Current and recommended practices for evaluating adverse drug events using electronic health records: A systematic review
Bibliographic record
Abstract
Abstract Electronic health records (EHR) are widely used sources of real‐world data in pharmacoepidemiologic research. As there is no end‐to‐end guidance for generating medication safety evidence with EHR, this study conducted a systematic review to determine the current and recommended practices in the literature. PubMed, Scopus, and CINAHL were searched for English articles published between 1 January 2010 and 11 June 2020. Selected articles were published in peer‐reviewed journals, conducted in the United States, analyzed structured EHR data, and defined drug exposure and adverse drug events (ADEs). The study evaluated methodological quality with a modified Newcastle‐Ottawa Scale (NOS) score ranging from 0 to 9 points. Data synthesis was performed with thematic analysis. Twenty‐six from 3885 articles were selected. The majority were cohort studies (85%). The studies were well designed, with a median NOS score of 9. Drug exposure was defined with dispensing (58%) and prescribing (31%) records. ADEs were defined across five categories: diagnosis codes (77%), validated outcome algorithms (35%), objective measures (35%), treatment procedures (19%), and antidotes (2%). Common covariates were age (89%), gender (85%), comorbidities (81%), and medication‐co‐medication use (73%). Four studies (15%) empirically defined covariates in a data‐driven manner. Twenty‐two (85%) analyzed covariates as confounders or effect modifiers in their analyses. Results were analyzed with either intention‐to‐treat (73%) or as‐treated (39%) approaches. Key recommendations include selecting dispensing rather than prescribing records, considering a proxy date of dispensation where applicable, selecting new instead of prevalent drug users, improving adoption of validated outcome algorithms, and not utilizing objective measures as the primary indicator of ADEs.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.132 | 0.353 |
| Meta-epidemiology (narrow) | 0.003 | 0.003 |
| Meta-epidemiology (broad) | 0.011 | 0.013 |
| Bibliometrics | 0.043 | 0.031 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.007 | 0.009 |
| Open science | 0.005 | 0.004 |
| Research integrity | 0.003 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".