GWAS of longitudinal trajectories at biobank scale.

Biobanks linked to massive, longitudinal electronic health record (EHR) data make numerous new genetic research questions feasible. One among these is the study of biomarker trajectories. For example, high blood pressure measurements over visits strongly predict stroke onset, and consistently high fasting glucose and Hb1Ac levels define diabetes. Recent research reveals that not only the mean level of biomarker trajectories but also their fluctuations, or within-subject (WS) variability, are risk factors for many diseases. Glycemic variation, for instance, is recently considered an important clinical metric in diabetes management. It is crucial to identify the genetic factors that shift the mean or alter the WS variability of a biomarker trajectory. Compared to traditional cross-sectional studies, trajectory analysis utilizes more data points and captures a complete picture of the impact of time-varying factors, including medication history and lifestyle. Currently, there are no efficient tools for genome-wide association studies (GWASs) of biomarker trajectories at the biobank scale, even for just mean effects. We propose TrajGWAS, a linear mixed effect model-based method for testing genetic effects that shift the mean or alter the WS variability of a biomarker trajectory. It is scalable to biobank data with 100,000 to 1,000,000 individuals and many longitudinal measurements and robust to distributional assumptions. Simulation studies corroborate that TrajGWAS controls the type I error rate and is powerful. Analysis of eleven biomarkers measured longitudinally and extracted from UK Biobank primary care data for more than 150,000 participants with 1,800,000 observations reveals loci that significantly alter the mean or WS variability.

Prediction of Venous Thromboembolism in Diverse Populations Using Machine Learning and Structured Electronic Health Records.

BACKGROUND: Venous thromboembolism (VTE) is a major cause of morbidity and mortality worldwide. Current risk assessment tools, such as the Caprini and Padua scores and Wells criteria, have limitations in their applicability and accuracy. This study aimed to develop machine learning models using structured electronic health record data to predict diagnosis and 1-year risk of VTE. METHODS: We trained and validated models on data from 159 001 participants in the Mount Sinai Data Warehouse. We then externally tested them on 401 723 participants in the UK Biobank and 123 039 participants in All of Us. All data sets contain populations of diverse ancestries and clinical histories. We used these data sets to develop small, medium, and large models with increasing features on a range of optimizing portability to maximizing performance. We make trained models publicly available in click-and-run format at https://doi.org/10.17632/tkwzysr4y6.6. RESULTS: In the holdout and external test sets, respectively, models achieved areas under the receiver operating characteristic curve of 0.80 to 0.83 and 0.72 to 0.82 for VTE diagnosis prediction and 0.76 to 0.78 and 0.64 to 0.69 for 1-year risk prediction, significantly outperforming the Padua score. Models also demonstrated robust performance across different VTE types and patient subsets, including ethnicity, age, and surgical and hospitalization status. Models identified both established and novel clinical features contributing to VTE risk, offering valuable insights into its underlying pathophysiology. CONCLUSIONS: Machine learning models using structured electronic health record data can significantly improve VTE diagnosis and 1-year risk prediction in diverse populations. Model probability scores exist on a continuum, affecting mortality risk in both healthy individuals and VTE cases. Integrating these models into electronic health record systems to generate real-time predictions may enhance VTE risk assessment, early detection, and preventative measures, ultimately reducing the morbidity and mortality associated with VTE.

The clinical and genetic spectrum of paediatric speech and language disorders.

