Search

Abstract The 12-lead electrocardiogram (ECG) is an integral component to the diagnosis of a multitude of cardiovascular conditions. It is performed using a complex set of skin surface electrodes, limiting its use outside traditional clinical settings. We developed an artificial intelligence algorithm, trained over 600,000 clinically acquired ECGs, to explore whether fewer leads as input are sufficient to reconstruct a 12-lead ECG. Two limb leads (I and II) and one precordial lead (V3) were required to generate a reconstructed 12-lead ECG highly correlated with the original ECG. An automatic algorithm for detection of ECG features consistent with acute myocardial infarction (MI) performed similarly for original and reconstructed ECGs (AUC = 0.95). When interpreted by cardiologists, reconstructed ECGs achieved an accuracy of 81.4 ± 5.0% in identifying ECG features of ST-segment elevation MI, comparable with the original 12-lead ECGs (accuracy 84.6 ± 4.6%). These results will impact development efforts to innovate ECG acquisition methods with simplified tools in non-specialized settings.

Abstract The annual cost of hospital care services in the US has risen to over $1 trillion despite relatively worse health outcomes compared to similar nations. These trends accentuate a growing need for innovative care delivery models that reduce costs and improve outcomes. HaH—a program that provides patients acute-level hospital care at home—has made significant progress over the past two decades. Technological advancements in remote patient monitoring, wearable sensors, health information technology infrastructure, and multimodal health data processing have contributed to its rise across hospitals. More recently, the COVID-19 pandemic brought HaH into the mainstream, especially in the US, with reimbursement waivers that made the model financially acceptable for hospitals and payors. However, HaH continues to face serious challenges to gain widespread adoption. In this review, we evaluate the peer-reviewed evidence and discuss the promises, challenges, and what it would take to tap into the future potential of HaH.

Abstract The rapid advancements in artificial intelligence (AI) have led to the development of sophisticated large language models (LLMs) such as GPT-4 and Bard. The potential implementation of LLMs in healthcare settings has already garnered considerable attention because of their diverse applications that include facilitating clinical documentation, obtaining insurance pre-authorization, summarizing research papers, or working as a chatbot to answer questions for patients about their specific data and concerns. While offering transformative potential, LLMs warrant a very cautious approach since these models are trained differently from AI-based medical technologies that are regulated already, especially within the critical context of caring for patients. The newest version, GPT-4, that was released in March, 2023, brings the potentials of this technology to support multiple medical tasks; and risks from mishandling results it provides to varying reliability to a new level. Besides being an advanced LLM, it will be able to read texts on images and analyze the context of those images. The regulation of GPT-4 and generative AI in medicine and healthcare without damaging their exciting and transformative potential is a timely and critical challenge to ensure safety, maintain ethical standards, and protect patient privacy. We argue that regulatory oversight should assure medical professionals and patients can use LLMs without causing harm or compromising their data or privacy. This paper summarizes our practical recommendations for what we can expect from regulators to bring this vision to reality.

Abstract The ability to identify who does or does not experience the intended immune response following vaccination could be of great value in not only managing the global trajectory of COVID-19 but also helping guide future vaccine development. Vaccine reactogenicity can potentially lead to detectable physiologic changes, thus we postulated that we could detect an individual’s initial physiologic response to a vaccine by tracking changes relative to their pre-vaccine baseline using consumer wearable devices. We explored this possibility using a smartphone app-based research platform that enabled volunteers (39,701 individuals) to share their smartwatch data, as well as self-report, when appropriate, any symptoms, COVID-19 test results, and vaccination information. Of 7728 individuals who reported at least one vaccination dose, 7298 received an mRNA vaccine, and 5674 provided adequate data from the peri-vaccine period for analysis. We found that in most individuals, resting heart rate (RHR) increased with respect to their individual baseline after vaccination, peaked on day 2, and returned to normal by day 6. This increase in RHR was greater than one standard deviation above individuals’ normal daily pattern in 47% of participants after their second vaccine dose. Consistent with other reports of subjective reactogenicity following vaccination, we measured a significantly stronger effect after the second dose relative to the first, except those who previously tested positive to COVID-19, and a more pronounced increase for individuals who received the Moderna vaccine. Females, after the first dose only, and those aged

