Male gender, elevated creatinine, decreased FEV(1.0)% predicted, mechanical ventilation, history of smoking, and acute use of systemic steroids during the hospitalization were associated with an increased risk of mortality.

CM&R 2011 : 3/4 (November) 170 HMORN -Selected Abstracts and those from people with comorbidities (cancer, chronic lung disease, or congestive heart failure). We estimated the sensitivity, specificity, and positive predictive value (PPV) of ONYX compared to manual review using multivariable logistic regression, accounting for clustering of reports within people. We examined how ONYX's performance varied by age and comorbidity. Results: ONYX classified 26% of reports (1,276/5,000; 38% [841/2,200] of true pneumonias and 16% [435/2,800] of non-pneumonias) as "requiring manual review" based on pre-defined criteria. Reports from older people were more likely to require manual review than those from younger people. Among reports that could be classified, ONYX had a sensitivity of 91% (1,242/1,359), specificity of 92% (2,170/2,365), and PPV of 81% (1,242/1,437, modeled based on pneumonia prevalence in the source database). Sensitivity and specificity were similar regardless of comorbidity. Conclusions: NLP offers potential for identifying pneumonia outcomes from EMR data. Next steps include 1) further training to decrease the proportion of reports requiring manual review and 2) evaluating the accuracy of ONYX in other health systems. Keywords: Natural language processing, Pneumonia, Objectives: Readmission and mortality among elderly patients admitted to inpatient care are critical outcome measures for healthcare systems. While the Veterans Health Administration (VA) has long made national administrative databases available for health services research, the HMORN Virtual Data Warehouse (VDW) at Scott & White (SWHP) has only been developed in the past year. The current study sought to utilize both data sources while examining outcomes of hospitalization for community-acquired pneumonia (CAP) in two systems of care in the same geographic region. Methods: Patients >65 years of age hospitalized with CAP (ICD9 codes [480][481][482][483][485][486][487] Dates of care and inpatient death were sought to determine outcomes subsequent to CAP admission. Covariates included sociodemographic measures (such as age, gender) and comorbidity. Results: There were 469 SWHP CAP patients and 152 VA CAP patients admitted in the year studied. The SWHP fiscal year differed from the VA fiscal year, potentially complicating the work of the SWHP data team. VA CAP patients averaged 77.2 years (SD 7.6), were 96% male, and experienced an inpatient death rate of 18%. SWHP CAP patients averaged 81.6 years (SD 8.2), and 56% were women. Date of death as an inpatient was not explicitly available in VDW data. Patients in the two systems varied significantly as VA patients were more likely to be male and younger. Conclusions: Patients in the federal VA system can be studied in conjunction with private-care patients in the same catchment area using newly available VDW data. The data presented required two teams of data analyst/programmers, one working within the VA system and the other working within the SWHP VDW system. Coordination was facilitated by t...

No abstract