Effect of integration of prescription drug monitoring program data in the electronic health record on queries by primary care providers.

Objective: To evaluate the impact of PDMP integration in the EHR on provider query rates within twelve primary care clinics in one academic medical center. Methods: Using linked data from the EHR and state PDMP program, we evaluated changes in PDMP query rates using a stepped-wedge observational design where integration was implemented in three waves (four clinics per wave) over a five-month period (May, July, September 2019). Multivariable negative binomial general estimating equations (GEE) models assessed changes in PDMP query rates, overall and across several provider and clinic-level subgroups. Results: Among 206 providers in PDMP integrated clinics, the average number of queries per provider per month increased significantly from 1.43 (95% CI 1.07 - 1.91) pre-integration to 3.94 (95% CI 2.96 - 5.24) post-integration, a 2.74-fold increase (95% CI 2.11 to 3.59; p < .0001). Those in the lowest quartile of PDMP use pre-integration increased 36.8-fold (95% CI 16.91 - 79.95) after integration, significantly more than other pre-integration PDMP use quartiles. Conclusions: Integration of the PDMP in the EHR significantly increased the use of the PDMP overall and across all studied subgroups. PDMP use increased to a greater degree among providers with lower PDMP use pre-integration.

Identifying subtypes of type 2 diabetes mellitus with machine learning: development, internal validation, prognostic validation and medication burden in linked electronic health records in 420 448 individuals.

INTRODUCTION: None of the studies of type 2 diabetes (T2D) subtyping to date have used linked population-level data for incident and prevalent T2D, incorporating a diverse set of variables, explainable methods for cluster characterization, or adhered to an established framework. We aimed to develop and validate machine learning (ML)-informed subtypes for type 2 diabetes mellitus (T2D) using nationally representative data. RESEARCH DESIGN AND METHODS: In population-based electronic health records (2006-2020; Clinical Practice Research Datalink) in individuals ≥18 years with incident T2D (n=420 448), we included factors (n=3787), including demography, history, examination, biomarkers and medications. Using a published framework, we identified subtypes through nine unsupervised ML methods (K-means, K-means++, K-mode, K-prototype, mini-batch, agglomerative hierarchical clustering, Birch, Gaussian mixture models, and consensus clustering). We characterized clusters using intracluster distributions and explainable artificial intelligence (AI) techniques. We evaluated subtypes for (1) internal validity (within dataset; across methods); (2) prognostic validity (prediction for 5-year all-cause mortality, hospitalization and new chronic diseases); and (3) medication burden. RESULTS: Development: We identified four T2D subtypes: metabolic, early onset, late onset and cardiometabolic. Internal validity: Subtypes were predicted with high accuracy (F1 score >0.98). Prognostic validity: 5-year all-cause mortality, hospitalization, new chronic disease incidence and medication burden differed across T2D subtypes. Compared with the metabolic subtype, 5-year risks of mortality and hospitalization in incident T2D were highest in late-onset subtype (HR 1.95, 1.85-2.05 and 1.66, 1.58-1.75) and lowest in early-onset subtype (1.18, 1.11-1.27 and 0.85, 0.80-0.90). Incidence of chronic diseases was highest in late-onset subtype and lowest in early-onset subtype. Medications: Compared with the metabolic subtype, after adjusting for age, sex, and pre-T2D medications, late-onset subtype (1.31, 1.28-1.35) and early-onset subtype (0.83, 0.81-0.85) were most and least likely, respectively, to be prescribed medications within 5 years following T2D onset. CONCLUSIONS: In the largest study using ML to date in incident T2D, we identified four distinct subtypes, with potential future implications for etiology, therapeutics, and risk prediction.

Relationship between youth cardiometabolic health and physical activity in medical records.

