Mortality and readmission risk can be predicted by the record-based Multidimensional Prognostic Index: a cohort study of medical inpatients older than 75years.

Abstract

To examine the predictive value of the record-based Multidimensional Prognostic Index (MPI) on mortality, readmission and length of hospital stay (LOS) among older medical inpatients. A cohort of medical inpatients aged ≥ 75years was rated using the record-based MPI to assess frailty retrospectively. 90-day and 1-year mortality hazard ratios (HRs) were calculated in a sex- and age-adjusted Cox proportional hazards model. 30-day readmission relative risk (RR) estimates were calculated in a binary regression model with mortality as a competing risk. Discrimination was expressed by the area under the receiver operating characteristic (ROC) curve. Median LOS was calculated using the non-parametric Kruskal-Wallis one-way ANOVA. In total, 1190 patients with a median age of 83years were included. 50% were male. 335 patients (28%) were categorized as non-frail (MPI score 0.0-0.33), 522 (44%) moderately frail (MPI score 0.34-0.66) and 333 (28%) severely frail (MPI score 0.67-1.0). 90-day mortality HR was 7.4 (95% confidence interval (CI) 2.9-18.6, p < 0.001) for the moderately frail and 18.5 (95% CI 7.5-46.1, p < 0.001) for the severely frail compared with the non-frail. ROC area was 0.76 (95% CI 0.72-0.80). Similarly, 1-year mortality HR was 3.3 (95% CI 2.2-5.0, p < 0.001) for the moderately frail and 7.1 (95% CI 4.7-10.6, p < 0.001) for the severely frail. 30-day readmission RR was 2.1 (95% CI 1.5-2.9, p < 0.001) for the moderately frail and 1.8 (95% CI 1.3-2.6, p = 0.001) for the severely frail. LOS was significantly longer with increasing MPI score (p < 0.001). The record-based MPI assessed at discharge predicts dose-dependent post-discharge mortality and readmission risk and is associated with LOS in older medical inpatients.