Search

BACKGROUND: Screening for islet autoantibodies in children and adolescents identifies individuals who will later develop type 1 diabetes, allowing patient and family education to prevent diabetic ketoacidosis at onset and to enable consideration of preventive therapies. We aimed to assess whether islet autoantibody screening is effective for predicting type 1 diabetes in adolescents aged 10−18 years with an increased risk of developing type 1 diabetes. METHODS: Data were harmonised from prospective studies from Finland (the Diabetes Prediction and Prevention study), Germany (the BABYDIAB study), and the USA (Diabetes Autoimmunity Study in the Young and the Diabetes Evaluation in Washington study). Autoantibodies against insulin, glutamic acid decarboxylase, and insulinoma-associated protein 2 were measured at each follow-up visit. Children who were lost to follow-up or diagnosed with type 1 diabetes before 10 years of age were excluded. Inverse probability censoring weighting was used to include data from remaining participants. Sensitivity and the positive predictive value of these autoantibodies, tested at one or two ages, to predict type 1 diabetes by the age of 18 years were the main outcomes. FINDINGS: Of 20 303 children with an increased type 1 diabetes risk, 8682 were included for the analysis with inverse probability censoring weighting. 1890 were followed up to 18 years of age or developed type 1 diabetes between the ages of 10 years and 18 years, and their median follow-up was 18·3 years (IQR 14·5–20·3). 442 (23·4%) of 1890 adolescents were positive for at least one islet autoantibody, and 262 (13·9%) developed type 1 diabetes. Time from seroconversion to diabetes diagnosis increased by 0·64 years (95% CI 0·34–0·95) for each 1-year increment of diagnosis age (Pearson’s correlation coefficient 0·88, 95% CI 0·50–0·97, p=0·0020). The median interval between the last prediagnostic sample and diagnosis was 0·3 years (IQR 0·1–1·3) in the 227 participants who were autoantibody positive and 6·8 years (1·6–9·9) for the 35 who were autoantibody negative. Single screening at the age of 10 years was 90% (95% CI 86–95) sensitive, with a positive predictive value of 66% (60–72) for clinical diabetes. Screening at two ages (10 years and 14 years) increased sensitivity to 93% (95% CI 89–97) but lowered the positive predictive value to 55% (49–60). INTERPRETATION: Screening of adolescents at risk for type 1 diabetes only once at 10 years of age for islet autoantibodies was highly effective to detect type 1 diabetes by the age of 18 years, which in turn could enable prevention of diabetic ketoacidosis and participation in secondary prevention trials. FUNDING: JDRF International.

Recent advances in the science and technology of artificial intelligence (AI) and growing numbers of deployed AI systems in healthcare and other services have called attention to the need for ethical principles and governance. We define and provide a rationale for principles that should guide the commission, creation, implementation, maintenance, and retirement of AI systems as a foundation for governance throughout the lifecycle. Some principles are derived from the familiar requirements of practice and research in medicine and healthcare: beneficence, nonmaleficence, autonomy, and justice come first. A set of principles follow from the creation and engineering of AI systems: explainability of the technology in plain terms; interpretability, that is, plausible reasoning for decisions; fairness and absence of bias; dependability, including “safe failure”; provision of an audit trail for decisions; and active management of the knowledge base to remain up to date and sensitive to any changes in the environment. In organizational terms, the principles require benevolence—aiming to do good through the use of AI; transparency, ensuring that all assumptions and potential conflicts of interest are declared; and accountability, including active oversight of AI systems and management of any risks that may arise. Particular attention is drawn to the case of vulnerable populations, where extreme care must be exercised. Finally, the principles emphasize the need for user education at all levels of engagement with AI and for continuing research into AI and its biomedical and healthcare applications.

The Collaborative Open Outcomes tooL (COOL) is a novel, highly configurable application to simulate, evaluate and compare potential population-level screening schedules. Its first application is type 1 diabetes (T1D) screening, where known biomarkers for risk exist but clinical application lags behind. COOL was developed with the T1DI Study Group, in order to assess screening schedules for islet autoimmunity development based on existing datasets. This work shows clinical research utility, but the tool can be applied in other contexts. COOL helps the user define and evaluate a domain knowledge-driven screening schedule, which can be further refined with data-driven insights. COOL can also compare performance of alternative schedules using adjusted sensitivity, specificity, PPV and NPV metrics. Insights from COOL may support a variety of needs in disease screening and surveillance.