Speech and language disorders are known to have a substantial genetic contribution. Although frequently examined as components of other conditions, research on the genetic basis of linguistic differences as separate phenotypic subgroups has been limited so far. Here, we performed an in-depth characterization of speech and language disorders in 52 143 individuals, reconstructing clinical histories using a large-scale data-mining approach of the electronic medical records from an entire large paediatric healthcare network. The reported frequency of these disorders was the highest between 2 and 5 years old and spanned a spectrum of 26 broad speech and language diagnoses. We used natural language processing to assess the degree to which clinical diagnoses in full-text notes were reflected in ICD-10 diagnosis codes. We found that aphasia and speech apraxia could be retrieved easily through ICD-10 diagnosis codes, whereas stuttering as a speech phenotype was coded in only 12% of individuals through appropriate ICD-10 codes. We found significant comorbidity of speech and language disorders in neurodevelopmental conditions (30.31%) and, to a lesser degree, with epilepsies (6.07%) and movement disorders (2.05%). The most common genetic disorders retrievable in our analysis of electronic medical records were STXBP1 (n = 21), PTEN (n = 20) and CACNA1A (n = 18). When assessing associations of genetic diagnoses with specific linguistic phenotypes, we observed associations of STXBP1 and aphasia (P = 8.57 × 10-7, 95% confidence interval = 18.62-130.39) and MYO7A with speech and language development delay attributable to hearing loss (P = 1.24 × 10-5, 95% confidence interval = 17.46-infinity). Finally, in a sub-cohort of 726 individuals with whole-exome sequencing data, we identified an enrichment of rare variants in neuronal receptor pathways, in addition to associations of UQCRC1 and KIF17 with expressive aphasia, MROH8 and BCHE with poor speech, and USP37, SLC22A9 and UMODL1 with aphasia. In summary, our study outlines the landscape of paediatric speech and language disorders, confirming the phenotypic complexity of linguistic traits and novel genotype-phenotype associations. Subgroups of paediatric speech and language disorders differ significantly with respect to the composition of monogenic aetiologies.

High mortality rates among individuals misdiagnosed with tuberculosis: a matched retrospective cohort study of individuals diagnosed with tuberculosis in Brazil.

We conducted a matched retrospective cohort study comparing mortality among individuals receiving a false-positive tuberculosis diagnosis (n=3701) to individuals correctly diagnosed with TB (n=8595) in Brazil from 2007-2016. Over an average 5.4-year follow-up period, we estimated a mortality rate ratio of 1.95 (95% confidence interval: 1.80, 2.11) for individuals incorrectly diagnosed with TB compared to controls. The leading causes of death among the misdiagnosed were malignant neoplasms (40.9%) and respiratory system disorders (15.9%), conditions with symptoms similar to tuberculosis. Our findings highlight the need for improved follow-up care after identification of false-positive cases to increase survival for this high-risk population.

Trajectories of BMI before and after diagnosis of type 2 diabetes in a real-world population.

AIMS/HYPOTHESIS: Few studies have examined the clinical characteristics associated with changes in weight before and after diagnosis of type 2 diabetes. Using a large real-world cohort, we derived trajectories of BMI before and after diabetes diagnosis, and examined the clinical characteristics associated with these trajectories, including assessing the impact of pre-diagnosis weight change on post-diagnosis weight change. METHODS: We performed an observational cohort study using electronic medical records from individuals in the Scottish Care Information Diabetes Collaboration database. Two trajectories were calculated, based on observed BMI measurements between 3 years and 6 months before diagnosis and between 1 and 5 years after diagnosis. In the post-diagnosis trajectory, each BMI measurement was time-dependently adjusted for the effects of diabetes medications and HbA1c change. RESULTS: A total of 2736 individuals were included in the study. There was a pattern of pre-diagnosis weight gain, with 1944 individuals (71%) gaining weight overall, and 875 (32%) gaining more than 0.5 kg/m2 per year. This was followed by a pattern of weight loss after diagnosis, with 1722 individuals (63%) losing weight. Younger age and greater social deprivation were associated with increased weight gain before diagnosis. Pre-diagnosis weight change was unrelated to post-diagnosis weight change, but post-diagnosis weight loss was associated with older age, female sex, higher BMI, higher HbA1c and weight gain during the peri-diagnosis period. When considering the peri-diagnostic period (defined as from 6 months before to 12 months after diagnosis), we identified 986 (36%) individuals who had a high HbA1c at diagnosis but who lost weight rapidly and were most aggressively treated at 1 year; this subgroup had the best glycaemic control at 5 years. CONCLUSIONS/INTERPRETATION: Average weight increases before diagnosis and decreases after diagnosis; however, there were significant differences across the population in terms of weight changes. Younger individuals gained weight pre-diagnosis, but, in older individuals, type 2 diabetes is less associated with weight gain, consistent with other drivers for diabetes aetiology in older adults. We have identified a substantial group of individuals who have a rapid deterioration in glycaemic control, together with weight loss, around the time of diagnosis, and who subsequently stabilise, suggesting that a high HbA1c at diagnosis is not inevitably associated with a poor outcome and may be driven by reversible glucose toxicity.

