Using Highly Detailed Administrative Data to Predict Pneumonia Mortality

Abstract

BackgroundMortality prediction models generally require clinical data or are derived from information coded at discharge, limiting adjustment for presenting severity of illness in observational studies using administrative data.ObjectivesTo develop and validate a mortality prediction model using administrative data available in the first 2 hospital days.Research DesignAfter dividing the dataset into derivation and validation sets, we created a hierarchical generalized linear mortality model that included patient demographics, comorbidities, medications, therapies, and diagnostic tests administered in the first 2 hospital days. We then applied the model to the validation set.SubjectsPatients aged ≥18 years admitted with pneumonia between July 2007 and June 2010 to 347 hospitals in Premier, Inc.’s Perspective database.MeasuresIn hospital mortality.ResultsThe derivation cohort included 200,870 patients and the validation cohort had 50,037. Mortality was 7.2%. In the multivariable model, 3 demographic factors, 25 comorbidities, 41 medications, 7 diagnostic tests, and 9 treatments were associated with mortality. Factors that were most strongly associated with mortality included receipt of vasopressors, non-invasive ventilation, and bicarbonate. The model had a c-statistic of 0.85 in both cohorts. In the validation cohort, deciles of predicted risk ranged from 0.3% to 34.3% with observed risk over the same deciles from 0.1% to 33.7%.ConclusionsA mortality model based on detailed administrative data available in the first 2 hospital days had good discrimination and calibration. The model compares favorably to clinically based prediction models and may be useful in observational studies when clinical data are not available.

Outcomes of Patients with Healthcare-Associated Pneumonia: Worse Disease or Sicker Patients?

“…10 This model had good discrimination (c statistic, 0.86) and calibration. We then adjusted for the patient's length of stay and added an indicator of HCAP status to this model to assess its additional value in predicting death.…”

Association between dementia and discharge status in patients hospitalized with pneumonia

“…Pneumonia is also reportedly associated with poor long-term prognosis in patients with dementia [8,9]. Several studies have assessed the association between independent comorbidities with in-hospital mortality [10][11][12] in patients with pneumonia; however, the influence of dementia is not completely understood.…”

Outcomes of Patients with Healthcare-Associated Pneumonia: Worse Disease or Sicker Patients?

“…To predict the case-fatality rate for each patient, we constructed a hierarchical generalized linear model (SAS GLIMMIX) with a random hospital effect, using a logit link that included demographics, co-morbidities, medications (excluding antimicrobials) and treatments administered in the first 2 days of hospitalization. 10 This model had good discrimination (c-statistic 0.86) and calibration. We then adjusted for the patient’s length of stay and added an indicator of HCAP status to this model to assess its additional value in predicting death.…”

“…HCAP patients were also sicker on admission, but even after adjustment for initial treatments—a measure of severity of illness that accurately predicts inpatient outcomes 10 —they still had a higher case-fatality rate. This difference in case-fatality rate among the two groups may be due to more virulent or resistant organisms, or additional unmeasured cofounders.…”