Large-Scale Assessment of a Smartwatch to Identify Atrial Fibrillation.

BACKGROUND: Optical sensors on wearable devices can detect irregular pulses. The ability of a smartwatch application (app) to identify atrial fibrillation during typical use is unknown. METHODS: Participants without atrial fibrillation (as reported by the participants themselves) used a smartphone (Apple iPhone) app to consent to monitoring. If a smartwatch-based irregular pulse notification algorithm identified possible atrial fibrillation, a telemedicine visit was initiated and an electrocardiography (ECG) patch was mailed to the participant, to be worn for up to 7 days. Surveys were administered 90 days after notification of the irregular pulse and at the end of the study. The main objectives were to estimate the proportion of notified participants with atrial fibrillation shown on an ECG patch and the positive predictive value of irregular pulse intervals with a targeted confidence interval width of 0.10. RESULTS: We recruited 419,297 participants over 8 months. Over a median of 117 days of monitoring, 2161 participants (0.52%) received notifications of irregular pulse. Among the 450 participants who returned ECG patches containing data that could be analyzed - which had been applied, on average, 13 days after notification - atrial fibrillation was present in 34% (97.5% confidence interval [CI], 29 to 39) overall and in 35% (97.5% CI, 27 to 43) of participants 65 years of age or older. Among participants who were notified of an irregular pulse, the positive predictive value was 0.84 (95% CI, 0.76 to 0.92) for observing atrial fibrillation on the ECG simultaneously with a subsequent irregular pulse notification and 0.71 (97.5% CI, 0.69 to 0.74) for observing atrial fibrillation on the ECG simultaneously with a subsequent irregular tachogram. Of 1376 notified participants who returned a 90-day survey, 57% contacted health care providers outside the study. There were no reports of serious app-related adverse events. CONCLUSIONS: The probability of receiving an irregular pulse notification was low. Among participants who received notification of an irregular pulse, 34% had atrial fibrillation on subsequent ECG patch readings and 84% of notifications were concordant with atrial fibrillation. This siteless (no on-site visits were required for the participants), pragmatic study design provides a foundation for large-scale pragmatic studies in which outcomes or adherence can be reliably assessed with user-owned devices. (Funded by Apple; Apple Heart Study ClinicalTrials.gov number, NCT03335800.).

Lessons learned in the Apple Heart Study and implications for the data management of future digital clinical trials.

The digital clinical trial is fast emerging as a pragmatic trial that can improve a trial's design including recruitment and retention, data collection and analytics. To that end, digital platforms such as electronic health records or wearable technologies that enable passive data collection can be leveraged, alleviating burden from the participant and study coordinator. However, there are challenges. For example, many of these data sources not originally intended for research may be noisier than traditionally obtained measures. Further, the secure flow of passively collected data and their integration for analysis is non-trivial. The Apple Heart Study was a prospective, single-arm, site-less digital trial designed to evaluate the ability of an app to detect atrial fibrillation. The study was designed with pragmatic features, such as an app for enrollment, a wearable device (the Apple Watch) for data collection, and electronic surveys for participant-reported outcomes that enabled a high volume of patient enrollment and accompanying data. These elements led to challenges including identifying the number of unique participants, maintaining participant-level linkage of multiple complex data streams, and participant adherence and engagement. Novel solutions were derived that inform future designs with an emphasis on data management. We build upon the excellent framework of the Clinical Trials Transformation Initiative to provide a comprehensive set of guidelines for data management of the digital clinical trial that include an increased role of collaborative data scientists in the design and conduct of the modern digital trial.

Discovery of Distinct Immune Phenotypes Using Machine Learning in Pulmonary Arterial Hypertension.