The role and limitations of electronic medical records versus patient interviews for determining symptoms of, underlying comorbidities of, and medication use by patients with COVID-19.

Electronic medical records (EMRs) are important for rapidly compiling information to determine disease characteristics (eg, symptoms) and risk factors (eg, underlying comorbidities, medications) for disease-related outcomes. To assess EMR data accuracy, agreement between EMR abstractions and patient interviews was evaluated. Symptoms, medical history, and medication use among patients with COVID-19 collected from EMRs and patient interviews were compared using overall agreement (ie, same answer in EMR and interview), reported agreement (yes answer in both EMR and interview among those who reported yes in either), and κ statistics. Overall, patients reported more symptoms in interviews than in EMR abstractions. Overall agreement was high (≥50% for 20 of 23 symptoms), but only subjective fever and dyspnea had reported agreement of ≥50%. The κ statistics for symptoms were generally low. Reported medical conditions had greater agreement with all condition categories (n = 10 of 10) having ≥50% overall agreement and half (n = 5 of 10) having ≥50% reported agreement. More nonprescription medications were reported in interviews than in EMR abstractions, leading to low reported agreement (28%). Discordance was observed for symptoms, medical history, and medication use between EMR abstractions and patient interviews. Investigations using EMRs to describe clinical characteristics and identify risk factors should consider the potential for incomplete data, particularly for symptoms and medications.

Impact of COVID-19 pandemic on depression incidence and healthcare service use among patients with depression: an interrupted time-series analysis from a 9-year population-based study.

BACKGROUND: Most studies on the impact of the COVID-19 pandemic on depression burden focused on the earlier pandemic phase specific to lockdowns, but the longer-term impact of the pandemic is less well-studied. In this population-based cohort study, we examined the short-term and long-term impacts of COVID-19 on depression incidence and healthcare service use among patients with depression. METHODS: Using the territory-wide electronic medical records in Hong Kong, we identified all patients aged ≥ 10 years with new diagnoses of depression from 2014 to 2022. We performed an interrupted time-series (ITS) analysis to examine changes in incidence of medically attended depression before and during the pandemic. We then divided all patients into nine cohorts based on year of depression incidence and studied their initial and ongoing service use patterns until the end of 2022. We applied generalized linear modeling to compare the rates of healthcare service use in the year of diagnosis between patients newly diagnosed before and during the pandemic. A separate ITS analysis explored the pandemic impact on the ongoing service use among prevalent patients with depression. RESULTS: We found an immediate increase in depression incidence (RR = 1.21, 95% CI: 1.10-1.33, p < 0.001) in the population after the pandemic began with non-significant slope change, suggesting a sustained effect until the end of 2022. Subgroup analysis showed that the increases in incidence were significant among adults and the older population, but not adolescents. Depression patients newly diagnosed during the pandemic used 11% fewer resources than the pre-pandemic patients in the first diagnosis year. Pre-existing depression patients also had an immediate decrease of 16% in overall all-cause service use since the pandemic, with a positive slope change indicating a gradual rebound over a 3-year period. CONCLUSIONS: During the pandemic, service provision for depression was suboptimal in the face of increased demand generated by the increasing depression incidence during the COVID-19 pandemic. Our findings indicate the need to improve mental health resource planning preparedness for future public health crises.

Development and Implementation of a Dynamically Updated Big Data Intelligence Platform Using Electronic Medical Records for Secondary Hypertension.