Context Rapid growth has been suggested to promote islet autoimmunity and progression to type 1 diabetes (T1D). Childhood growth has not been analyzed separately from the infant growth period in most previous studies, but it may have distinct features due to differences between the stages of development. Objective We aimed to analyze the association of childhood growth with development of islet autoimmunity and progression to T1D diagnosis in children 1 to 8 years of age. Methods Longitudinal data of childhood growth and development of islet autoimmunity and T1D were analyzed in a prospective cohort study including 10 145 children from Finland, Germany, Sweden, and the United States, 1-8 years of age with at least 3 height and weight measurements and at least 1 measurement of islet autoantibodies. The primary outcome was the appearance of islet autoimmunity and progression from islet autoimmunity to T1D. Results Rapid increase in height (cm/year) was associated with increased risk of seroconversion to glutamic acid decarboxylase autoantibody, insulin autoantibody, or insulinoma-like antigen-2 autoantibody (hazard ratio [HR] = 1.26 [95% CI = 1.05, 1.51] for 1-3 years of age and HR = 1.48 [95% CI = 1.28, 1.73] for >3 years of age). Furthermore, height rate was positively associated with development of T1D (HR = 1.80 [95% CI = 1.15, 2.81]) in the analyses from seroconversion with insulin autoantibody to diabetes. Conclusion Rapid height growth rate in childhood is associated with increased risk of islet autoimmunity and progression to T1D. Further work is needed to investigate the biological mechanism that may explain this association.

The potential value of AI to healthcare, and nursing in particular, ranges from improving quality and efficiency of care to delivering on the promise of personalized and precision medicine. AI systems may become virtually indispensable as ever more data is amassed about every aspect of health. AI can help reduce variability in care, while improving precision, accelerating discovery and reducing disparities. AI can empower patients and potentially allow healthcare professionals to relate to their patients as healers supported by the combined wisdom of the best medical research and analytic technology. There are, however, many challenges to understanding the optimal uses of AI; addressing the technological, systemic, regulatory and attitudinal roadblocks to successful implementation; and integrating AI into the fabric of health care. This paper provides a grounding in the origins and fundamental building blocks of AI, applications in healthcare and for nursing, and the critical challenges facing implementation in healthcare.

Four birth-cohort studies in the U.S. (DAISY, DEWIT), Sweden (DiPiS) and Finland (DIPP) were initiated in 1990s to early 2000. General population newborns and young first-degree relatives (FDR) of T1D patients, preselected on increased HLA risk, were followed for islet autoantibodies (AAb)and T1D. We harmonized their common variables, including AAb to insulin, IA-2, GAD, or ZnT8, HLA-DR-DQ risk genotypes, socio-demographics, diet, family history, glucose/HbA1c and anthropometric measures. The combined cohort included 22,312 subjects with AAb measured in 256,243 visits (mean 11±10 per subject) at 3-12 month intervals. During follow-up, 2279 (10%) subjects developed at least one AAb, 914 (4%) multiple AAb and 603 (3%) T1D. The cumulative risk of T1D differed by the number of AAb at seroconversion (Figure). T1D developed in 26% (95% CI 24-32%) of 1AAb, 62% (55-68%) of 2AAb and 84% (75-88%) of >=3AAb positive subjects in the 15y after seroconversion. T1D risk varied by HLA genotype: 40% (35-48%) in DR3/4-DQ2/8, 22% (18-31%) in DR4/x-DQ8/x and 21% (18-35%) in DR3/x-DQ2/x of 1AAb subjects and, respectively, in 82% (68-88%), 61%(50-65%)and 87% (32-98%)of 2AAb subjects. Once ≥2 AAb positive, HLA genotype was less predictive of progression to T1D. This harmonized dataset offers an unparalleled opportunity to validate and improve upon reported risk estimates for T1D development via novel analytical methods. Disclosure V. Anand: None. B. Liu: None. M. Ghalwash: None. E. Koski: None. K. Ng: None. J.L. Dunne: None. W. Hagopian: Research Support; Self; Novo Nordisk A/S. M. Lundgren: None. Å. Lernmark: None. J. Ilonen: None. R. Veijola: None. M. Rewers: None. Funding JDRF

The development of multiple islet autoantibodies (IA) predicts progression to type 1 diabetes (T1D), and forms the basis of current staging of T1D. However, age at seroconversion, persistence of IA, and frequency of sampling influence the determination of multiple IA status. Individual IA may appear at different timepoints and fluctuate. This analysis aims to understand how differences in the stringency of multiple IA status definition impact risk estimates of progression to stage 3 T1D. We combined data from three birth-cohort studies (N=18,537): DAISY (U.S.), DiPiS (Sweden) and DIPP (Finland), defining four groups by IA positivity pattern and accounting for variability in persistence and timing in concurrent and sequential patterns of IAA, IA2A, and GADA. Subjects were placed exclusively in their most stringent IA group. Cumulative incidence of progression to T1D was estimated by survival analysis and log-rank test, which showed group differences (p Disclosure B. Liu: None. M. Ghalwash: None. V. Anand: None. E. Koski: None. K. Ng: None. J.L. Dunne: None. M. Lundgren: None. H. Elding Larsson: None. W. Hagopian: Research Support; Self; Novo Nordisk A/S. R. Veijola: None. B.I. Frohnert: None. Funding JDRF