RATIONALE: Accumulating evidence implicates inflammation in pulmonary arterial hypertension (PAH) and therapies targeting immunity are under investigation, although it remains unknown if distinct immune phenotypes exist. OBJECTIVE: Identify PAH immune phenotypes based on unsupervised analysis of blood proteomic profiles. METHODS AND RESULTS: In a prospective observational study of group 1 PAH patients evaluated at Stanford University (discovery cohort; n=281) and University of Sheffield (validation cohort; n=104) between 2008 and 2014, we measured a circulating proteomic panel of 48 cytokines, chemokines, and factors using multiplex immunoassay. Unsupervised machine learning (consensus clustering) was applied in both cohorts independently to classify patients into proteomic immune clusters, without guidance from clinical features. To identify central proteins in each cluster, we performed partial correlation network analysis. Clinical characteristics and outcomes were subsequently compared across clusters. Four PAH clusters with distinct proteomic immune profiles were identified in the discovery cohort. Cluster 2 (n=109) had low cytokine levels similar to controls. Other clusters had unique sets of upregulated proteins central to immune networks-cluster 1 (n=58; TRAIL [tumor necrosis factor-related apoptosis-inducing ligand], CCL5 [C-C motif chemokine ligand 5], CCL7, CCL4, MIF [macrophage migration inhibitory factor]), cluster 3 (n=77; IL [interleukin]-12, IL-17, IL-10, IL-7, VEGF [vascular endothelial growth factor]), and cluster 4 (n=37; IL-8, IL-4, PDGF-ß [platelet-derived growth factor beta], IL-6, CCL11). Demographics, PAH clinical subtypes, comorbidities, and medications were similar across clusters. Noninvasive and hemodynamic surrogates of clinical risk identified cluster 1 as high-risk and cluster 3 as low-risk groups. Five-year transplant-free survival rates were unfavorable for cluster 1 (47.6%; 95% CI, 35.4%-64.1%) and favorable for cluster 3 (82.4%; 95% CI, 72.0%-94.3%; across-cluster P<0.001). Findings were replicated in the validation cohort, where machine learning classified 4 immune clusters with comparable proteomic, clinical, and prognostic features. CONCLUSIONS: Blood cytokine profiles distinguish PAH immune phenotypes with differing clinical risk that are independent of World Health Organization group 1 subtypes. These phenotypes could inform mechanistic studies of disease pathobiology and provide a framework to examine patient responses to emerging therapies targeting immunity.

Clinical trials in a COVID-19 pandemic: Shared infrastructure for continuous learning in a rapidly changing landscape.

BACKGROUND: Clinical trials, conducted efficiently and with the utmost integrity, are a key component in identifying effective vaccines, therapies, and other interventions urgently needed to solve the COVID-19 crisis. Yet launching and implementing trials with the rigor necessary to produce convincing results is a complicated and time-consuming process. Balancing rigor and efficiency involves relying on designs that employ flexible features to respond to a fast-changing landscape, measuring valid endpoints that result in translational actions and disseminating findings in a timely manner. We describe the challenges involved in creating infrastructure with potential utility for shared learning. METHODS: We have established a shared infrastructure that borrows strength across multiple trials. The infrastructure includes an endpoint registry to aid in selecting appropriate endpoints, a registry to facilitate establishing a Data & Safety Monitoring Board, common data collection instruments, a COVID-19 dedicated design and analysis team, and a pragmatic platform protocol, among other elements. RESULTS: The authors have relied on the shared infrastructure for six clinical trials for which they serve as the Data Coordinating Center and have a design and analysis team comprising 15 members who are dedicated to COVID-19. The authors established a pragmatic platform to simultaneously investigate multiple treatments for the outpatient with adaptive features to add or drop treatment arms. CONCLUSION: The shared infrastructure provides appealing opportunities to evaluate disease in a more robust manner with fewer resources and is especially valued during a pandemic where efficiency in time and resources is crucial. The most important element of the shared infrastructure is the pragmatic platform. While it may be the most challenging of the elements to establish, it may provide the greatest benefit to both patients and researchers.

Genetic analyses uncover pleiotropic compensatory roles for Drosophila Nucleobindin-1 in inositol trisphosphate-mediated intracellular calcium homeostasis.