Background: The accurate identification and diagnosis of secondary hypertension is critical,especially while cardiovascular heart disease continues to be the leading cause of death. To develop a big data intelligence platform for secondary hypertension using electronic medical records to contribute to future basic and clinical research. Methods: Using hospital data, the platform, named Hypertension DATAbase at Urumchi (UHDATA), included patients diagnosed with hypertension at the People's Hospital of Xinjiang Uygur Autonomous Region since December 2004. The electronic data acquisition system, the database synchronization technology, and data warehouse technology (extract-transform-load, ETL) for the scientific research big data platform were used to synchronize and extract the data from each business system in the hospital. Standard data elements were established for the platform, including demographic and medical information. To facilitate the research, the database was also linked to the sample database system, which includes blood samples, urine specimens, and tissue specimens. Results: From December 17, 2004, to August 31, 2022, a total of 295,297 hypertensive patients were added to the platform, with 53.76% being males, with a mean age of 59 years, and 14% with secondary hypertension. However, 75,802 patients visited the Hypertension Center at our hospital, with 43% (32,595 patients) being successfully diagnosed with secondary hypertension. The database contains 1458 elements, with an average fill rate of 90%. The database can continuously include the data for new hypertensive patients and add new data for existing hypertensive patients, including post-discharge follow-up information, and the database updates every 2 weeks. Presently, some studies that are based on the platform have been published. Conclusions: Using computer information technology, we developed and implemented a big database of dynamically updating electronic medical records for patients with hypertension, which is helpful in promoting future research on secondary hypertension.

Noninvasive Risk Prediction Models for Heart Failure Using Proportional Jaccard Indices and Comorbidity Patterns.

Background: In the post-coronavirus disease 2019 (COVID-19) era, remote diagnosis and precision preventive medicine have emerged as pivotal clinical medicine applications. This study aims to develop a digital health-monitoring tool that utilizes electronic medical records (EMRs) as the foundation for performing a non-random correlation analysis among different comorbidity patterns for heart failure (HF). Methods: Novel similarity indices, including proportional Jaccard index (PJI), multiplication of the odds ratio proportional Jaccard index (OPJI), and alpha proportional Jaccard index (APJI), provide a fundamental framework for constructing machine learning models to predict the risk conditions associated with HF. Results: Our models were constructed for different age groups and sexes and yielded accurate predictions of high-risk HF across demographics. The results indicated that the optimal prediction model achieved a notable accuracy of 82.1% and an area under the curve (AUC) of 0.878. Conclusions: Our noninvasive HF risk prediction system is based on historical EMRs and provides a practical approach. The proposed indices provided simple and straightforward comparative indicators of comorbidity pattern matching within individual EMRs. All source codes developed for our noninvasive prediction models can be retrieved from GitHub.

Undetected suicide attempts among U.S. soldiers: results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS).

BACKGROUND: While previous studies have reported high rates of documented suicide attempts (SAs) in the U.S. Army, the extent to which soldiers make SAs that are not identified in the healthcare system is unknown. Understanding undetected suicidal behavior is important in broadening prevention and intervention efforts. METHODS: Representative survey of U.S. Regular Army enlisted soldiers (n = 24 475). Reported SAs during service were compared with SAs documented in administrative medical records. Logistic regression analyses examined sociodemographic characteristics differentiating soldiers with an undetected SA v. documented SA. Among those with an undetected SA, chi-square tests examined characteristics associated with receiving a mental health diagnosis (MH-Dx) prior to SA. Discrete-time survival analysis estimated risk of undetected SA by time in service. RESULTS: Prevalence of undetected SA (unweighted n = 259) was 1.3%. Annual incidence was 255.6 per 100 000 soldiers, suggesting one in three SAs are undetected. In multivariable analysis, rank ⩾E5 (OR = 3.1[95%CI 1.6-5.7]) was associated with increased odds of undetected v. documented SA. Females were more likely to have a MH-Dx prior to their undetected SA (Rao-Scott χ21 = 6.1, p = .01). Over one-fifth of undetected SAs resulted in at least moderate injury. Risk of undetected SA was greater during the first four years of service. CONCLUSIONS: Findings suggest that substantially more soldiers make SAs than indicated by estimates based on documented attempts. A sizable minority of undetected SAs result in significant injury. Soldiers reporting an undetected SA tend to be higher ranking than those with documented SAs. Undetected SAs require additional approaches to identifying individuals at risk.

