Mortality Predictions on Admission as a Context for Organizing Care Activities

Mark E. Cowen, MD, SM; Robert L. Strawderman, ScD; Jennifer L. Czerwinski, BA; , Mary Jo Smith, RN, MS; Lakshmi K. Halasyamani, MD

Disclosures

Journal of Hospital Medicine. 2013;8(5):229-235. 

In This Article

Results

Table 1 displays the risk factors used in the 30-day mortality prediction rule, their distribution in the populations of interest, and the frequency of the outcomes of interest. The derivation (D1) and validation (V1) populations were clinically similar; the patients of hospital V2 differed in the proportion of risk factors and outcomes. The scoring weights or parameter estimates for the risk factors are given in the Appendix (see Supporting Information, Appendix I, in the online version of this article).

Predicting 30-Day Mortality

The areas under the ROC (95% confidence interval [CI]) for the D1, V1, and V2 populations were 0.876 (95% CI, 0.870–0.882), 0.885 (95% CI, 0.877–0.893), and 0.883 (95% CI, 0.875–0.892), respectively. The calibration curves for all 3 populations are shown in Figure 1. The overlap of symbols indicates that the level of predicted risk matched actual mortality for most intervals, with slight underprediction for those in the highest risk percentiles.

Figure 1.

Calibration. The horizontal axis displays 20 intervals of risk, containing 5-percentile increments of the predicted mortality based on the D1 population. The vertical axis displays the actual proportion of patients within the interval who died within 30 days. The cluster of 3 symbols represent the mean predicted chance of dying for the derivation and 2 validation populations, respectively. The crosshatches represent the actual proportion of patients within each interval who died, with the 95% binomial confidence limits represented by the length of the vertical bar. The 20 intervals (named for the highest percentile within the interval) with corresponding probabilities of death: 5th percentile (probability 0–0.0008); 10th percentile (probability 0.0008–0.0011); 15th percentile (probability 0.0011–0.0021); 20 (0.0021–0.0033); 25 (0.0033–0.0049); 30 (0.0049–0.0067); 35 (0.0067–0.0087); 40 (0.0087–0.0108); 45 (0.0108–0.0134); 50 (0.0134–0.0165); 55 (0.0165–0.0201); 60 (0.0201–0.0247); 65 (0.0247–0.0308); 70 (0.0308–0.0392); 75 (0.0392–0.0503); 80 (0.0503–0.0669); 85 (0.0669–0.0916); 90 (0.0916–0.1308); 95 (0.1308–0.2186); 100 (0.2186–1.0).

Example of Risk Strata

Figure 2 displays the relationship between the predicted probability of dying within 30 days and the outcomes of interest for V1, and illustrates the Pareto principle for defining high- and low-risk subgroups. Most of the 30-day deaths (74.7% of D1, 74.2% of V1, and 85.3% of V2) occurred in the small subset of patients with a predicted probability of death exceeding 0.067 (the top quintile of risk of D1, the top 18 % of V1, and the top 29.8% of V2). In contrast, the mortality rate for those with a predicted risk of ≤0.0033 was 0.02% for the lowest quintile of risk in D1, 0.07% for the 19.3% having the lowest risk in V1, and 0% for the 9.7% of patients with the lowest risk in V2. Figure 3 indicates that the risk for dying peaks within the first few days of the hospitalization. Moreover, those in the high-risk group remained at elevated risk relative to the lower risk strata for at least 100 days.

Figure 2.

Risk of outcomes within intervals of mortality risk (validation hospital V1). The curves for the other 2 populations (D1, V2) were similar (see the Supporting information, Appendix II, in the online version of this article). Examples of possible risk strata are indicated.

Figure 3.

Instantaneous risk of death (hazard function) following hospital admission—validation hospital V1. For sake of clarity, 5 ordinal categories of predicted risk are shown. The curves for the other 2 populations (D1, V2) were similar and are shown in the Appendix II (see the Supporting information, Appendix I, in the online version of this article).

Relationships With Other Outcomes of Interest

The graphical curves of Figure 2 represent the occurrence of adverse events. The rising slopes indicate the risk for other events increases with the risk of dying within 30 days (for details and data for D1 and V2, see the Supporting Information, Appendix II, in the online version of this article). The strength of these relationships is quantified by the areas under the ROC curve (Table 2). The probability of 30-day mortality strongly predicted the occurrence of in-hospital death, palliative care status, and death within 180 days; modestly predicted having an unplanned transfer to an ICU within the first 24 hours of the hospitalization and undergoing resuscitative efforts for cardiopulmonary arrest; and weakly predicted intensive care unit use at some point in the hospitalization, occurrence of a condition not present on admission (complication), and being rehospitalized within 30 days

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....