Nucleobindin-1 is an EF-hand calcium-binding protein with a distinctive profile, predominantly localized to the Golgi in insect and wide-ranging vertebrate cell types, alike. Its putative involvements in intracellular calcium (Ca2+) homeostasis have never been phenotypically characterized in any model organism. We have analyzed an adult-viable mutant that completely disrupts the G protein α-subunit binding and activating (GBA) motif of Drosophila Nucleobindin-1 (dmNUCB1). Such disruption does not manifest any obvious fitness-related, morphological/developmental, or behavioral abnormalities. A single copy of this mutation or the knockdown of dmnucb1 in restricted sets of cells variously rescues pleiotropic mutant phenotypes arising from impaired inositol 1,4,5-trisphosphate receptor (IP3R) activity (in turn depleting cytoplasmic Ca2+ levels across diverse tissue types). Additionally, altered dmNUCB1 expression or function considerably reverses lifespan and mobility improvements effected by IP3R mutants, in a Drosophila model of amyotrophic lateral sclerosis. Homology modeling-based analyses further predict a high degree of conformational conservation in Drosophila, of biochemically validated structural determinants in the GBA motif that specify in vertebrates, the unconventional Ca2+-regulated interaction of NUCB1 with Gαi subunits. The broad implications of our findings are hypothetically discussed, regarding potential roles for NUCB1 in GBA-mediated, Golgi-associated Ca2+ signaling, in health and disease.

Paper-based device for the colorimetric assay of bilirubin based on in-situ formation of gold nanoparticles.

A paper-based colorimetric assay for the determination of bilirubin has been developed. The method is based on the in-situ reduction of chloroauric acid to form gold nanoparticles. A chromatographic paper was patterned using a wax printer. Chloroauric acid was drop-cast onto the reagent zone. In the presence of bilirubin, gold(III) ions are reduced and form gold nanoparticles. This leads to a color change from yellow to purple. The intensity of the purple color (peak at 530 nm) increases with bilirubin concentration in the 5.0 to 1000 mg L-1 range. The detection limit is 1.0 mg L-1. For the quantification of bilirubin, images were captured using a digital camera, and data were processed with the help of machine learning-based supervised prediction using Random Forest classification. The method was applied to the determination of bilirubin in urine samples. The spiked urine samples exhibit more than 95% recovery. Graphical abstractSchematic representation of the paper-based colorimetric assay for the detection of bilirubin based on the in-situ formation of gold nanoparticles. A color band is generated for visual interpretation and used for the testing of bilirubin in urine.

Readying the workforce: evaluation of VHA's comprehensive women's health primary care provider initiative.

BACKGROUND: Veterans Health Administration (VHA) primary care providers (PCPs) often see few women, making it challenging to maintain proficiency in women's health (WH). Therefore, VHA in 2010 established Designated WH Providers, who would maintain proficiency in comprehensive WH care and be preferentially assigned women patients. OBJECTIVE: To evaluate early implementation of this national policy. METHODS: At each VHA health care system (N=140), the Women Veterans Program Manager completed a Fiscal Year 2012 workforce capacity assessment (response rate, 100%), representing the first time the national Designated WH Provider workforce had been identified. Assessment data were linked to administrative data. RESULTS: Of all VHA PCPs, 23% were Designated WH Providers; 100% of health care systems and 83% of community clinics had at least 1 Designated WH Provider. On average, women veterans comprised 19% (SD=27%) of the patients Designated WH Providers saw in primary care, versus 5% (SD=7%) for Other PCPs (P<0.001). For women veterans using primary care (N=313,033), new patients were less likely to see a Designated WH Provider than established women veteran patients (52% vs. 64%; P<0.001). CONCLUSIONS: VHA has achieved its goal of a Designated WH Provider in every health care system, and is approaching its goal of a Designated WH Provider at every hospital/community clinic. Designated WH Providers see more women than do Other PCPs. However, as the volume of women patients remains low for many providers, attention to alternative approaches to maintaining proficiency may prove necessary, and barriers to assigning new women patients to Designated WH Providers merit attention.

Reproductive health diagnoses of women veterans using department of Veterans Affairs health care.