Health Data Sciences and Cardiovascular Diseases in South Asia: Innovations and Challenges in Digital Health.

PURPOSE OF REVIEW: Health data sciences can help mitigate high burden of cardiovascular disease (CVD) management in South Asia by increasing availability and affordability of healthcare services. This review explores the current landscape, challenges, and strategies for leveraging digital health technologies to improve CVD outcomes in the region. RECENT FINDINGS: Several South Asian countries are implementing national digital health strategies that aim to provide unique health account numbers for patients, creating longitudinal digital health records while others aim to digitize healthcare services and improve health outcomes. Significant challenges impede progress, including lack of interoperability, inadequate training of healthcare workers, cultural barriers, and data privacy concerns. Leveraging digital health for CVD management involves using big data for early detection, employing artificial intelligence for diagnostics, and integrating multiomics data for health insights. Addressing these challenges through policy frameworks, capacity building, and international cooperation is crucial for improving CVD outcomes in region.

A Comparison of Sodium-Glucose Co-Transporter 2 Inhibitor Kidney Outcome Trial Participants with a Real-World Chronic Kidney Disease Primary Care Population.

BACKGROUND/HYPOTHESIS: Observational studies suggest sodium-glucose co-transporter-2 (SGLT2) inhibitor kidney outcome trials are not representative of the broader population of people with chronic kidney disease (CKD). However, there are limited data on the generalisability to those without co-existing type 2 diabetes (T2D), and the representativeness of the EMPA-KIDNEY trial has not been adequately explored. We hypothesised that SGLT2 inhibitor kidney outcome trials are more representative of people with co-existing T2D than those without, and that EMPA-KIDNEY is more representative than previous trials. METHODS: A cross-sectional analysis of adults with CKD in English primary care was conducted using the Oxford-Royal College of General Practitioners Clinical Information Digital Hub. The proportions that met the eligibility criteria of SGLT2 inhibitor kidney outcome trials were determined, and their characteristics described. Logistic regression analyses were performed to identify factors associated with trial eligibility. RESULTS: Of 6,670,829 adults, 516,491 (7.7%) with CKD were identified. In the real-world CKD population, 0.9%, 2.2%, and 8.0% met the CREDENCE, DAPA-CKD, and EMPA-KIDNEY eligibility criteria, respectively. All trials were more representative of people with co-existing T2D than those without T2D. Trial participants were 9-14 years younger than the real-world CKD population, and had more advanced CKD, including higher levels of albuminuria. A higher proportion of the CREDENCE (100%), DAPA-CKD (67.6%) and EMPA-KIDNEY (44.5%) trial participants had T2D compared to the real-world CKD population (32.8%). Renin-angiotensin system inhibitors were prescribed in almost all trial participants, compared to less than half of the real-world CKD population. Females were under-represented and less likely to be eligible for the trials. CONCLUSION: SGLT2 inhibitor kidney outcome trials represent a sub-group of people with CKD at high risk of adverse kidney events. Out study highlights the importance of complementing trials with real-world studies, exploring the effectiveness of SGLT2 inhibitors in the broader population of people with CKD.

Three-phase generalized raking and multiple imputation estimators to address error-prone data.

Validation studies are often used to obtain more reliable information in settings with error-prone data. Validated data on a subsample of subjects can be used together with error-prone data on all subjects to improve estimation. In practice, more than one round of data validation may be required, and direct application of standard approaches for combining validation data into analyses may lead to inefficient estimators since the information available from intermediate validation steps is only partially considered or even completely ignored. In this paper, we present two novel extensions of multiple imputation and generalized raking estimators that make full use of all available data. We show through simulations that incorporating information from intermediate steps can lead to substantial gains in efficiency. This work is motivated by and illustrated in a study of contraceptive effectiveness among 83 671 women living with HIV, whose data were originally extracted from electronic medical records, of whom 4732 had their charts reviewed, and a subsequent 1210 also had a telephone interview to validate key study variables.

Effectiveness of patient reminders on influenza vaccination coverage among adults with chronic conditions: A feasibility study in Australian general practices.