Clinical type 1 diabetes (T1D) is preceded by asymptomatic islet-autoimmunity (IA) with presence of islet autoantibodies (AAb). However, both AAb profiles (clusters) during IA phase and progression rate to T1D are heterogeneous. It is unclear if AAb profiles affect the progression to T1D similarly in different countries. We analyzed data from 3 prospective birth cohorts, DAISY (U.S.), DiPiS (Sweden), and DIPP (Finland), to address this question. We developed a similarity algorithm that considers subject-level temporal IA profiles (based on 4 AAbs - IAA, GADA, IA2A, and ZnT8A positivity or negativity), age at which AAb developed, and imbalance in AAb positivity by weighting positive AAb matches higher than negative ones. Using this subject-level similarity measure, we performed hierarchical clustering of AAb positive subjects having at least 3 visits. We used a log-rank test to identify the number of clusters in 1125 DIPP subjects (329 T1D cases), 309 DAISY subjects (70 cases), and 226 DiPiS subjects (33 cases). Each dataset was individually clustered. We identified three clusters in each dataset associated with different rate of progression from seroconversion to T1D - slow progressors (SP), moderate progressors (MP) and fast progressors (FP). These were distributed as follows: DIPP (654 SP with 3% 10-year T1D risk, 356 MP with 60% risk, 115 FP with 96% risk); DAISY (215 SP with 6% 10-year risk, 92 MP with 62% risk, 2 FP with 100% risk); and DiPiS (160 SP with 3% 10-year risk, 62 MP with 58% risk, 4 FP with 100% risk). Finally, we tested the generalizability of the clusters by 1) learning a clustering from one data set “learned clusters,” 2) classifying subjects from another (independent) data set with the “learned clusters,” and 3) measuring how often the subjects were assigned to the same cluster (SP, MP, FP) as the “internal” clustering. The average accuracy of cluster transfer was high (mean = 87.9%, sd = 1.1%) suggesting that the IA profile predicts progression rate to T1D independently of the population. Disclosure M. Ghalwash: None. V. Anand: None. B. Liu: None. E. Koski: None. K. Ng: None. J.L. Dunne: None. M. Lundgren: None. M. Rewers: None. R. Veijola: None. Funding JDRF

Importance The lack of standards in methods to reduce bias for clinical algorithms presents various challenges in providing reliable predictions and in addressing health disparities. Objective To evaluate approaches for reducing bias in machine learning models using a real-world clinical scenario. Design, setting, and participants Health data for this cohort study were obtained from the IBM MarketScan Medicaid Database. Eligibility criteria were as follows: (1) Female individuals aged 12 to 55 years with a live birth record identified by delivery-related codes from January 1, 2014, through December 31, 2018; (2) greater than 80% enrollment through pregnancy to 60 days post partum; and (3) evidence of coverage for depression screening and mental health services. Statistical analysis was performed in 2020. Exposures Binarized race (Black individuals and White individuals). Main outcomes and measures Machine learning models (logistic regression [LR], random forest, and extreme gradient boosting) were trained for 2 binary outcomes: postpartum depression (PPD) and postpartum mental health service utilization. Risk-adjusted generalized linear models were used for each outcome to assess potential disparity in the cohort associated with binarized race (Black or White). Methods for reducing bias, including reweighing, Prejudice Remover, and removing race from the models, were examined by analyzing changes in fairness metrics compared with the base models. Baseline characteristics of female individuals at the top-predicted risk decile were compared for systematic differences. Fairness metrics of disparate impact (DI, 1 indicates fairness) and equal opportunity difference (EOD, 0 indicates fairness). Results Among 573 634 female individuals initially examined for this study, 314 903 were White (54.9%), 217 899 were Black (38.0%), and the mean (SD) age was 26.1 (5.5) years. The risk-adjusted odds ratio comparing White participants with Black participants was 2.06 (95% CI, 2.02-2.10) for clinically recognized PPD and 1.37 (95% CI, 1.33-1.40) for postpartum mental health service utilization. Taking the LR model for PPD prediction as an example, reweighing reduced bias as measured by improved DI and EOD metrics from 0.31 and -0.19 to 0.79 and 0.02, respectively. Removing race from the models had inferior performance for reducing bias compared with the other methods (PPD: DI = 0.61; EOD = -0.05; mental health service utilization: DI = 0.63; EOD = -0.04). Conclusions and relevance Clinical prediction models trained on potentially biased data may produce unfair outcomes on the basis of the chosen metrics. This study's results suggest that the performance varied depending on the model, outcome label, and method for reducing bias. This approach toward evaluating algorithmic bias can be used as an example for the growing number of researchers who wish to examine and address bias in their data and models.