BACKGROUND: Little is known regarding the reproductive health needs of women Veterans using Department of Veterans Affairs (VA) health care. OBJECTIVE: To describe the reproductive health diagnoses of women Veterans using VA health care, how these diagnoses differ across age groups, and variations in sociodemographic and clinical characteristics by presence of reproductive health diagnoses. RESEARCH DESIGN: This study is a cross-sectional analysis of VA administrative and clinical data. SUBJECTS: The study included women Veterans using VA health care in FY10. MEASURES: Reproductive health diagnoses were identified through presence of International Classification of Disease, 9th Revision (ICD-9) codes in VA clinical and administrative records. The prevalence of specific diagnosis categories were examined by age group (18-44, 45-64, ≥65 y) and the most frequent diagnoses for each age group were identified. Sociodemographic and clinical characteristics were compared by presence of at least 1 reproductive health diagnosis. RESULTS: The most frequent reproductive health diagnoses were menstrual disorders and endometriosis among those aged 18-44 years (n=16,658, 13%), menopausal disorders among those aged 45-64 years (n=20,707, 15%), and osteoporosis among those aged ≥65 years (n=8365, 22%). Compared with women without reproductive health diagnoses, those with such diagnoses were more likely to have concomitant mental health (46% vs. 37%, P<0.001) and medical conditions (75% vs. 63%, P<0.001). CONCLUSIONS: Women Veterans using VA health care have diverse reproductive health diagnoses. The high prevalence of comorbid medical and mental health conditions among women Veterans with reproductive health diagnoses highlights the importance of integrating reproductive health expertise into all areas of VA health care, including primary, mental health, and specialty care.

"We don't wanna birth it here": A qualitative study of Southern jail personnel approaches to pregnancy.

Each year, approximately 55,000 pregnant people are incarcerated in US jails. To learn about pregnancy and postpartum care in jails, we analyzed 34 qualitative interviews with jail personnel from facilities in five Southeastern US states. Themes included jail processes unique to pregnancy and burden on jails produced by liability and limited resources. Societal attitudes such as stigma, distrust of pregnant people, and a focus on fetal well-being were also important themes. Jail-community partnerships may mitigate the effects of scarce resources and improve jail perinatal care. Better community safety nets that decrease contact with jails are needed to improve pregnancy outcomes.

Words Matter: Examining Gender Differences in the Language Used to Evaluate Pediatrics Residents.

BACKGROUND: Gender disparities in academic medicine continue to be pervasive. Written evaluations of residents may provide insight into perceptions of residents by faculty, which may influence letters of recommendation for positions beyond residency and reinforce perceived stereotype threat experienced by trainees. OBJECTIVE: To examine language used in faculty evaluations of pediatrics residents to determine if there are differences in language used with respect to gender of resident. DESIGN/METHODS: All faculty evaluations of residents in 3 consecutive intern classes from 2016 to 2018 were collected and redacted for name and gender identifiers. We performed a qualitative analysis of written comments in 2 mandatory free text sections. The study team initially coded text collectively, generating a code book, then individually to apply the coding scheme. Next, evaluations were unblinded to gender. Code applications were aggregated by resident, and frequencies of code application by resident were compared by standardized mean differences to detect imbalances between genders. RESULTS: A total of 448 evaluations were analyzed: 88 evaluations of 17 male residents, and 360 evaluations of 70 female residents. Codes more frequently applied to women included "enthusiasm," and "caring," while codes more frequently applied to men included "intelligence," and "prepared." A conceptual model was created to reflect potential impacts of these differences using a lens of social role theory. CONCLUSIONS: We identified differences in the way male and female residents are evaluated by faculty, which may have negative downstream effects on female residents, who may experience negative self-perception, differential development of clinical skills, and divergent career opportunities as a result.

Are You Really the Doctor? Physician Experiences with Gendered Microaggressions from Patients.