BACKGROUND: Thers is limited research examining modifiable cardiometabolic risk factors with a single-item health behavior question obtained during a clinic visit. Such information could support clinicians in identifying patients at risk for adverse cardiometabolic health. We investigated if children meeting physical activity or screen time recommendations, collected during clinic visits, have better cardiometabolic health than children not meeting recommendations. We hypothesized that children meeting either recommendation would have fewer cardiometabolic risk factors. METHODS AND FINDINGS: This cross-sectional study used data from electronic medical records (EMRs) between January 1, 2013 through December 30, 2017 from children (2-18 years) with a well child visits and data for ≥1 cardiometabolic risk factor (i.e., systolic and diastolic blood pressure, glycated hemoglobin, alanine transaminase, high-density and low-density lipoprotein, total cholesterol, and/or triglycerides). Physical activity and screen time were patient/caregiver-reported. Analyses included EMRs from 63,676 well child visits by 30,698 unique patients (49.3% female; 41.7% Black, 31.5% Hispanic). Models that included data from all visits indicated children meeting physical activity recommendations had reduced risk for abnormal blood pressure (odds ratio [OR] = 0.91, 95%CI 0.86, 0.97; p = 0.002), glycated hemoglobin (OR = 0.83, 95%CI 0.75, 0.91; p = 0.00006), alanine transaminase (OR = 0.85, 95%CI 0.79, 0.92; p = 0.00001), high-density lipoprotein (OR = 0.88, 95%CI 0.82, 0.95; p = 0.0009), and triglyceride values (OR = 0.89, 95%CI 0.83, 0.96; p = 0.002). Meeting screen time recommendations was not associated with abnormal cardiometabolic risk factors. CONCLUSION: Collecting information on reported adherence to meeting physical activity recommendations can provide clinicians with additional information to identify patients with a higher risk of adverse cardiometabolic health.

Use of electronic health record data to examine administrations of pro re nata analgesics during hip fracture post-acute care.

PURPOSE: Medications prescribed to older adults in US skilled nursing facilities (SNF) and administrations of pro re nata (PRN) "as needed" medications are unobservable in Medicare insurance claims. There is an ongoing deficit in our understanding of medication use during post-acute care. Using SNF electronic health record (EHR) datasets, including medication orders and barcode medication administration records, we described patterns of PRN analgesic prescribing and administrations among SNF residents with hip fracture. METHODS: Eligible participants resided in SNFs owned by 11 chains, had a diagnosis of hip fracture between January 1, 2018 to August 2, 2021, and received at least one administration of an analgesic medication in the 100 days after the hip fracture. We described the scheduling of analgesics, the proportion of available PRN doses administered, and the proportion of days with at least one PRN analgesic administration. RESULTS: Among 24 038 residents, 57.3% had orders for PRN acetaminophen, 67.4% PRN opioids, 4.2% PRN non-steroidal anti-inflammatory drugs, and 18.6% PRN combination products. The median proportion of available PRN doses administered per drug was 3%-50% and the median proportion of days where one or more doses of an ordered PRN analgesic was administered was 25%-75%. Results differed by analgesic class and the number of administrations ordered per day. CONCLUSIONS: EHRs can be leveraged to ascertain precise analgesic exposures during SNF stays. Future pharmacoepidemiology studies should consider linking SNF EHRs to insurance claims to construct a longitudinal history of medication use and healthcare utilization prior to and during episodes of SNF care.

Implementation and use of electronic patient records in the Brazilian Air Force: a cross-sectional study1.

OBJECTIVE: The objective was to analyze the implementation and use of the electronic patient record in the health services of the Brazilian Air Force. METHODS: This is a cross-sectional study carried out with 234 physicians, between March and May 2021. The data collection instrument was sent by email. The electronic patient record was implemented in the Air Force approximately 3 years ago (64.5%), and about 81% of the physicians received training to operate it. RESULTS: The most common records involve data related to consultations (90.1%) and interviews with physical examination (67.1%). Physicians cited that information storage (75.6%), agility, and feasibility of recording (55.1%) were the main advantages of the electronic patient record. As disadvantages, problems in electronic equipment (69.7%) and system errors (65%) were reported. Most participants considered that the implementation had a positive impact on work dynamics (75.6%) and productivity (66.7%), mainly regarding the components "Work processes" (57.3%) and "Amount of carried out activities" (21.4%). Keeping records was significantly associated with the job position (p<0.001), type of unit (p=0.008), time of implementation (p<0.001), and participation in training (p=0.028). CONCLUSION: The implementation of the electronic patient record in the Air Force was recently done, and just over half of the physicians were trained prior to the implementation. The tool is considered compatible with work processes and has a positive effect on productivity.

Insomnia symptom prevalence in England: a comparison of cross-sectional self-reported data and primary care records in the UK Biobank.

