Derivation of a bedside score (MASH-P) to predict 6-month mortality in tuberculous meningitis

Background: Tuberculous meningitis is commonly associated with a poor outcome. Simple bedside prognostic scores can help immensely in predicting the outcome. Materials and method: A total of 721 patients, from 5 of our previous studies, were included. With primary outcome measure as death, a prognostic model was derived using binary logistic regression. The model was assessed using discrimination and calibration, and internally validated using the bootstrap method. A bedside prognostic score was derived by rounding of the regression coefficients to the nearest integers. Results: A total of 126 (17.48%) patients died. The final model found that higher age, stage III disease, baseline MBI <= 12, papilledema and hydrocephalus were significant predictors of death. The final model showed good discrimination as evident by an AUC = 83.1% (95% confidence interval 79.5%-86.7%, P < .001) and good calibration (Hosmer and Lemeshow test P = .579). The model remained valid after internal validation by boot strapping. A simple bedside score with the acronym MASH-P to denote variables baseline MBI (M), age (A), stage (S), hydrocephalus (H) and papilledema (P), was thus derived. The score can range from 0 to 10. Higher the score, higher is the probability of death; a score of 0 carries a predicted probability of just 1.7% while a score of 10 corresponds to a predicted probability of 65%. An electronic ready reckoner has also been developed to aid prognostication on the go. Conclusion: MASH-P is a simple prognostic scoring model that can be used at bedside and aid in decision making as well as counselling.