Background: In contrast to physician implicit bias toward patients, bias and microaggressions from patients toward physicians have received comparatively less attention. Materials and Methods: We captured physician experiences of gendered microaggressions from patients by conducting a mixed-methods survey-based study of physicians at a single academic health care institution in May 2019. A quantitative portion assessed the frequency of gendered microaggressions (microaggression experiences [ME] score) and the association with measures of perceived impacts (job satisfaction, burnout, perceived career impacts, behavioral modifications). A one-tailed Wilcoxon rank sum test compared distributional frequencies of microaggressions by gender, and by gender and race. Chi-square tests measured the associations between gendered microaggressions and perceived impacts. Welch two-sample t-tests assessed differences in ME scores by rank and specialty. Linear regression assessed the association of ME scores and job satisfaction/burnout. A qualitative portion solicited anecdotal experiences, analyzed by inductive thematic analysis. Results: There were 297 completed surveys (response rate 27%). Female physicians experienced a significantly higher frequency of gendered microaggressions (p < 0.001) compared with male physicians. Microaggressions were significantly associated with job satisfaction (chi-square 6.83, p = 0.009), burnout (chi-square 8.76, p = 0.003), perceived career impacts (chi-square 18.67, p < 0.001), and behavioral modifications (chi-square 19.96, p < 0.001). Trainees experienced more microaggressions (p = 0.009) and burnout (p = 0.009) than faculty. Higher ME scores predicted statistically significant increases in burnout (p < 0.0001) and reduced job satisfaction (p = 0.02). Twelve microaggressions themes emerged from the qualitative responses, including role questioning and assumption of inexperience. The frequency of microaggressions did not vary significantly by race; however, qualitative responses described race as a factor. Conclusions: Physicians experience gendered microaggressions from patients, which may influence job satisfaction, burnout, career perceptions, and behavior. Future research may explore the multidirectionality of microaggressions and tools for responding at the individual and institutional level.

Trends in Utilization of Magnetic Resonance Imaging for Stroke Patients With Cardiac Rhythm Devices.

Background and Purpose: Magnetic resonance imaging (MRI) is safe for most patients with cardiovascular implantable electronic devices (CIEDs). However, patients presenting with acute ischemic stroke or transient ischemic attack (AIS/TIA) who have CIEDs may undergo MRI less frequently than patients without devices. We assessed contemporary use of MRI for patients with AIS/TIA and the effect of a recent coverage revision by the Center for Medicare and Medicaid Services (CMS) on MRI utilization. Methods: Using Optum® claims data from January 2012 to June 2019, we performed an interrupted time series analysis of MRI utilization during AIS/TIA hospitalizations with the April 2018 CMS coverage revision serving as the intervention. For patients treated after the coverage revision, we used multivariable logistic regression to determine the association between lack of CIED and MRI utilization for AIS/TIA. Results: We identified 417,899 patient hospitalizations for AIS/TIA, of which 30,425 (7%) had a CIED present (CIED vs non-CIED patients: age 77.6 ± 9.8 vs 72.7 ± 12.3 years; 45.5% vs 54.3% female). From 2012 to 2019, annual MRI utilization increased from 3% to 20% for CIED patients and 58% to 66% for non-CIED patients. The CMS coverage revision was associated with a 4.2% absolute additional increase in MRI utilization for CIED patients. Non-CIED patients treated after the CMS coverage revision were substantially more likely than CIED patients to undergo MRI (adjusted OR 6.7, 95% CI: 6.3-7.1, P<.001). Conclusions: MRI utilization has increased for stroke patients with CIEDs but remains far lower than in similar patients without devices.

Inflammatory but not respiratory symptoms are associated with ongoing upper airway viral shedding in outpatients with uncomplicated COVID-19.

Although the vast majority of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) infections are uncomplicated, our understanding of predictors of symptom resolution and viral shedding cessation remains limited. We characterized symptom trajectories and oropharyngeal viral shedding among 120 outpatients with uncomplicated Coronavirus Disease of 2019 (COVID-19) enrolled in a clinical trial of Peginterferon Lambda, which demonstrated no clinical or virologic benefit compared with placebo. In the combined trial cohort, objective fever was uncommon, inflammatory symptoms (myalgias, fatigue) peaked at 4 to 5 days postsymptom onset, and cough peaked at 9 days. The median time to symptom resolution from earliest symptom onset was 17 days (95% confidence interval 14-18). SARS-CoV-2 IgG seropositivity at enrollment was associated with hastened resolution of viral shedding (hazard ratio 1.80, 95% confidence interval 1.05-3.1, P = 0.03), but not with symptom resolution. Inflammatory symptoms were associated with a significantly greater odds of oropharyngeal SARS-CoV-2 RNA detection; respiratory symptoms were not. These findings have important implications for COVID-19 screening approaches and trial design.