OBJECTIVES: We aimed to use a large dataset to compare self-reported and primary care measures of insomnia symptom prevalence in England and establish whether they identify participants with similar characteristics. DESIGN: Cross-sectional study with linked electronic health records (EHRs). SETTING: Primary care in England. PARTICIPANTS: 163 748 UK Biobank participants in England (aged 38-71 at baseline) with linked primary care EHRs. OUTCOME MEASURES: We compared the percentage of those self-reporting 'usually' having insomnia symptoms at UK Biobank baseline assessment (2006-2010) to those with a Read code for insomnia symptoms in their primary care records prior to baseline. We stratified prevalence in both groups by sociodemographic, lifestyle, sleep and health characteristics. RESULTS: We found that 29% of the sample self-reported having insomnia symptoms, while only 6% had a Read code for insomnia symptoms in their primary care records. Only 10% of self-reported cases had an insomnia symptom Read code, while 49% of primary care cases self-reported having insomnia symptoms. In both primary care and self-reported data, prevalence of insomnia symptom cases was highest in females, older participants and those with the lowest household incomes. However, while snorers and risk takers were more likely to be a primary care case, they were less likely to self-report insomnia symptoms than non-snorers and non-risk takers. CONCLUSIONS: Only a small proportion of individuals experiencing insomnia symptoms have an insomnia symptom Read code in their primary care record. However, primary care data do provide a clinically meaningful measure of insomnia prevalence. In addition, the sociodemographic characteristics of people attending primary care with insomnia were consistent with those with self-reported insomnia, thus primary care records are a valuable data source for studying risk factors for insomnia. Further studies should replicate our findings in other populations and examine ways to increase discussions about sleep health in primary care.

Children and young people's body mass index measures derived from routine data sources: A national data linkage study in Wales.

BACKGROUND: Routine monitoring of Body Mass Index (BMI) in general practice, and via national surveillance programmes, is essential for the identification, prevention, and management of unhealthy childhood weight. We examined and compared the presence and representativeness of children and young people's (CYPs) BMI recorded in two routinely collected administrative datasets: general practice electronic health records (GP-BMI) and the Child Measurement Programme for Wales (CMP-BMI), which measures height and weight in 4-5-year-old school children. We also assessed the feasibility of combining GP-BMI and CMP-BMI data for longitudinal analyses. METHODS: We accessed de-identified population-level GP-BMI data for calendar years 2011 to 2019 for 246,817 CYP, and CMP-BMI measures for 222,772 CYP, held within the Secure Anonymised Information Linkage Databank. We examined the proportion of CYP in Wales with at least one GP-BMI record, its distribution by child socio-demographic characteristics, and trends over time. We compared GP-BMI and CMP-BMI distributions. We quantified the proportion of children with a CMP-BMI measure and a follow-up GP-BMI recorded at an older age and explored the representativeness of these measures. RESULTS: We identified a GP-BMI record in 246,817 (41%) CYP, present in a higher proportion of females (54.2%), infants (20.7%) and adolescents. There was no difference in the deprivation profile of those with a GP-BMI measurement. 31,521 CYP with a CMP-BMI had at least one follow-up GP-BMI; those with a CMP-BMI considered underweight or very overweight were 87% and 70% more likely to have at least one follow-up GP-BMI record respectively compared to those with a healthy weight, as were males and CYP living in the most deprived areas of Wales. CONCLUSIONS: Records of childhood weight status extracted from general practice are not representative of the population and are biased with respect to weight status. Linkage of information from the national programme to GP records has the potential to enhance discussions around healthy weight at the point of care but does not provide a representative estimate of population level weight trajectories, essential to provide insights into factors determining a healthy weight gain across the early life course. A second CMP measurement is required in Wales.

The impact of hospitalization on inappropriate prescribing and polypharmacy in older patients: A descriptive cross-sectional study.

