Natural language processing-driven state machines to extract social factors from unstructured clinical documentation.

Objective: This study sought to create natural language processing algorithms to extract the presence of social factors from clinical text in 3 areas: (1) housing, (2) financial, and (3) unemployment. For generalizability, finalized models were validated on data from a separate health system for generalizability. Materials and Methods: Notes from 2 healthcare systems, representing a variety of note types, were utilized. To train models, the study utilized n-grams to identify keywords and implemented natural language processing (NLP) state machines across all note types. Manual review was conducted to determine performance. Sampling was based on a set percentage of notes, based on the prevalence of social need. Models were optimized over multiple training and evaluation cycles. Performance metrics were calculated using positive predictive value (PPV), negative predictive value, sensitivity, and specificity. Results: PPV for housing rose from 0.71 to 0.95 over 3 training runs. PPV for financial rose from 0.83 to 0.89 over 2 training iterations, while PPV for unemployment rose from 0.78 to 0.88 over 3 iterations. The test data resulted in PPVs of 0.94, 0.97, and 0.95 for housing, financial, and unemployment, respectively. Final specificity scores were 0.95, 0.97, and 0.95 for housing, financial, and unemployment, respectively. Discussion: We developed 3 rule-based NLP algorithms, trained across health systems. While this is a less sophisticated approach, the algorithms demonstrated a high degree of generalizability, maintaining >0.85 across all predictive performance metrics. Conclusion: The rule-based NLP algorithms demonstrated consistent performance in identifying 3 social factors within clinical text. These methods may be a part of a strategy to measure social factors within an institution.

Selecting EHR-driven recruitment strategies: An evidence-based decision guide.

Participant recruitment for research is a persistent bottleneck that can be improved by leveraging electronic health records (EHRs). Despite emerging evidence for various EHR-driven approaches, guidance for those attempting to select and use such approaches is limited. The national Recruitment Innovation Center established the EHR Recruitment Consult Resource (ERCR) service line to support multisite studies through implementation of EHR-driven recruitment strategies. As the ERCR, we evolved a guide through 17 consultations over 3 years with multisite studies recruiting in diverse biomedical research domains. We assessed literature and engaged domain experts to identify five key EHR-driven recruitment strategies: direct to patient messages, candidate lists for mailings/calls, direct to research alerts, point of care alerts, and participant registries. Differentiating factors were grouped into factors of study population, study protocol and recruitment workflows, and recruitment site capabilities. The decision matrix indicates acceptable or preferred strategies based on the differentiating factors. Across the ERCR consultations, candidate lists for mailing or calls were most common, participant registries were least frequently recommended, and for some studies no EHR-driven recruitment was recommended. Comparative effectiveness research is needed to refine further evidence for these and potentially new strategies to come.

Placental findings in low-risk, singleton, term pregnancies after uncomplicated deliveries.

Should all placentas be sent to pathology for examination after delivery room triage? A cohort of 88 placentas was prospectively obtained and examined from low-risk, singleton, term pregnancies after uneventful delivery. All patients had a normal prenatal testing and anatomy ultrasound. Fifty-one placentas (58%) were normal. Thirty-seven of the placental cohort (42%) had abnormal findings. Thirteen of the abnormal placentas (35.1%) showed pathology unassociated with fetal compromise. Twenty-four of the placentas (27.3% of the total cohort and 64.9% of the abnormal placentas) showed findings associated with fetal compromise. The most common pathologies were marginal cord insertion, chorioamnionitis, and abruption. Routine placental examination is not indicated, according to our data, in low-risk, singleton, and term pregnancy unless the placenta is determined to be abnormal at delivery examination.