Abstract We developed a smartphone application, MyGeneRank, to conduct a prospective observational cohort study (NCT03277365) involving the automated generation, communication, and electronic capture of response to a polygenic risk score (PRS) for coronary artery disease (CAD). Adults with a smartphone and an existing 23andMe genetic profiling self-referred to the study. We evaluated self-reported actions taken in response to personal CAD PRS information, with special interest in the initiation of lipid-lowering therapy. 19% (721/3,800) of participants provided complete responses for baseline and follow-up use of lipid-lowering therapy. 20% (n = 19/95) of high CAD PRS vs 7.9% (n = 8/101) of low CAD PRS participants initiated lipid-lowering therapy at follow-up (p-value = 0.002). Both the initiation of statin and non-statin lipid-lowering therapy was associated with degree of CAD PRS: 15.2% (n = 14/92) vs 6.0% (n = 6/100) for statins (p-value = 0.018) and 6.8% (n = 8/118) vs 1.6% (n = 2/123) for non-statins (p-value = 0.022) in high vs low CAD PRS, respectively. High CAD PRS was also associated with earlier initiation of lipid lowering therapy (average age of 52 vs 65 years in high vs low CAD PRS respectively, p-value = 0.007). Overall, degree of CAD PRS was associated with use of any lipid-lowering therapy at follow-up: 42.4% (n = 56/132) vs 28.5% (n = 37/130) (p-value = 0.009). We find that digital communication of personal CAD PRS information is associated with increased and earlier lipid-lowering initiation in individuals of high CAD PRS. Loss to follow-up is the primary limitation of this study. Alternative communication routes, and long-term studies with EHR-based outcomes are needed to understand the generalizability and durability of this finding.

Abstract Individual smartwatch or fitness band sensor data in the setting of COVID-19 has shown promise to identify symptomatic and pre-symptomatic infection or the need for hospitalization, correlations between peripheral temperature and self-reported fever, and an association between changes in heart-rate-variability and infection. In our study, a total of 38,911 individuals (61% female, 15% over 65) have been enrolled between March 25, 2020 and April 3, 2021, with 1118 reported testing positive and 7032 negative for COVID-19 by nasopharyngeal PCR swab test. We propose an explainable gradient boosting prediction model based on decision trees for the detection of COVID-19 infection that can adapt to the absence of self-reported symptoms and to the available sensor data, and that can explain the importance of each feature and the post-test-behavior for the individuals. We tested it in a cohort of symptomatic individuals who exhibited an AUC of 0.83 [0.81–0.85], or AUC = 0.78 [0.75–0.80] when considering only data before the test date, outperforming state-of-the-art algorithm in these conditions. The analysis of all individuals (including asymptomatic and pre-symptomatic) when self-reported symptoms were excluded provided an AUC of 0.78 [0.76–0.79], or AUC of 0.70 [0.69–0.72] when considering only data before the test date. Extending the use of predictive algorithms for detection of COVID-19 infection based only on passively monitored data from any device, we showed that it is possible to scale up this platform and apply the algorithm in other settings where self-reported symptoms can not be collected.