BACKGROUND: Polypharmacy and the use of potentially inappropriate medications (PIMs) in older individuals are widespread phenomena that are associated with an increase in morbidity and mortality. The Beers Criteria is a tool that helps to identify patients that are prescribed with PIMs, thereby reducing the risk of associated harm. Amongst other populations, the criteria identify drugs that should not be used by the majority of older patients. AIM: Determining the proportion of older inpatients who were discharged from hospitalization with polypharmacy (a prescription for more than seven drugs), or with a PIM as defined by the Beers Criteria. METHODS: A descriptive cross-sectional study based on patients aged 65 and over who were hospitalized in the years 2019-2021 in the internal medicine, orthopedic and surgical wards at a medium-size hospital. Demographic information and details about drug treatment were collected from the electronic patient records system. Patients who died during hospitalization were excluded from the study group. MAIN OUTCOME MEASURES: The proportion of inpatients with polypharmacy or a PIM as part of their regular prescription, at the time of admission and at discharge. RESULTS: 49 564 patients were included in the study cohort. At discharge, 19% of the patients were given a prescription for a PIM, with a small but significant decrease compared with the rate admission (22.1%). At discharge, 42.8% of patients had polypharmacy, representing a small but significant increase compared with the rate on admission (40.6%). CONCLUSIONS: The study demonstrated high baseline rates of PIM prescription and polypharmacy. Hospitalization was associated with a decrease in PIM prescription and an increase in polypharmacy. This highlights the importance of medication review during admission to reduce the potential risk to older adults from polypharmacy and PIM prescription.

Electronic Immunization Registry in Rwanda: Qualitative Study of Health Worker Experiences.

BACKGROUND: Monitoring childhood immunization programs is essential for health systems. Despite the introduction of an electronic immunization registry called e-Tracker in Rwanda, challenges such as lacking population denominators persist, leading to implausible reports of coverage rates of more than 100%. OBJECTIVE: This study aimed to assess the extent to which the immunization e-Tracker responds to stakeholders' needs and identify key areas for improvement. METHODS: In-depth interviews were conducted with all levels of e-Tracker users including immunization nurses, data managers, and supervisors from health facilities in 5 districts of Rwanda. We used an interview guide based on the constructs of the Human, Organization, and Technology-Fit (HOT-Fit) framework, and we analyzed and summarized our findings using the framework. RESULTS: Immunization nurses reported using the e-Tracker as a secondary data entry tool in addition to paper-based forms, which resulted in considerable dissatisfaction among nurses. While users acknowledged the potential of a digital tool compared to paper-based systems, they also reported the need for improvement of functionalities to support their work, such as digital client appointment lists, lists of defaulters, search and register functions, automated monthly reports, and linkages to birth notifications and the national identity system. CONCLUSIONS: Reducing dual documentation for users can improve e-Tracker use and user satisfaction. Our findings can help identify additional digital health interventions to support and strengthen the health information system for the immunization program.

The diabetes mellitus multimorbidity network in hospitalized patients over 50 years of age in China: data mining of medical records.

OBJECTIVE: Many diabetes mellitus (DM) patients suffer from multimorbidity. Understanding the DM multimorbidity network should be given priority. The purpose of this study is characterize the DM multimorbidity network in people over 50 years. METHODS: Data on 75 non-communicable diseases (NCDs) were extracted from electronic medical records of 309,843 hospitalized patients older than 50 years who had at least one NCD. The association rules analysis was used as a novel classification method and combined with the Chi-square tests to identify associations between NCDs and DM. RESULT: A total of 12 NCDs were closely related to DM, {cholelithiasis, DM} was an unexpected combination. {dyslipidemia, DM} and {gout, DM} had the largest lift in the male and female groups, respectively. The negative related group included 7 NCDs. There were 9 NCDs included in the strong association rules. Most combinations were different by age and sex. In males, the strongest rule was {peripheral vascular disease (PVD), dyslipidemia, DM}, while {hypertension, dyslipidemia, chronic liver disease (CLD), DM} was the strongest in females. In patients younger than 70 years, hypertension, CLD, and dyslipidemia were the most dominant NCDs in the DM multimorbidity network. In patients 70 years or older, chronic kidney disease (CKD), CVD, CHD, and heart disease (HD) frequently co-occurred with DM. CONCLUSION: Future primary healthcare policies for DM should be formulated based on age and sex. In patients younger than 70 years, more attention to hypertension, CLD, and dyslipidemia is required, while attention to CKD, CVD, CHD and HD is needed in patients older than 70 years.

The Role of Telehealth in Improving Care Connections and Outcomes for Community Health Center Patients with Diabetes.