The development of a mobile app-focused deduplication strategy for the Apple Heart Study that informs recommendations for future digital trials.

An app-based clinical trial enrolment process can contribute to duplicated records, carrying data management implications. Our objective was to identify duplicated records in real time in the Apple Heart Study (AHS). We leveraged personal identifiable information (PII) to develop a dissimilarity score (DS) using the Damerau-Levenshtein distance. For computational efficiency, we focused on four types of records at the highest risk of duplication. We used the receiver operating curve (ROC) and resampling methods to derive and validate a decision rule to classify duplicated records. We identified 16,398 (4%) duplicated participants, resulting in 419,297 unique participants out of a total of 438,435 possible. Our decision rule yielded a high positive predictive value (96%) with negligible impact on the trial's original findings. Our findings provide principled solutions for future digital trials. When establishing deduplication procedures for digital trials, we recommend collecting device identifiers in addition to participant identifiers; collecting and ensuring secure access to PII; conducting a pilot study to identify reasons for duplicated records; establishing an initial deduplication algorithm that can be refined; creating a data quality plan that informs refinement; and embedding the initial deduplication algorithm in the enrolment platform to ensure unique enrolment and linkage to previous records.

Arrhythmias Other Than Atrial Fibrillation in Those With an Irregular Pulse Detected With a Smartwatch: Findings From the Apple Heart Study.

[Figure: see text].

Peginterferon Lambda-1a for treatment of outpatients with uncomplicated COVID-19: a randomized placebo-controlled trial.

Type III interferons have been touted as promising therapeutics in outpatients with coronavirus disease 2019 (COVID-19). We conducted a randomized, single-blind, placebo-controlled trial (NCT04331899) in 120 outpatients with mild to moderate COVID-19 to determine whether a single, 180 mcg subcutaneous dose of Peginterferon Lambda-1a (Lambda) within 72 hours of diagnosis could shorten the duration of viral shedding (primary endpoint) or symptoms (secondary endpoint). In both the 60 patients receiving Lambda and 60 receiving placebo, the median time to cessation of viral shedding was 7 days (hazard ratio [HR] = 0.81; 95% confidence interval [CI] 0.56 to 1.19). Symptoms resolved in 8 and 9 days in Lambda and placebo, respectively, and symptom duration did not differ significantly between groups (HR 0.94; 95% CI 0.64 to 1.39). Both Lambda and placebo were well-tolerated, though liver transaminase elevations were more common in the Lambda vs. placebo arm (15/60 vs 5/60; p = 0.027). In this study, a single dose of subcutaneous Peginterferon Lambda-1a neither shortened the duration of SARS-CoV-2 viral shedding nor improved symptoms in outpatients with uncomplicated COVID-19.

Postpartum Depression Screening and Referrals in Special Supplemental Nutrition Program for Women, Infants, and Children Clinics.

OBJECTIVES: To examine the relationships among participants' demographic, social, and health characteristics and positive screening scores for symptoms of postpartum depression (PPD); to examine the feasibility of referring to a case management program women with symptoms of PPD who are accessing Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) services; and to identify barriers to screening and treatment programs for women with symptoms of PPD. DESIGN: Descriptive, cross-sectional study followed by a process evaluation. SETTING: Two WIC clinics in a large southeastern U.S. city. PARTICIPANTS: One group (n = 302) included women with infants younger than 12 months who were screened for symptoms of PPD. The second group (n = 31) included case managers (n = 7), nutritionists (n = 12), advisory board members (n = 7), and student volunteers (n = 5) who participated in focus groups. METHODS: We conducted an initial screening of women for symptoms of PPD using the two-item Patient Health Questionnaire (PHQ-2). Participants with scores of 2 or greater (n = 73) were asked to complete the nine-item PHQ (PHQ-9) and the Edinburgh Postnatal Depression Scale. Participants were referred for case management services if they scored 10 or greater on the PHQ-9 or Edinburgh Postnatal Depression Scale (n = 29) and agreed to the referral (n = 19). We transcribed and analyzed the qualitative data recorded during focus groups. RESULTS: Participants with no health insurance and limited support in caring for their infants were more likely to report symptoms of PPD. Overall, 302 women were screened for PPD, indicating the feasibility of PPD screening in WIC clinics. Of the 19 participants referred to case management, 47% (n = 9) accessed care. The results of focus groups illuminated barriers to screening and treatment programs for women at the individual, local, and macrosystem levels. CONCLUSION: Our findings show the feasibility of PPD screening in WIC clinics. However, some participants did not receive mental health services after referral because of various barriers, which highlights the need to integrate mental health providers into WIC clinics.