Background: Screening for asymptomatic, undiagnosed atrial fibrillation (AF) has the potential to allow earlier treatment, possibly resulting in prevention of strokes, but also to increase medical resource utilization. Objective: To compare healthcare utilization rates during the year following initiation of screening among participants screened for AF by electrocardiogram (ECG) sensor patch compared with a matched observational control group. Methods: A total of 1718 participants recruited from a health care plan based on age and comorbidities who were screened with an ECG patch (actively monitored group) as part of a prospective, pragmatic research trial were matched by age, sex, and CHA2DS2-VASc score with 3371 members from the same health plan (observational control group). Healthcare utilization, including visits, prescriptions, procedures, and diagnoses, during the 1 year following screening was compared between the groups using health plan claims data. Results: Overall, the actively monitored group had significantly higher rates of cardiology visits (adjusted incidence rate ratio [aIRR] [95% confidence interval (CI)]: 1.43 [1.27, 1.60]), no difference in primary care provider visits (aIRR [95% CI]: 1.0 [0.95, 1.05]), but lower rates of emergency department (ED) visits and hospitalizations (aIRR [95% CI]: 0.80 [0.69, 0.92]) compared with controls. Among those with newly diagnosed AF, the reduction in ED visits and hospitalizations was even greater (aIRR [95% CI]: 0.27 [0.17, 0.43]). Conclusion: AF screening in an asymptomatic, moderate-risk population with an ECG patch was associated with an increase in cardiology outpatient visits but also significantly lower rates of ED visits and hospitalizations over the 1 year following screening.

Aim Healthcare workers (HCWs) were among the first group of people vaccinated with the Pfizer-BioNTech Covid-19 vaccine (BNT162b2). Characterization of the kinetics of antibody response to vaccination is important to devise future vaccination strategies. To better characterize the antibody response to BNT162b2, we analyzed the kinetics of IgG and IgM antibody response to 5 different SARS-CoV-2 epitopes over a period of 6 months. Methods and results An observational single-centered study was conducted to evaluate the temporal dynamics of anti-SARS-CoV-2 antibodies following immunization with two doses of BNT162b2. Anti-SARS-CoV-2 antibodies were assessed using the Maverick SARS-CoV-2 multi-antigen panel (Genalyte Inc.). Healthcare workers aged ≥18 receiving BNT162b2 vaccination who self-reported no prior symptoms of COVID-19 nor prior COVID-19 PCR test positivity, were included in this study. HCWs developed an IgG antibody response to SARS-CoV-2 Spike S1, Spike S1 receptor binding domain (RBD), Spike S1S2 and Spike S2 after vaccination. IgG response was observed at two weeks following immunization in most participant samples and continued to increase at week 4, but subsequently decreased significantly starting at 3 months and up to 6 months. In contrast, IgM response to respective epitopes was minimal. Conclusion Multiplex results demonstrate that, contrary to natural infection, immunization with BNT162b2 produces minimal anti-Spike IgM response. Polyclonal IgG response to Spike declined at 3 months and continued to do so up to 6 months.

Abstract Background Whole-exome sequencing (WES) has become an efficient diagnostic test for patients with likely monogenic conditions such as rare idiopathic diseases or sudden unexplained death. Yet, many cases remain undiagnosed. Here, we report the added diagnostic yield achieved for 101 WES cases re-analyzed 1 to 7 years after initial analysis. Methods Of the 101 WES cases, 51 were rare idiopathic disease cases and 50 were postmortem “molecular autopsy” cases of early sudden unexplained death. Variants considered for reporting were prioritized and classified into three groups: (1) diagnostic variants, pathogenic and likely pathogenic variants in genes known to cause the phenotype of interest; (2) possibly diagnostic variants, possibly pathogenic variants in genes known to cause the phenotype of interest or pathogenic variants in genes possibly causing the phenotype of interest; and (3) variants of uncertain diagnostic significance, potentially deleterious variants in genes possibly causing the phenotype of interest. Results Initial analysis revealed diagnostic variants in 13 rare disease cases (25.4%) and 5 sudden death cases (10%). Re-analysis resulted in the identification of additional diagnostic variants in 3 rare disease cases (5.9%) and 1 sudden unexplained death case (2%), which increased our molecular diagnostic yield to 31.4% and 12%, respectively. Conclusions The basis of new findings ranged from improvement in variant classification tools, updated genetic databases, and updated clinical phenotypes. Our findings highlight the potential for re-analysis to reveal diagnostic variants in cases that remain undiagnosed after initial WES.