INTRODUCTION: Does telehealth decrease health disparities by improving connections to care or simply result in new barriers for vulnerable populations who often lack access to technology? This study aims to better understand the role of telehealth and social determinants of health in improving care connections and outcomes for Community Health Center patients with diabetes. METHODS: This retrospective analysis of Electronic Health Record (EHR) data examined the relationship between telehealth utilization and glycemic control and consistency of connection to the health care team ("connectivity"). EHR data were collected from 20 Community Health Centers from July 1, 2019 through December 31, 2021. Descriptive statistics were calculated, and multivariable linear regression was used to assess the associations between telehealth use and engagement in care and glycemic control. RESULTS: The adjusted analysis found positive, statistically significant associations between telehealth use and each of the 2 primary outcomes. Telehealth use was associated with 0.89 additional months of hemoglobin A1c (HbA1c) control (95% confidence interval [CI], 0.73 to 1.04) and 4.49 additional months of connection to care (95% CI, 4.27 to 4.70). DISCUSSION: The demonstrated increased engagement in primary care for telehealth users is significant and encouraging as Community Health Center populations are at greater risk of lapses in care and loss to follow up. CONCLUSIONS: Telehealth can be a highly effective, patient-centered form of care for people with diabetes. Telehealth can play a critical role in keeping vulnerable patients with diabetes connected to their care team and involved in care and may be an important tool for reducing health disparities.

Clinician Barriers to Ordering Pulmonary Function Tests for Adults with Suspected Asthma.

BACKGROUND: Primary care clinicians do not adhere to national and international guidelines recommending pulmonary function testing (PFTs) in patients with suspected asthma. Little is known about why that occurs. Our objective was to assess clinician focused barriers to ordering PFTs. METHODS: An internet-based 11-item survey of primary care clinicians at a large safety-net institution was conducted between August 2021 and November 2021. This survey assessed barriers and possible electronic health record (EHR) solutions to ordering PFTs. One of the survey questions contained an open-ended question about barriers which was analyzed qualitatively. RESULTS: The survey response rate was 59% (117/200). The top 3 reported barriers included beliefs that testing will not change management, distance to testing site, and the physical effort it takes to complete testing. Clinicians were in favor of an EHR intervention to prompt them to order PFTs. Responses to the open-ended question also conveyed that objective testing does not change management. DISCUSSION: PFTs improve diagnostic accuracy and reduce inappropriate therapies. Of the barriers we identified, the most modifiable is to educate clinicians about how PFTs can change management. That in conjunction with an EHR prompt, which clinicians approved of, may lead to guideline congruent and improved quality in asthma care.

Variable Impact of Medical Scribes on Physician Electronic Health Record Documentation Practices: A Quantitative Analysis Across a Large, Integrated Health-System.

BACKGROUND: Medical scribes have been utilized to reduce electronic health record (EHR) associated documentation burden. Although evidence suggests benefits to scribes, no large-scale studies have quantitatively evaluated scribe impact on physician documentation across clinical settings. This study aimed to evaluate the effect of scribes on physician EHR documentation behaviors and performance. METHODS: This retrospective cohort study used EHR audit log data from a large academic health system to evaluate clinical documentation for all ambulatory encounters between January 2014 and December 2019 to evaluate the effect of scribes on physician documentation behaviors. Scribe services were provided on a first-come, first-served basis on physician request. Based on a physician's scribe use, encounters were grouped into 3 categories: never using a scribe, prescribe (before scribe use), or using a scribe. Outcomes included chart closure time, the proportion of delinquent charts, and charts closed after-hours. RESULTS: Three hundred ninety-five physicians (23% scribe users) across 29 medical subspecialties, encompassing 1,132,487 encounters, were included in the analysis. At baseline, scribe users had higher chart closure time, delinquent charts, and after-hours documentation than physicians who never used scribes. Among scribe users, the difference in outcome measures postscribe compared with baseline varied, and using a scribe rarely resulted in outcome measures approaching a range similar to the performance levels of nonusing physicians. In addition, there was variability in outcome measures across medical specialties and within similar subspecialties. CONCLUSION: Although scribes may improve documentation efficiency among some physicians, not all will improve EHR-related documentation practices. Different strategies may help to optimize documentation behaviors of physician-scribe dyads and maximize outcomes of scribe implementation.

The Ability of Primary Care Practices to Measure and Report on Care Quality.

