Prediction of 30-day mortality in older patients with a first acute myocardial infarction

This study sought predictors of mortality in patients aged >or=75 years with a first ST-segment elevation myocardial infarction (STEMI) and evaluated the validity of the GUSTO-I and TIMI risk models. Clinical variables, treatment and mortality data from 433 consecutive patients were collected. Univariable and multivariable logistic regression analyses were applied to identify baseline factors associated with 30-day mortality. Subsequently a model predicting 30-day mortality was created and compared with the performance of the GUSTO-I and TIMI models. After adjustment, a higher Killip class was the most important predictor (OR 16.1; 95% CI 5.7-45.6). Elevated heart rate, longer time delay to admission, hyperglycemia and older age were also associated with increased risk. Patients with hypercholesterolemia had a significantly lower risk (OR 0.46; 95% CI 0.24-0.86). Discrimination (c-statistic 0.79, 95% CI 0.75-0.84) and calibration (Hosmer-Lemeshow 6, p = 0.5) of our model were good. The GUSTO-I and TIMI risk scores produced adequate discrimination within our dataset (c-statistic 0.76, 95% CI 0.71-0.81, and c-statistic 0.77, 95% CI 0.72-0.82, respectively), but calibration was not satisfactory (HL 21.8, p = 0.005 for GUSTO-I, and HL 20.6, p = 0.008 for TIMI). In conclusion, short-term mortality in elderly patients with a first STEMI depends most importantly on initial clinical and hemodynamic status. The GUSTO-I and TIMI models are insufficiently adequate for providing an exact estimate of 30-day mortality risk.

Using multivariate sequential patterns to improve survival prediction in intensive care burn unit

Resuscitation and stabilization are key issues in Intensive Care Burn Units and early survival predictions help to decide the best clinical action during these phases. Current survival scores of burns focus on clinical variables such as age or the body surface area. However, the evolution of other parameters (e.g. diuresis or fluid balance) during the first days is also valuable knowledge. In this work we suggest a methodology and we propose a Temporal Data Mining algorithm to estimate the survival condition from the patient’s evolution. Experiments conducted on 480 patients show the improvement of survival prediction.