Abstract Background Diagnosis of monogenic as well as atypical forms of diabetes mellitus has important clinical implications for their specific diagnosis, prognosis, and targeted treatment. Single gene mutations that affect beta-cell function represent 1–2% of all cases of diabetes. However, phenotypic heterogeneity and lack of family history of diabetes can limit the diagnosis of monogenic forms of diabetes. Next-generation sequencing technologies provide an excellent opportunity to screen large numbers of individuals with a diagnosis of diabetes for mutations in disease-associated genes. Methods We utilized a targeted sequencing approach using the Illumina HiSeq to perform a case-control sequencing study of 22 monogenic diabetes genes in 4016 individuals with type 2 diabetes (including 1346 individuals diagnosed before the age of 40 years) and 2872 controls. We analyzed protein-coding variants identified from the sequence data and compared the frequencies of pathogenic variants (protein-truncating variants and missense variants) between the cases and controls. Results A total of 40 individuals with diabetes (1.8% of early onset sub-group and 0.6% of adult onset sub-group) were carriers of known pathogenic missense variants in the GCK, HNF1A, HNF4A, ABCC8, and INS genes. In addition, heterozygous protein truncating mutations were detected in the GCK, HNF1A, and HNF1B genes in seven individuals with diabetes. Rare missense mutations in the GCK gene were significantly over-represented in individuals with diabetes (0.5% carrier frequency) compared to controls (0.035%). One individual with early onset diabetes was homozygous for a rare pathogenic missense variant in the WFS1 gene but did not have the additional phenotypes associated with Wolfram syndrome. Conclusion Targeted sequencing of genes linked with monogenic diabetes can identify disease-relevant mutations in individuals diagnosed with type 2 diabetes not suspected of having monogenic forms of the disease. Our data suggests that GCK-MODY frequently masquerades as classical type 2 diabetes. The results confirm that MODY is under-diagnosed, particularly in individuals presenting with early onset diabetes and clinically labeled as type 2 diabetes; thus, sequencing of all monogenic diabetes genes should be routinely considered in such individuals. Genetic information can provide a specific diagnosis, inform disease prognosis and may help to better stratify treatment plans.

Abstract Coronary artery disease (CAD) is the leading cause of death, and genetic factors contribute significantly to risk of CAD. This study aims to identify new CAD genetic loci through a large-scale linkage analysis of 24 large and multigenerational families with 433 family members (GeneQuest II). All family members were genotyped with markers spaced by every 10 cM and a model-free nonparametric linkage (NPL-all) analysis was carried out. Two highly significant CAD loci were identified on chromosome 17q21.2 (NPL score of 6.20) and 7p22.2 (NPL score of 5.19). We also identified four loci with significant NPL scores between 4.09 and 4.99 on 2q33.3, 3q29, 5q13.2 and 9q22.33. Similar analyses in individual families confirmed the six significant CAD loci and identified seven new highly significant linkages on 9p24.2, 9q34.2, 12q13.13, 15q26.1, 17q22, 20p12.3, and 22q12.1, and two significant loci on 2q11.2 and 11q14.1. Two loci on 3q29 and 9q22.33 were also successfully replicated in our previous linkage analysis of 428 nuclear families. Moreover, two published risk variants, SNP rs46522 in UBE2Z and SNP rs6725887 in WDR12 by GWAS, were found within the 17q21.2 and 2q33.3 loci. These studies lay a foundation for future identification of causative variants and genes for CAD.