BACKGROUND: Creating useful clinical quality measure (CQM) reports in a busy primary care practice is known to depend on the capability of the electronic health record (EHR). Two other domains may also contribute: supportive leadership to prioritize the work and commit the necessary resources, and individuals with the necessary health information technology (IT) skills to do so. Here we describe the results of an assessment of the above 3 domains and their associations with successful CQM reporting during an initiative to improve smaller primary care practices' cardiovascular disease CQMs. METHODS: The study took place within an AHRQ EvidenceNOW initiative of external support for smaller practices across Washington, Oregon and Idaho. Practice facilitators who provided this support completed an assessment of the 3 domains previously described for each of their assigned practices. Practices submitted 3 CQMs to the study team: appropriate aspirin prescribing, use of statins when indicated, blood pressure control, and tobacco screening/cessation. RESULTS: Practices with advanced EHR reporting capability were more likely to report 2 or more CQMs. Only one-third of practices were "advanced" in this domain, and this domain had the highest proportion of practices (39.1%) assessed as "basic." The presence of advanced leadership or advanced skills did not appreciably increase the proportion of practices that reported 2 or more CQMs. CONCLUSIONS: Our findings support previous reports of limited EHR reporting capabilities within smaller practices but extend these findings by demonstrating that practices with advanced capabilities in this domain are more likely to produce CQM reports.

Life, death, and statins: association of statin prescriptions and survival in older general practice patients.

AIMS: This study serves as an exemplar to demonstrate the scalability of a research approach using survival analysis applied to general practice electronic health record data from multiple sites. Collection of these data, the subsequent analysis, and the preparation of practice-specific reports were performed using a bespoke distributed data collection and analysis software tool. BACKGROUND: Statins are a very commonly prescribed medication, yet there is a paucity of evidence for their benefits in older patients. We examine the relationship between statin prescriptions for general practice patients over 75 and all-cause mortality. METHODS: We carried out a retrospective cohort study using survival analysis applied to data extracted from the electronic health records of five Australian general practices. FINDINGS: The data from 8025 patients were analysed. The median duration of follow-up was 6.48 years. Overall, 52 015 patient-years of data were examined, and the outcome of death from any cause was measured in 1657 patients (21%), with the remainder being censored. Adjusted all-cause mortality was similar for participants not prescribed statins versus those who were (HR 1.05, 95% CI 0.92-1.20, P = 0.46), except for patients with diabetes for whom all-cause mortality was increased (HR = 1.29, 95% CI: 1.00-1.68, P = 0.05). In contrast, adjusted all-cause mortality was significantly lower for patients deprescribed statins compared to those who were prescribed statins (HR 0.81, 95% CI 0.70-0.93, P < 0.001), including among females (HR = 0.75, 95% CI: 0.61-0.91, P < 0.001) and participants treated for secondary prevention (HR = 0.72, 95% CI: 0.60-0.86, P < 0.001). This study demonstrated the scalability of a research approach using survival analysis applied to general practice electronic health record data from multiple sites. We found no evidence of increased mortality due to statin-deprescribing decisions in primary care.

Long-Term Outcomes Associated With ß-Lactam Allergies.

