Adverse drug event presentation and tracking (ADEPT): semiautomated, high throughput pharmacovigilance using real-world data

Abstract

Objective To advance use of real-world data (RWD) for pharmacovigilance, we sought to integrate a high-sensitivity natural language processing (NLP) pipeline for detecting potential adverse drug events (ADEs) with easily interpretable output for high-efficiency human review and adjudication of true ADEs. Materials and methods The adverse drug event presentation and tracking (ADEPT) system employs an open source NLP pipeline to identify in clinical notes mentions of medications and signs and symptoms potentially indicative of ADEs. ADEPT presents the output to human reviewers by highlighting these drug-event pairs within the context of the clinical note. To measure incidence of seizures associated with sildenafil, we applied ADEPT to 149 029 notes for 982 patients with pediatric pulmonary hypertension. Results Of 416 patients identified as taking sildenafil, NLP found 72 [17%, 95% confidence interval (CI) 14–21] with seizures as a potential ADE. Upon human review and adjudication, only 4 (0.96%, 95% CI 0.37–2.4) patients with seizures were determined to have true ADEs. Reviewers using ADEPT required a median of 89 s (interquartile range 57–142 s) per patient to review potential ADEs. Discussion ADEPT combines high throughput NLP to increase sensitivity of ADE detection and human review, to increase specificity by differentiating true ADEs from signs and symptoms related to comorbidities, effects of other medications, or other confounders. Conclusion ADEPT is a promising tool for creating gold standard, patient-level labels for advancing NLP-based pharmacovigilance. ADEPT is a potentially time savings platform for computer-assisted pharmacovigilance based on RWD.

Real‐world evidence to advance knowledge in pulmonary hypertension: Status, challenges, and opportunities. A consensus statement from the Pulmonary Vascular Research Institute's Innovative Drug Development Initiative's Real‐world Evidence Working Group

“…NLP is particularly relevant for RWD sources, as more than 80% of EHR data are unstructured 165 . NLP has been applied in a variety of clinical contexts, including detecting potential adverse drug events not identified through standard reporting systems, 166 screening candidates for clinical trial eligibility, 167 and improving the identification of diseases, 168 among others. While NLP has been widely employed in other cardiovascular and pulmonary diseases, 169 at present, its use in PH remains limited.…”

Key Contributions in Clinical Research Informatics

“…CRI researchers are progressively and intensively applying modern machine learning (ML) algorithms in the real-world and especially with EHR data. The fourth best paper from Geva et al describes the development of a natural language processing (NLP) pipeline for detecting potential adverse drug events (ADEs) in EHRs combined with a tool for human review and adjudication of true ADEs [9]. The ADE presentation and tracking (ADEPT) solution appears as an efficient advance on the use of real-world data for pharmacovigilance by computer-assisting expert reviewers with annotated candidate mentions in clinical documents.…”

Medication Errors: Detection Methodology Matters

“…The role of new technologies such as artificial intelligence or natural language processing to augment detection needs to be studied and may help expand our knowledge in this field. 4,5 For example, an artificial intelligence system that compares home medications with the inpatient medication record can both gather data on frequency of discrepancies as well as provide clinical decision support by prompting clinicians to correct such discrepancies. Alternatively using natural language processing, we could design an expert system to correlate clinical documentation of the medication plan (e.g., recommendations by consultants) with the electronic medication orders or even the medication administration record.…”

“…Future efforts to develop a more comprehensive detection of medication errors in healthcare are desirable. The role of new technologies such as artificial intelligence or natural language processing to augment detection needs to be studied and may help expand our knowledge in this field 4,5 . For example, an artificial intelligence system that compares home medications with the inpatient medication record can both gather data on frequency of discrepancies as well as provide clinical decision support by prompting clinicians to correct such discrepancies.…”