Neurology
- Pediatrics
- Biostatistics
Background:
- Medical coma in children presents a significant challenge in predicting patient outcomes.
- Early identification of prognosis is crucial for effective clinical management and resource allocation.
Purpose of the Study:
- To evaluate the utility of multivariate analysis in predicting the outcome of medical coma in pediatric patients.
- To identify key clinical variables that independently predict survival, neurological deficits, or death.
Main Methods:
- Prospective investigation of 116 children with medical coma.
- Application of stepwise logistic regression and discriminant analysis on 27 clinical variables recorded within 12 hours of admission.
- Development of a classification function to predict patient outcomes.
Main Results:
- Multivariate analysis identified 17 of 27 variables with independent significance for predicting outcome.
- The developed classification function achieved 92% accuracy in correctly classifying patient outcomes (survived intact, survived with deficits, died).
- A small subset of eight children (four survivors, four deaths) were significantly misclassified, highlighting potential limitations.
Conclusions:
- Multivariate analysis of early clinical data is a valuable tool for predicting medical coma outcomes in children.
- This approach can provide crucial predictive information for individual patients, guiding clinical decision-making.
- Further refinement of predictive models may improve accuracy and reduce misclassification rates.
Abstract:
During a recent prospective investigation of 116 children with medical coma in Ibadan, 52 (45%) survived intact, 23 (20%) developed residual neurological deficits and the remaining 41 (35%) died. This series is concerned with the use of multivariate analysis in predicting outcome of medical coma in an individual patient, based mostly on clinical information (27 variables) obtained within 12 hours of admission. Stepwise logistic regression analysis revealed that only 17 of the 27 variables have independent significance in predicting outcome. When discriminant analysis was performed on these 17 variables so as to obtain a classification function of outcome (survived intact, survived with deficits and dead), 92% of the cases were correctly classified. Eight children were, however, seriously misclassified. These were four intact survivors and four dead cases who were classified as dead and survived without neurological sequelae, respectively. These results suggest that multivariate analysis of clinical and laboratory information obtained in the comatose state can provide predictive information.