Importance: ß-lactam (BL) allergies are the most common drug allergy worldwide, but most are reported in error. BL allergies are also well-established risk factors for adverse drug events and antibiotic-resistant infections during inpatient health care encounters, but the understanding of the long-term outcomes of patients with BL allergies remains limited. Objective: To evaluate the long-term clinical outcomes of patients with BL allergies. Design, Setting, and Participants: This longitudinal retrospective cohort study was conducted at a single regional health care system in western Pennsylvania. Electronic health records were analyzed for patients who had an index encounter with a diagnosis of sepsis, pneumonia, or urinary tract infection between 2007 and 2008. Patients were followed-up until death or the end of 2018. Data analysis was performed from January 2022 to January 2024. Exposure: The presence of any BL class antibiotic in the allergy section of a patient's electronic health record, evaluated at the earliest occurring observed health care encounter. Main Outcomes and Measures: The primary outcome was all-cause mortality, derived from the Social Security Death Index. Secondary outcomes were defined using laboratory and microbiology results and included infection with methicillin-resistant Staphylococcus aureus (MRSA), Clostridium difficile, or vancomycin-resistant Enterococcus (VRE) and severity and occurrence of acute kidney injury (AKI). Generalized estimating equations with a patient-level panel variable and time exposure offset were used to evaluate the odds of occurrence of each outcome between allergy groups. Results: A total of 20 092 patients (mean [SD] age, 62.9 [19.7] years; 12 231 female [60.9%]), of whom 4211 (21.0%) had BL documented allergy and 15 881 (79.0%) did not, met the inclusion criteria. A total of 3513 patients (17.5%) were Black, 15 358 (76.4%) were White, and 1221 (6.0%) were another race. Using generalized estimating equations, documented BL allergies were not significantly associated with the odds of mortality (odds ratio [OR], 1.02; 95% CI, 0.96-1.09). BL allergies were associated with increased odds of MRSA infection (OR, 1.44; 95% CI, 1.36-1.53), VRE infection (OR, 1.18; 95% CI, 1.05-1.32), and the pooled rate of the 3 evaluated antibiotic-resistant infections (OR, 1.33; 95% CI, 1.30-1.36) but were not associated with C difficile infection (OR, 1.04; 95% CI, 0.94-1.16), stage 2 and 3 AKI (OR, 1.02; 95% CI, 0.96-1.10), or stage 3 AKI (OR, 1.06; 95% CI, 0.98-1.14). Conclusions and Relevance: Documented BL allergies were not associated with the long-term odds of mortality but were associated with antibiotic-resistant infections. Health systems should emphasize accurate allergy documentation and reduce unnecessary BL avoidance.

Identification of patients' smoking status using an explainable AI approach: a Danish electronic health records case study.

BACKGROUND: Smoking is a critical risk factor responsible for over eight million annual deaths worldwide. It is essential to obtain information on smoking habits to advance research and implement preventive measures such as screening of high-risk individuals. In most countries, including Denmark, smoking habits are not systematically recorded and at best documented within unstructured free-text segments of electronic health records (EHRs). This would require researchers and clinicians to manually navigate through extensive amounts of unstructured data, which is one of the main reasons that smoking habits are rarely integrated into larger studies. Our aim is to develop machine learning models to classify patients' smoking status from their EHRs. METHODS: This study proposes an efficient natural language processing (NLP) pipeline capable of classifying patients' smoking status and providing explanations for the decisions. The proposed NLP pipeline comprises four distinct components, which are; (1) considering preprocessing techniques to address abbreviations, punctuation, and other textual irregularities, (2) four cutting-edge feature extraction techniques, i.e. Embedding, BERT, Word2Vec, and Count Vectorizer, employed to extract the optimal features, (3) utilization of a Stacking-based Ensemble (SE) model and a Convolutional Long Short-Term Memory Neural Network (CNN-LSTM) for the identification of smoking status, and (4) application of a local interpretable model-agnostic explanation to explain the decisions rendered by the detection models. The EHRs of 23,132 patients with suspected lung cancer were collected from the Region of Southern Denmark during the period 1/1/2009-31/12/2018. A medical professional annotated the data into 'Smoker' and 'Non-Smoker' with further classifications as 'Active-Smoker', 'Former-Smoker', and 'Never-Smoker'. Subsequently, the annotated dataset was used for the development of binary and multiclass classification models. An extensive comparison was conducted of the detection performance across various model architectures. RESULTS: The results of experimental validation confirm the consistency among the models. However, for binary classification, BERT method with CNN-LSTM architecture outperformed other models by achieving precision, recall, and F1-scores between 97% and 99% for both Never-Smokers and Active-Smokers. In multiclass classification, the Embedding technique with CNN-LSTM architecture yielded the most favorable results in class-specific evaluations, with equal performance measures of 97% for Never-Smoker and measures in the range of 86 to 89% for Active-Smoker and 91-92% for Never-Smoker. CONCLUSION: Our proposed NLP pipeline achieved a high level of classification performance. In addition, we presented the explanation of the decision made by the best performing detection model. Future work will expand the model's capabilities to analyze longer notes and a broader range of categories to maximize its utility in further research and screening applications.

Disparities in Patient Portal Engagement Among Patients With Hypertension Treated in Primary Care.