BACKGROUND: Influenza vaccination is recommended for Australians 18+ years old with medical risk factors, but coverage is suboptimal. We aimed to examine whether automatic, opportunistic patient reminders (SMS and/or printed) before appointments with a general practitioner increased influenza vaccination uptake. METHODS: This clustered non-randomised feasibility study in Australian general practice included patients aged 18-64 years with at least one medical risk factor attending participating practices between May and September 2021. Software installed at intervention practices identified unvaccinated eligible patients when they booked an appointment, sent vaccination reminders (SMS on booking and 1 h before appointments), and printed automatic reminders on arrival. Control practices provided usual care. Clustered analyses adjusted for sociodemographic differences among practices were performed using logistic regression. RESULTS: A total of 12,786 at-risk adults attended 16 intervention practices (received reminders = 4066; 'internal control' receiving usual care = 8720), and 5082 individuals attended eight control practices. Baseline influenza vaccination uptake (2020) was similar in intervention and control practices (∼34%). After the intervention, uptake was similar in all groups (control practices = 29.3%; internal control = 30.0%; intervention = 31.6% (p-value = 0.203). However, SMS 1 h before appointments increased vaccination coverage (39.3%, adjusted OR = 1.65; 95%CI 1.20;2.27; number necessary to treat = 13), especially when combined with other reminder forms. That effect was more evident among adults with chronic respiratory, rheumatologic, or inflammatory bowel disease. CONCLUSION: These findings indicate that automated SMS reminders delivered at proximate times to appointments are a low-cost strategy to increase influenza vaccination among adults at higher risk of severe disease attending Australian general practices.

Validity of ICD-10 diagnosis codes for placenta accreta spectrum disorders.

BACKGROUND: The 10th revision of the International Classification of Diseases, Clinical Modification (ICD-10) includes diagnosis codes for placenta accreta spectrum for the first time. These codes could enable valuable research and surveillance of placenta accreta spectrum, a life-threatening pregnancy complication that is increasing in incidence. OBJECTIVE: We sought to evaluate the validity of placenta accreta spectrum diagnosis codes that were introduced in ICD-10 and assess contributing factors to incorrect code assignments. METHODS: We calculated sensitivity, specificity, positive predictive value and negative predictive value of the ICD-10 placenta accreta spectrum code assignments after reviewing medical records from October 2015 to March 2020 at a quaternary obstetric centre. Histopathologic diagnosis was considered the gold standard. RESULTS: Among 22,345 patients, 104 (0.46%) had an ICD-10 code for placenta accreta spectrum and 51 (0.23%) had a histopathologic diagnosis. ICD-10 codes had a sensitivity of 0.71 (95% CI 0.56, 0.83), specificity of 0.98 (95% CI 0.93, 1.00), positive predictive value of 0.61 (95% CI 0.48, 0.72) and negative predictive value of 1.00 (95% CI 0.96, 1.00). The sensitivities of the ICD-10 codes for placenta accreta spectrum subtypes- accreta, increta and percreta-were 0.55 (95% CI 0.31, 0.78), 0.33 (95% CI 0.12, 0.62) and 0.56 (95% CI 0.31, 0.78), respectively. Cases with incorrect code assignment were less morbid than cases with correct code assignment, with a lower incidence of hysterectomy at delivery (17% vs 100%), blood transfusion (26% vs 75%) and admission to the intensive care unit (0% vs 53%). Primary reasons for code misassignment included code assigned to cases of occult placenta accreta (35%) or to cases with clinical evidence of placental adherence without histopatholic diagnostic (35%) features. CONCLUSION: These findings from a quaternary obstetric centre suggest that ICD-10 codes may be useful for research and surveillance of placenta accreta spectrum, but researchers should be aware of likely substantial false positive cases.

Topical amitriptyline in burning mouth syndrome: A retrospective real-world evidence study.