Objectives: The advent of large databases, wearable technology, and novel communications methods has the potential to expand the pool of candidate research participants and offer them the flexibility and convenience of participating in remote research. However, reports of their effectiveness are sparse. We assessed the use of various forms of outreach within a nationwide randomized clinical trial being conducted entirely by remote means. Methods: Candidate participants at possibly higher risk for atrial fibrillation were identified by means of a large insurance claims database and invited to participate in the study by their insurance provider. Enrolled participants were randomly assigned to one of two groups testing a wearable sensor device for detection of the arrhythmia. Results: Over 10 months, the various outreach methods used resulted in enrollment of 2659 participants meeting eligibility criteria. Starting with a baseline enrollment rate of 0.8% in response to an email invitation, the recruitment campaign was iteratively optimized to ultimately include website changes and the use of a five-step outreach process (three short, personalized emails and two direct mailers) that highlighted the appeal of new technology used in the study, resulting in an enrollment rate of 9.4%. Messaging that highlighted access to new technology outperformed both appeals to altruism and appeals that highlighted accessing personal health information. Conclusions: Targeted outreach, enrollment, and management of large remote clinical trials is feasible and can be improved with an iterative approach, although more work is needed to learn how to best recruit and retain potential research participants. Trial registration: Clinicaltrials.gov NCT02506244. Registered 23 July 2015. Keywords: Clinical trials, Clinical research, Digital technology, Remote enrollment, Remote monitoring, Outreach

The Scripps molecular autopsy study seeks to incorporate genetic testing into the postmortem examination of cases of sudden death in the young (

Background: Digital medicine and smartphone-enabled health technologies provide a novel source of human health and human biology data. However, in part due to its intricacies, few methods have been established to analyze and interpret data in this domain. We previously conducted a six-month interventional trial examining the efficacy of a comprehensive smartphone-based health monitoring program for individuals with chronic disease. This included 38 individuals with hypertension who recorded 6,290 blood pressure readings over the trial. Methods: In the present study, we provide a hypothesis testing framework for unstructured time series data, typical of patient-generated mobile device data. We used a mixed model approach for unequally spaced repeated measures using autoregressive and generalized autoregressive models, and applied this to the blood pressure data generated in this trial. Results: We were able to detect, roughly, a 2 mmHg decrease in both systolic and diastolic blood pressure over the course of the trial despite considerable intra- and inter-individual variation. Furthermore, by supplementing this finding by using a sequential analysis approach, we observed this result over three months prior to the official study end—highlighting the effectiveness of leveraging the digital nature of this data source to form timely conclusions. Conclusions: Health data generated through the use of smartphones and other mobile devices allow individuals the opportunity to make informed health decisions, and provide researchers the opportunity to address innovative health and biology questions. The hypothesis testing framework we present can be applied in future studies utilizing digital medicine technology or implemented in the technology itself to support the quantified self.

Plasma HDL cholesterol levels (HDL-C) are an independent predictor of coronary artery disease (CAD). We have completed a genome-wide linkage scan for HDL-C in a US cohort consisting of 388 multiplex families with premature CAD (GeneQuest). The heritability of HDL-C in GeneQuest was 0.37 with gender and age as covariates (P = 5.1 × 10−4). Two major quantitative trait loci (QTL) for log-transformed HDL-C adjusted for age and gender were identified onto chromosomes 7p22 and 15q25 with maximum multipoint logarithm of odds (LOD) scores of 3.76 and 6.69, respectively. Fine mapping decreased the 7p22 LOD score to a nonsignificant level of 3.09 and split the 15q25 QTL into two loci, one minor QTL on 15q22 (LOD = 2.73) that spanned the LIPC gene, and the other at 15q25 (LOD = 5.63). A family-based quantitative transmission disequilibrium test (QTDT) revealed significant association between variant rs1800588 in LIPC and HDL-C in the GeneQuest population (P = 0.0067), which may account for the minor QTL on 15q22. The 15q25 QTL is the most significant locus identified for HDL-C to date, and these results provide a framework for the ultimate identification of the underlying HDL-C variant and gene on chromosomes 15q25, which will provide insights into novel regulatory mechanisms of HDL-C metabolism.