Importance: Patient portals are increasingly used for patient-clinician communication and to introduce interventions aimed at improving blood pressure control. Objective: To characterize patient portal use among patients with hypertension managed in primary care. Design, Settings, and Participants: This retrospective cohort study used electronic health records linked with patient portal log file data from a large, diverse Midwestern health care system. Patients with hypertension who had a primary care visit from January 1, 2021, to December 31, 2021, were included. The first visit in 2021 was considered the baseline visit; patient portal engagement was evaluated during the following year. Multivariate logistic regressions, presented as odds ratios (ORs) and 95% CIs, were used to evaluate associations between patient characteristics and patient portal engagement, adjusting for potential confounders. Exposures: Primary exposures included 4 sociodemographic factors routinely collected in the electronic health record: race and ethnicity, insurance, preferred language, and smoking status. Main Outcomes and Measures: Indicators of patient engagement with the patient portal included accessing the patient portal at least once, accessing the portal within 7 days of at least 50.0% of primary care physician (PCP) visits, frequent logins (<28 vs ≥28), messaging (<2 vs ≥2), and sharing home blood pressure readings. Results: Among 366 871 patients (mean [SD], 63.5 [12.6] years), 52.8% were female, 3.4% were Asian, 7.8% were Hispanic, 19.7% were non-Hispanic Black, 66.9% were non-Hispanic White, and 2.3% were of other race or ethnicity. During the 1-year study period starting in 2021, 70.5% accessed the patient portal at least once, 60.2% accessed around the time of their PCP visits, 35.7% accessed the portal frequently, 28.9% engaged in messaging, and 8.7% shared home blood pressure readings. Compared with White patients, non-Hispanic Black and Hispanic patients had lower odds of any access (Black: OR, 0.53; 95% CI, 0.52-0.54; Hispanic: OR, 0.66; 95% CI, 0.64-0.68), access around PCP visit time (Black: OR, 0.49; 95% CI, 0.48-0.50; Hispanic: OR, 0.62; 95% CI, 0.60-0.64), frequent access (Black: OR, 0.56; 95% CI, 0.55-0.57; Hispanic: OR, 0.71; 95% CI, 0.69-0.73), and messaging (Black: OR, 0.63; 95% CI, 0.61-0.64); Hispanic: OR, 0.71; 95% CI, 0.69-0.73). Conclusions and Relevance: This cohort study of patients with hypertension found clear sociodemographic disparities in patient portal engagement among those treated in primary care. Without special efforts to engage patients with portals, interventions that use patient portals to target hypertension may exacerbate disparities.

Identification of mycoplasma pneumonia in children based on fusion of multi-modal clinical free-text description and structured test data.

Mycoplasma pneumonia may lead to hospitalizations and pose life-threatening risks in children. The automated identification of mycoplasma pneumonia from electronic medical records holds significant potential for improving the efficiency of hospital resource allocation. In this study, we proposed a novel method for identifying mycoplasma pneumonia by integrating multi-modal features derived from both free-text descriptions and structured test data in electronic medical records. Our approach begins with the extraction of free-text and structured data from clinical records through a systematic preprocessing pipeline. Subsequently, we employ a pre-trained transformer language model to extract features from the free-text, while multiple additive regression trees are used to transform features from the structured data. An attention-based fusion mechanism is then applied to integrate these multi-modal features for effective classification. We validated our method using clinic records of 7157 patients, retrospectively collected for training and testing purposes. The experimental results demonstrate that our proposed multi-modal fusion approach achieves significant improvements over other methods across four key performance metrics.

Application of a data continuity prediction algorithm to an electronic health record-based pharmacoepidemiology study.

BACKGROUND AND OBJECTIVES: Use of algorithms to identify patients with high data-continuity in electronic health records (EHRs) may increase study validity. Practical experience with this approach remains limited. METHODS: We developed and validated four algorithms to identify patients with high data continuity in an EHR-based data source. Selected algorithms were then applied to a pharmacoepidemiologic study comparing rates of COVID-19 hospitalization in patients exposed to insulin versus noninsulin antidiabetic drugs. RESULTS: A model using a short list of five EHR-derived variables performed as well as more complex models to distinguish high- from low-data continuity patients. Higher data continuity was associated with more accurate ascertainment of key variables. In the pharmacoepidemiologic study, patients with higher data continuity had higher observed rates of the COVID-19 outcome and a large unadjusted association between insulin use and the outcome, but no association after propensity score adjustment. DISCUSSION: We found that a simple, portable algorithm to predict data continuity gave comparable performance to more complex methods. Use of the algorithm significantly impacted the results of an empirical study, with evidence of more valid results at higher levels of data continuity.