OBJECTIVE: To evaluate the effectiveness, tolerability, and safety of topical amitriptyline as a potential route of administration for the management of burning mouth syndrome. BACKGROUND: Burning mouth syndrome is a complex, idiopathic, and debilitating orofacial pain disorder that impairs quality of life, with a prevalence of up to 18% in menopausal women. Available drugs to alleviate its burning sensation have inconsistent and limited efficacy. Given its physicochemical properties, excellent tolerability, and ability to target peripheral pathways, topical amitriptyline seems a promising mechanistically specific analgesic drug for burning mouth syndrome. METHODS: In this retrospective cross-sectional real-world evidence study, patients with burning mouth syndrome who were prescribed topical amitriptyline for 8 weeks were identified. Eligibility criteria stemmed from ICHD-3, ICOP, and consensus definitions. The primary outcome measure was mean daily pain intensity (on a 0-10 scale); secondary outcomes included adverse events and patient global impression of improvement. Data are given as the mean ± SD. RESULTS: A total of 15 patients fulfilling the eligibility criteria were included and analyzed. Mean daily pain was 6.7 ± 2.1 at baseline and 3.7 ± 2.3 after treatment, with a mean reduction of 3.1 ± 2.8 (p = 0.002). Half of the patients experienced a decrease in pain by at least 50% (p = 0.008). Several mild adverse events were reported, such as somnolence or dry mouth. CONCLUSIONS: Topical amitriptyline may be a safe and potent route of administration in the treatment of burning mouth syndrome, a hypothesis to be tested in further controlled trials.

Weather conditions and legionellosis: a nationwide case-crossover study among Medicare recipients.

Legionellosis is a respiratory infection caused by Legionella sp. that is found in water and soil. Infection may cause pneumonia (Legionnaires' Disease) and a milder form (Pontiac Fever). Legionella colonizes water systems and results in exposure by inhalation of aerosolized bacteria. The incubation period ranges from 2 to 14 days. Precipitation and humidity may be associated with increased risk. We used Medicare records from 1999 to 2020 to identify hospitalizations for legionellosis. Precipitation, temperature, and relative humidity were obtained from the PRISM Climate Group for the zip code of residence. We used a time-stratified bi-directional case-crossover design with lags of 20 days. Data were analyzed using conditional logistic regression and distributed lag non-linear models. A total of 37 883 hospitalizations were identified. Precipitation and relative humidity at lags 8 through 13 days were associated with an increased risk of legionellosis. The strongest association was precipitation at day 10 lag (OR = 1.08, 95% CI = 1.05-1.11 per 1 cm). Over 20 days, 3 cm of precipitation increased the odds of legionellosis over four times. The association was strongest in the Northeast and Midwest and during summer and fall. Precipitation and humidity were associated with hospitalization among Medicare recipients for legionellosis at lags consistent with the incubation period for infection.

Learning difficulties often not documented in newly diagnosed focal epilepsy.

OBJECTIVE: A previous investigation of people with newly diagnosed focal epilepsy participating in the Human Epilepsy Project 1 (HEP1) revealed an association between learning difficulties and structural brain differences, suggesting an underlying relationship prior to seizure onset. To investigate physicians' practices of documentation learning difficulties during clinical encounters, we conducted a review of initial epileptologist encounter notes from HEP1 participants who self-reported early life learning difficulties separately as part of study enrollment. METHODS: HEP1 enrolled 67 North American participants between June 2012 and November 2017 who self-reported one or more difficulties with learning (i.e., having repeated grade, receiving learning support/remediation, and/or formal diagnosis of a learning disability) prior to epilepsy diagnosis as part of the study enrollment. The epileptologist's initial encounter note was then reviewed in detail for each of these participants. Documentation of learning issues and specific diagnoses of learning disabilities was compared to participant characteristics. Regression analysis was used to test for any independent associations between participant characteristics and physician documentation of learning difficulties. RESULTS: There were significant independent relationships between age, sex, and physician documentation of learning difficulties. On average, participants ages 22 and younger were 12.12 times more likely to have their learning difficulties documented compared to those 23 years and older (95 % CI: 2.226 to 66.02, p = 0.004). Additionally, male participants had 7.2 times greater odds of having their learning difficulty documented compared to female participants (95 % CI: 1.538 to 33.717, p = 0.012). There were no significant independent associations between race, language, employment, or geographical region. SIGNIFICANCE: These findings highlight disparities in physician documentation for people with newly diagnosed focal epilepsy and a history of learning difficulties. In the HEP1 cohort, physicians were more likely to document learning difficulties in males and in younger individuals. Systematic practice standards are important for reducing healthcare disparities across populations, improving clinical care to individuals, as well as enabling more accurate retrospective study of clinical phenomenon.