An increased plasma triglyceride (TG) level is associated with coronary artery disease (CAD) and myocardial infarction (MI) and is a key characteristic of the metabolic syndrome. Here, we used a genome-wide linkage scan to identify a novel genetic locus that influences the plasma TG level. We genotyped 714 persons in 388 multiplex Caucasian families with premature CAD and MI with 408 polymorphic microsatellite markers that cover the entire human genome. The genome-wide scan identified positive linkage for the quantitative TG trait to a novel locus on chromosome 1p31-32 [peak single-point logarithm of odds (LOD) = 3.57, peak multipoint LOD = 3.12]. For single-point linkage analysis, two markers, D1S1728 and D1S551, showed LOD scores of 2.42 and 3.57, respectively. For multipoint linkage analysis, three markers, D1S3736, D1S1728, and D1S551, showed LOD scores of 2.43, 3.03, and 3.12, respectively. No other chromosomal regions showed a LOD score of >2.2. This study identifies a new genetic locus for TG on chromosome 1p31-32. Future studies of the candidate genes at this locus will identify a specific gene influencing the TG, which will provide insights into novel regulatory mechanisms of TG metabolism and may be important for the development of therapies to prevent CAD.

Background. Mobile health and digital medicine technologies are becoming increasingly used by individuals with common, chronic diseases to monitor their health. Numerous devices, sensors, and apps are available to patients and consumers–some of which have been shown to lead to improved health management and health outcomes. However, no randomized controlled trials have been conducted which examine health care costs, and most have failed to provide study participants with a truly comprehensive monitoring system. Methods. We conducted a prospective randomized controlled trial of adults who had submitted a 2012 health insurance claim associated with hypertension, diabetes, and/or cardiac arrhythmia. The intervention involved receipt of one or more mobile devices that corresponded to their condition(s) (hypertension: Withings Blood Pressure Monitor; diabetes: Sanofi iBGStar Blood Glucose Meter; arrhythmia: AliveCor Mobile ECG) and an iPhone with linked tracking applications for a period of 6 months; the control group received a standard disease management program. Moreover, intervention study participants received access to an online health management system which provided participants detailed device tracking information over the course of the study. This was a monitoring system designed by leveraging collaborations with device manufacturers, a connected health leader, health care provider, and employee wellness program–making it both unique and inclusive. We hypothesized that health resource utilization with respect to health insurance claims may be influenced by the monitoring intervention. We also examined health-self management. Results & Conclusions. There was little evidence of differences in health care costs or utilization as a result of the intervention. Furthermore, we found evidence that the control and intervention groups were equivalent with respect to most health care utilization outcomes. This result suggests there are not large short-term increases or decreases in health care costs or utilization associated with monitoring chronic health conditions using mobile health or digital medicine technologies. Among secondary outcomes there was some evidence of improvement in health self-management which was characterized by a decrease in the propensity to view health status as due to chance factors in the intervention group.

There have been a number of recent successes in the use of whole genome sequencing and sophisticated bioinformatics techniques to identify pathogenic DNA sequence variants responsible for individual idiopathic congenital conditions. However, the success of this identification process is heavily influenced by the ancestry or genetic background of a patient with an idiopathic condition. This is so because potential pathogenic variants in a patient’s genome must be contrasted with variants in a reference set of genomes made up of other individuals’ genomes of the same ancestry as the patient. We explored the effect of ignoring the ancestries of both an individual patient and the individuals used to construct reference genomes. We pursued this exploration in two major steps. We first considered variation in the per-genome number and rates likely functional derived (i.e., non-ancestral, based on the chimp genome) single nucleotide variants and small indels in 52 individual whole human genomes sampled from 10 different global populations. We took advantage of a suite of computational and bioinformatics techniques to predict the functional effect of over 24 million genomic variants, both coding and non-coding, across these genomes. We found that the typical human genome harbors ~5.5-6.1 million total derived variants, of which ~12,000 are likely to have a functional effect (~5000 coding and ~7000 non-coding). We also found that the rates of functional genotypes per the total number of genotypes in individual whole genomes differ dramatically between human populations. We then created tables showing how the use of comparator or reference genome panels comprised of genomes from individuals that do not have the same ancestral background as a patient can negatively impact pathogenic variant identification. Our results have important implications for clinical sequencing initiatives.