Reducing Central Nervous System-Active Medications to Prevent Falls and Injuries Among Older Adults: A Cluster Randomized Clinical Trial.

Importance: High-risk medications that contribute to adverse health outcomes are frequently prescribed to older adults. Deprescribing interventions reduce their use, but studies are often not designed to examine effects on patient-relevant health outcomes. Objective: To test the effect of a health system-embedded deprescribing intervention targeting older adults and their primary care clinicians for reducing the use of central nervous system-active drugs and preventing medically treated falls. Design, Setting, and Participants: In this cluster randomized, parallel-group, clinical trial, 18 primary care practices from an integrated health care delivery system in Washington state were recruited from April 1, 2021, to June 16, 2022, to participate, along with their eligible patients. Randomization occurred at the clinic level. Patients were community-dwelling adults aged 60 years or older, prescribed at least 1 medication from any of 5 targeted medication classes (opioids, sedative-hypnotics, skeletal muscle relaxants, tricyclic antidepressants, and first-generation antihistamines) for at least 3 consecutive months. Intervention: Patient education and clinician decision support. Control arm participants received usual care. Main Outcomes and Measures: The primary outcome was medically treated falls. Secondary outcomes included medication discontinuation, sustained medication discontinuation, and dose reduction of any and each target medication. Serious adverse drug withdrawal events involving opioids or sedative-hypnotics were the main safety outcome. Analyses were conducted using intent-to-treat analysis. Results: Among 2367 patient participants (mean [SD] age, 70.6 [7.6] years; 1488 women [63%]), the adjusted cumulative incidence rate of a first medically treated fall at 18 months was 0.33 (95% CI, 0.29-0.37) in the intervention group and 0.30 (95% CI, 0.27-0.34) in the usual care group (estimated adjusted hazard ratio, 1.11 (95% CI, 0.94-1.31) (P = .11). There were significant differences favoring the intervention group in discontinuation, sustained discontinuation, and dose reduction of tricyclic antidepressants at 6 months (discontinuation adjusted rate: intervention group, 0.23 [95% CI, 0.18-0.28] vs usual care group, 0.13 [95% CI, 0.09-0.17]; adjusted relative risk, 1.79 [95% CI, 1.29-2.50]; P = .001) and secondary time points (9, 12, and 15 months). Conclusions and Relevance: In this randomized clinical trial of a health system-embedded deprescribing intervention targeting community-dwelling older adults prescribed central nervous system-active medications and their primary care clinicians, the intervention was no more effective than usual care in reducing medically treated falls. For health systems that attend to deprescribing as part of routine clinical practice, additional interventions may confer modest benefits on prescribing without a measurable effect on clinical outcomes. Trial Registration: ClinicalTrials.gov Identifier: NCT05689554.

Improving Methods of Identifying Anaphylaxis for Medical Product Safety Surveillance Using Natural Language Processing and Machine Learning.

We sought to determine whether machine learning and natural language processing (NLP) applied to electronic medical records could improve performance of automated health-care claims-based algorithms to identify anaphylaxis events using data on 516 patients with outpatient, emergency department, or inpatient anaphylaxis diagnosis codes during 2015-2019 in 2 integrated health-care institutions in the Northwest United States. We used one site's manually reviewed gold-standard outcomes data for model development and the other's for external validation based on cross-validated area under the receiver operating characteristic curve (AUC), positive predictive value (PPV), and sensitivity. In the development site 154 (64%) of 239 potential events met adjudication criteria for anaphylaxis compared with 180 (65%) of 277 in the validation site. Logistic regression models using only structured claims data achieved a cross-validated AUC of 0.58 (95% CI: 0.54, 0.63). Machine learning improved cross-validated AUC to 0.62 (0.58, 0.66); incorporating NLP-derived covariates further increased cross-validated AUCs to 0.70 (0.66, 0.75) in development and 0.67 (0.63, 0.71) in external validation data. A classification threshold with cross-validated PPV of 79% and cross-validated sensitivity of 66% in development data had cross-validated PPV of 78% and cross-validated sensitivity of 56% in external data. Machine learning and NLP-derived data improved identification of validated anaphylaxis events.