Standardizing visual display of gastrointestinal pathology results for improved clinical interpretation.

OBJECTIVES: Pathology is an essential component of disease diagnosis and management in pediatric gastroenterology. Pathology reports have not been standardized in some areas of pediatric gastrointestinal pathology and pathology reporting varies. Development of electronic medical record (EMR) pathology synoptic report templates (PSRT) enables pathology data collection in a specific format and can help standardize pathology reporting. We developed, implemented, and evaluated EMR PSRTs for eosinophilic esophagitis (EoE) and inflammatory bowel disease (IBD). METHODS: PSRTs were developed by a multidisciplinary team of pediatric experts of allergy, gastroenterology, and pathology for both EoE and IBD based on available literature and validated scales. Likert surveys (range 1 low acceptance to 5 high acceptance) based on the Technology Acceptance Model assessed user acceptance of the developed PSRTs. The use of PSRTs was monitored via control charts. RESULTS: Overall, evaluation questionnaires achieved >80% response rates. Clinicians and pathologists reported moderate-to-high levels of Perceived Usefulness (median (interquartile range) for EoE PSRT: clinicians 4.0 (4.0, 5.0) and pathologists 3.5 (3.5, 4.0); and IBD PSRT: clinicians 4.0 (3.0, 4.0) and pathologists 4.0 (4.0, 5.0)) and Perceived Ease of Use (EoE PSRT: clinicians 4.5 (4.0, 5.0) and pathologists 4.0 (4.0, 4.0); and IBD PSRT: clinicians 4.0 (4.0, 5.0) and pathologists 4.0 (4.0, 5.0)) of the developed PSRTs. Control charts demonstrated 100% utilization by 2-5 months from launch. CONCLUSIONS: We demonstrate successful implementation of synoptic reporting for both pediatric EoE and IBD pathology. EMR synoptic reporting provides standardization of pathology reporting and improved methods of pathology data presentation, which could potentially optimize provider efficiency, clinician interpretation of pathology results and disease trajectory, patient care, and clinician satisfaction.

Validation of GPT-4 for clinical event classification: A comparative analysis with ICD codes and human reviewers.

BACKGROUND AND AIM: Effective clinical event classification is essential for clinical research and quality improvement. The validation of artificial intelligence (AI) models like Generative Pre-trained Transformer 4 (GPT-4) for this task and comparison with conventional methods remains unexplored. METHODS: We evaluated the performance of the GPT-4 model for classifying gastrointestinal (GI) bleeding episodes from 200 medical discharge summaries and compared the results with human review and an International Classification of Diseases (ICD) code-based system. The analysis included accuracy, sensitivity, and specificity evaluation, using ground truth determined by physician reviewers. RESULTS: GPT-4 exhibited an accuracy of 94.4% in identifying GI bleeding occurrences, outperforming ICD codes (accuracy 63.5%, P < 0.001). GPT-4's accuracy was either slightly lower or statistically similar to individual human reviewers (Reviewer 1: 98.5%, P < 0.001; Reviewer 2: 90.8%, P = 0.170). For location classification, GPT-4 achieved accuracies of 81.7% and 83.5% for confirmed and probable GI bleeding locations, respectively, with figures that were either slightly lower or comparable with those of human reviewers. GPT-4 was highly efficient, analyzing the dataset in 12.7 min at a cost of 21.2 USD, whereas human reviewers required 8-9 h each. CONCLUSION: Our study indicates GPT-4 offers a reliable, cost-efficient, and faster alternative to current clinical event classification methods, outperforming the conventional ICD coding system and performing comparably to individual expert human reviewers. Its implementation could facilitate more accurate and granular clinical research and quality audits. Future research should explore scalability, prompt and model tuning, and ethical implications of high-performance AI models in clinical data processing.