Using the Electronic Medical Record to Identify Community-Acquired Pneumonia: Toward a Replicable Automated Strategy

Abstract

BackgroundTimely information about disease severity can be central to the detection and management of outbreaks of acute respiratory infections (ARI), including influenza. We asked if two resources: 1) free text, and 2) structured data from an electronic medical record (EMR) could complement each other to identify patients with pneumonia, an ARI severity landmark.MethodsA manual EMR review of 2747 outpatient ARI visits with associated chest imaging identified x-ray reports that could support the diagnosis of pneumonia (kappa score = 0.88 (95% CI 0.82∶0.93)), along with attendant cases with Possible Pneumonia (adds either cough, sputum, fever/chills/night sweats, dyspnea or pleuritic chest pain) or with Pneumonia-in-Plan (adds pneumonia stated as a likely diagnosis by the provider). The x-ray reports served as a reference to develop a text classifier using machine-learning software that did not require custom coding. To identify pneumonia cases, the classifier was combined with EMR-based structured data and with text analyses aimed at ARI symptoms in clinical notes.Results370 reference cases with Possible Pneumonia and 250 with Pneumonia-in-Plan were identified. The x-ray report text classifier increased the positive predictive value of otherwise identical EMR-based case-detection algorithms by 20–70%, while retaining sensitivities of 58–75%. These performance gains were independent of the case definitions and of whether patients were admitted to the hospital or sent home. Text analyses seeking ARI symptoms in clinical notes did not add further value.ConclusionSpecialized software development is not required for automated text analyses to help identify pneumonia patients. These results begin to map an efficient, replicable strategy through which EMR data can be used to stratify ARI severity.

Total duration of antimicrobial therapy in veterans hospitalized with uncomplicated pneumonia: Results of a national medication utilization evaluation

“…Finally, manual chart review to classify uncomplicated cases and related outcomes was laborious, and automated case identification is technologically plausible and should be explored. 39 In conclusion, this national VHA MUE found that patients with uncomplicated pneumonia were commonly prescribed antimicrobials for the duration of therapy in excess of guideline recommendations. Patients with uncomplicated pneumonia who received therapy duration consistent with guideline recommendations did not have significantly different all-cause readmission and death rates compared to patients receiving prolonged treatment.…”

“…Other work should include an in‐depth assessment of clinical outcomes related to treatment duration, investigation of provider rationale for prolonged treatment, and duration of antimicrobial therapy prescribed upon discharge for other common disease states. Finally, manual chart review to classify uncomplicated cases and related outcomes was laborious, and automated case identification is technologically plausible and should be explored …”

Epidemic Surveillance Using an Electronic Medical Record: An Empiric Approach to Performance Improvement

“…Information potentially useful to public health surveillance has long been extracted from the free text narrative of chief complaints [37][38][39][40][41], laboratory or imaging reports [37,38,[41][42][43][44], hospital discharge summaries [45,46] or outpatient clinical notes [19,47,48]. To date however, little is known about how information extracted from free text EMR fields should be combined with structured data to accelerate outbreak detection.…”

Incidence rates of hospital-acquired urinary tract and bloodstream infections generated by automated compilation of electronically available healthcare data

“…24 The authors are currently investigating the use of text-mining algorithms in order to extract unstructured data from the free text information documented in the EPR and convert this information to structured variables, such as previously described for community-acquired pneumonia by DeLisle et al Adding other types of HAI as well as employing a more complex set of definitions, such as those used by Shepard et al for catheter-associated UTI, may refine the HAIR algorithms and provide more informative data. 15,25 Although there are few commercial systems for electronic surveillance of HAI in general, none of these is available in Denmark. Commercial systems are expensive to procure and maintain; these costs may be avoided if existing databases within hospitals can be used as demonstrated in this study.…”