Study Finds Electronic Clinical Decision Support Tool Improves Mortality in Patients With Pneumonia

An electronic pneumonia clinical decision support tool (ePNa) can help improve patient care and lower the risk of mortality for patients with pneumonia in the emergency department (ED), according to a study published in American Journal of Respiratory and Critical Care Medicine. The authors of this study aimed to report patient outcomes and processes of care after deployment of ePNa.

The ePNa tool has a knowledge base formed by the scientific guidelines used to help clinicians diagnose and treat pneumonia. ePNa extracts real-time and historical data from the electronic health record (EHR) to guide diagnosis, risk stratification, ordering of microbiology studies, site of care decisions, and treatment.

For this study, the researchers deployed ePNa into 6 geographic clusters of 16 Intermountain Healthcare hospital EDs at 2-month intervals between December 2017 and November 2018. Data were collected until June 2019, 18 months after the first cluster deployment.

All patients 18 years and older with radiographic pneumonia on ED chest imaging and International Classification of Diseases, Tenth Revision (ICD-10) discharge codes for pneumonia were included in this study. Mortality data were collected from Intermountain medical records and death certificates from the state departments of health. Missing data were identified by manual chart review; only less than 1% of total data were missing.

Because the EHR did not reliably capture ePNA use, ePNa utilization was calculated by physician review of individual ED clinician notes identified by pneumonia ICD-10 codes and ED chest imaging.

The primary analysis used a mixed effects model to evaluate the relationship between ePNa deployment and severity-adjusted 30-day mortality. Patient disposition from the ED was compared with what ePNa would have recommended, regardless of whether ePNa was used or not. Analyses of hospital length of stay included only the patients who survived to hospital discharge.

The researchers also conducted several post hoc sensitivity analyses. They first truncated data to 12 months before and 12 months after the washout period of each cluster, then further truncated to 6 months before and 6 months after. They also performed a sensitivity analysis to compare predicted mortality with observed mortality.

Of the 7641 patients with pneumonia discharge codes confirmed by ED chest imaging, 342 were excluded from washout periods, 37 were excluded due to in-hospital death or transfer to hospice, 1 patient was excluded for missing mortality information, and 413 were excluded due to transfer to a non-Intermountain hospital from which subsequent data were not available. This left 6848 patients for this study: 4536 from before ePNa deployment and 2312 from after deployment.

Median (interquartile range) age of the patients was 67 (50-79) years; 48% were women; 94% were White, including 7% of Hispanic, Latino, or Spanish origin; and 64.8% were initially admitted into the hospital.

Overall, 30-day all-cause mortality, including both outpatients and inpatients, was 8.6% before deployment and 4.8% after deployment of ePNa. A mixed effects regression model showed an odds ratio (OR) of 0.62 (95% CI, 0.49-0.79; P < .001) for mortality after deployment.

Sensitivity analyses that truncated data to 12 months before and after the washout period produced an OR of 0.66 (95% CI, 0.51-0.87; P = .003). When truncated to 6 months, the OR was 0.64 (95% CI, 0.44-0.93, P = .02).

In the mixed effects sensitivity model, observed mortality decreased by 3.8% after ePNa deployment (8.6% predeployment vs 4.8% post). This was greater than the 1.4% difference in predicted mortality (7.8% predeployment vs 6.4% post). Mortality reduction was greatest in patients directly admitted to the intensive care unit (ICU) from the ED (OR, 0.32; 95% CI, 0.14-0.77; P = .01) compared with those admitted to the medical floor (OR, 0.53; 95% CI, 0.25-1.1; P = .09) and with outpatient disposition.

Among patients admitted to the hospital, medication prescribing concordant with ePNa recommendations increased from 79.5% to 87.9% in the 8 hours following ED arrival. Use of broad-spectrum antibiotics was not significantly different before and after ePNa deployment (27% vs 25%). However, the use of antibiotics active against methicillin resistant Staphylococcus aureus decreased from 13% before deployment to 10% after deployment: from 15% to 8% between 8 and 72 hours and from 6% to 3% after 72 hours. Mean times from ED admission to first antibiotic were 159.4 (95% CI, 156.9-161.9) minutes before deployment and 150.9 (95% CI, 144.1-157.8) minutes after ePNa deployment.

Outpatient disposition for treatment of pneumonia in the ED increased from 29.2% before ePNa deployment to 46.9%. A similar increase was noted for patients who were recommended outpatient care by the ePNa (49.2% predeployment vs 66.6% post). Hospital ward disposition (57.3% vs 47%) and ICU disposition (13.5% to 6.1%) both decreased after ePNa deployment. Despite increased outpatient disposition, neither 7-day secondary hospital admission or adjusted 30-day mortality were significantly different before vs after ePNa deployment.

Overall, ePNa was utilized by the ED clinician for 67% of eligible patients with pneumonia after the tool was deployed. Utilization was 69% in the 6 larger hospitals but only 36% in the 10 smaller, rural hospitals.

There were some limitations to this study. The study was confined to a single health care system in a region of the United States with a patient population that may differ from other areas. The decision not to randomize by cluster may have affected the results of the study. The researchers noted that their results may have differed if they had included patients with pneumonia who had other discharge codes or lacked radiographic imaging. The inclusion criteria also excluded patients who were admitted to the hospital with a pneumonia diagnosis but had a different diagnosis upon discharge. The researchers were also not able to perform a sensitivity analysis limited to patients treated with ePNa.

The researchers concluded that the benefits of using ePNa included lower mortality rates and improved processes of care in patients with pneumonia. They plan to develop an interoperable version of the tool for use across different EHRs at health systems other than Intermountain.

Reference

Dean NC, Vines CG, Carr JR, et al. A pragmatic stepped-wedge, cluster-controlled trial of real-time pneumonia clinical decision support. Am J Respir Crit Care Med. Published online March 8, 2022. doi:10.1164/rccm.202109-2092OC.