Prognostication in acutely admitted older patients by nurses and physicians

Abstract

Background: The process of prognostication has not been described for acutely hospitalized older patients.

Objective: To investigate (1) which factors are associated with 90-day mortality risk in a group of acutely hospitalized older medical patients, and (2) whether adding a clinical impression score of nurses or physicians improves the discriminatory ability of mortality prediction.

Design: Prospective cohort study.

Participants: Four hundred and sixty-three medical patients 65 years or older acutely admitted from November 1, 2002, through July 1, 2005, to a 1024-bed tertiary university teaching hospital.

Measurements: At admission, the attending nurse and physician were asked to give a clinical impression score for the illness the patient was admitted for. This score ranged from 1 (high possibility of a good outcome) until 10 (high possibility of a bad outcome, including mortality). Of all patients baseline characteristics and clinical parameters were collected. Mortality was registered up to 90 days after admission.

Main results: In total, 23.8% (n = 110) of patients died within 90 days of admission. Four parameters were significantly associated with mortality risk: functional impairment, diagnosis malignancy, co-morbidities and high urea nitrogen serum levels. The AUC for the baseline model which included these risk factors (model 1) was 0.76 (95% CI 0.71 to 0.82). The AUC for the model using the risk factors and the clinical impression score of the physician (model 2) was 0.77 (0.71 to 0.82). The AUC for the model using the risk factors and the clinical impression score of the nurse (model 3) was 0.76 (0.71 to 0.82) and the AUC for the model, including the baseline covariates and the clinical impression score of both nurses and physicians was 0.77 (0.72 to 0.82). Adding clinical impression scores to model 1 did not significantly improve its accuracy.

Conclusion: A set of four clinical variables predicted mortality risk in acutely hospitalized older patients quite well. Adding clinical impression scores of nurses, physicians or both did not improve the discriminating ability of the model.

Figure 1

Clinical impression scores of nurses and physicians,  = 463 patients. Size of the bubble indicates the number of times a specific combination of nurse and physician score was given. The smallest bubble indicates a frequency of 1; the largest bubble indicates a frequency of 36.

Figure 2

ROC curves for four models predicting mortality. B is baseline set of clinical variables (model 1), B + P is model 1 with clinical impression score of physician (model 2), B + N is model 1 with clinical impression score of nurse (model 3) and B + P + N is model 1 with clinical impression score of both physicians and nurses (model 4).