Predicting Nondiagnostic Home Sleep Apnea Tests Using Machine Learning.

STUDY OBJECTIVES: Home sleep apnea testing (HSAT) is an efficient and cost-effective method of diagnosing obstructive sleep apnea (OSA). However, nondiagnostic HSAT necessitates additional tests that erode these benefits, delaying diagnoses and increasing costs. Our objective was to optimize this diagnostic pathway by using predictive modeling to identify patients who should be referred directly to polysomnography (PSG) due to their high probability of nondiagnostic HSAT. METHODS: HSAT performed as the initial test for suspected OSA within the Veterans Administration Greater Los Angeles Healthcare System was analyzed retrospectively. Data were extracted from pre-HSAT questionnaires and the medical record. Tests were diagnostic if there was a respiratory event index (REI) ≥ 5 events/h. Tests with REI < 5 events/h or technical inadequacy-two outcomes requiring additional testing with PSG-were considered nondiagnostic. Standard logistic regression models were compared with models trained using machine learning techniques. RESULTS: Models were trained using 80% of available data and validated on the remaining 20%. Performance was evaluated using partial area under the precision-recall curve (pAUPRC). Machine learning techniques consistently yielded higher pAUPRC than standard logistic regression, which had pAUPRC of 0.574. The random forest model outperformed all other models (pAUPRC 0.862). Preferred calibration of this model yielded the following: sensitivity 0.46, specificity 0.95, positive predictive value 0.81, negative predictive value 0.80. CONCLUSIONS: Compared with standard logistic regression models, machine learning models improve prediction of patients requiring in-laboratory PSG. These models could be implemented into a clinical decision support tool to help clinicians select the optimal test to diagnose OSA.

Extended cardiac ambulatory rhythm monitoring in adults with congenital heart disease: Arrhythmia detection and impact of extended monitoring.

BACKGROUND: Arrhythmias are a leading cause of death in adults with congenital heart disease (ACHD). While 24-48-hour monitors are often used to assess arrhythmia burden, extended continuous ambulatory rhythm monitors (ECAM) can record 2 weeks of data. The utility of this device and the arrhythmia burden identified beyond 48-hour monitoring have not been evaluated in the ACHD population. Additionally, the impact of ECAM has not been studied to determine management recommendations. OBJECTIVE: To address the preliminary question, we hypothesized that clinically significant arrhythmias would be detected on ECAM beyond 48 hours and this would lead to clinical management changes. METHODS: A single center retrospective cohort study of ACHD patients undergoing ECAM from June 2013 to May 2016 was performed. The number and type of arrhythmias detected within and beyond the first 48 hours of monitoring were compared using Kaplan-Meier curves and Cox proportional hazard models. RESULTS: Three hundred fourteen patients had monitors performed [median age 31 (IQR 25-41) years, 61% female). Significant arrhythmias were identified in 156 patients (50%), of which 46% were noted within 48 hours. A management change based on an arrhythmia was made in 49 patients (16%). CONCLUSIONS: ECAM detects more clinically significant arrhythmias than standard 48-hour monitoring in ACHD patients. Management changes, including medication changes, further testing or imaging, and procedures, were made based on results of ECAM. Recommendations and guidelines have been made based on arrhythmias on 48-hour monitoring; the predictive ability and clinical consequence of arrhythmias found on ECAM are not yet known.

Predictors of ductal closure and intestinal complications in very low birth weight infants treated with indomethacin.

OBJECTIVE: To describe factors associated with failure of patent ductus arteriosus closure and development of gastrointestinal complications in subjects treated with indomethacin. STUDY DESIGN: Infants