Using highly detailed administrative data to predict pneumonia mortality

• Published in: PloS one 2014-01, Vol.9 (1), p.e87382-e87382
• Main Authors: Rothberg, Michael B, Pekow, Penelope S, Priya, Aruna, Zilberberg, Marya D, Belforti, Raquel, Skiest, Daniel, Lagu, Tara, Higgins, Thomas L, Lindenauer, Peter K
• Format: Article
• Language: English
• Subjects: Adolescent; Adult; Aged; Aged, 80 and over; Analysis; Automation; Bicarbonate; Bicarbonates; Carbonates; Codes; Comorbidity; Computer Science; Demographics; Demography; Derivation; Diagnostic systems; Engineering; Female; Health aspects; Hospital Information Systems; Hospital Mortality; Hospitals; Humans; Hydrologic data; Illnesses; Male; Mathematical models; Medical diagnosis; Medical research; Medicine; Middle Aged; Models, Biological; Mortality; Observational studies; Patients; Pneumonia; Pneumonia - mortality; Pneumonia - therapy; Prediction models; Prognosis; Quality; Research design; Retrospective Studies; Risk Factors; Sepsis; Studies; Ventilation
• ISSN: 1932-6203; 1932-6203
• DOI: 10.1371/journal.pone.0087382

Mortality 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. To develop and validate a mortality prediction model using administrative data available in the first 2 hospital days. After 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. Patients aged ≥ 18 years admitted with pneumonia between July 2007 and June 2010 to 347 hospitals in Premier, Inc.'s Perspective database. In hospital mortality. The 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%. A 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.