Randomized controlled trials can benefit from proactive assessment of how well their participant selection strategies during the design of eligibility criteria can influence the study generalizability. In this paper, we present a quantitative metric called generalizability index for study traits 2.0 (GIST 2.0) to assess the a priori generalizability (based on population representativeness) of a clinical trial by accounting for the dependencies among multiple eligibility criteria. The metric was evaluated on 16 sepsis trials identified from ClinicalTrials.gov, with their adverse event reports extracted from the trial results sections. The correlation between GIST scores and adverse events was analyzed. We found that the GIST 2.0 score was significantly correlated with total adverse events and serious adverse events (weighted correlation coefficients of 0.825 and 0.709, respectively, with P < 0.01). This study exemplifies the promising use of Big Data in electronic health records and ClinicalTrials.gov for optimizing eligibility criteria design for clinical studies.

Randomized controlled trials can benefit from proactive assessment of how well their participant selection strategies during the design of eligibility criteria can influence the study generalizability. In this paper, we present a quantitative metric called generalizability index for study traits 2.0 (GIST 2.0) to assess the a priori generalizability (based on population representativeness) of a clinical trial by accounting for the dependencies among multiple eligibility criteria. The metric was evaluated on 16 sepsis trials identified from ClinicalTrials.gov, with their adverse event reports extracted from the trial results section. The correlation between GIST scores and adverse events was analyzed. We found that the GIST 2.0 score was significantly correlated with total adverse events and serious adverse events (weighted correlation coefficients of 0.825 and 0.709, respectively, with P < 0.01). This study exemplifies the promising use of Big Data in electronic health records and ClinicalTrials.gov for optimizing eligibility criteria design for clinical studies.

Syndromic surveillance refers to methods relying on detection of individual and population health indicators that are discernible before confirmed diagnoses are made. In particular, prior to the laboratory confirmation of an infectious disease, ill persons may exhibit behavioral patterns, symptoms, signs, or laboratory findings that can be tracked through a variety of data sources. Syndromic surveillance systems are being developed locally, regionally, and nationally. The efforts have been largely directed at facilitating the early detection of a covert bioterrorist attack, but the technology may also be useful for general public health, clinical medicine, quality improvement, patient safety, and research. This report, authored by developers and methodologists involved in the design and deployment of the first wave of syndromic surveillance systems, is intended to serve as a guide for informaticians, public health managers, and practitioners who are currently planning deployment of such systems in their regions.

This chapter provides general background on the types of data used for biosurveillance, syndromic surveillance and traditional public health surveillance, with particular emphasis on the use of clinical laboratory data. Overviews of types of surveillance, the history of clinical laboratory testing and different types of laboratories are provided as background. A more detailed discussion of the roles and characteristics of clinical laboratory data with respect to biosurveillance includes comments on data standards and analytic requirements, with particular emphasis on understanding possible sources of variability or artifact during data analysis and interpretation.

APPARENT DISPARITY IN TESTING OF PATIENTS WITH CHRONIC KIDNEY DISEASE (CKD), PLUS DIABETES AND/OR HYPERTENSION AND IMPACT OF GENDER Eileen Koski, Xiaohua Huang, Marsha Green, Harvey Kaufman, Herman Hurwitz, Richard Furlanetto, and Joseph Vassalotti . Quest Diagnostics Incorporated: Lyndhurst NJ, West Norriton PA, Madison NJ, Philadelpia PA, Chantilly VA, National Kidney Foundation and Mount Sinai School of Medicine New York NY This study represents a joint effort of Quest Diagnostics and the National Kidney Foundation to examine the relationship between CKD and two major diagnostic risk factors for CKD, diabetes mellitus (DM) and hypertension (HTN), in US adults. This analysis is based on national data from the Quest Informatics Data Warehouse, specifically 3.4 million eGFR results in October 2006, and 2.8 million urinary albumin-creatinine ratio results (ACR) from November 2005 to October 2006. The analysis identified 629,761 patients with eGFR 70, possibly an artifact of older patients having already been diagnosed with CKD thus no longer screened with ACR. The study suggests a need for physician education to improve compliance with evidence-based guidelines in patients with CKD, and DM, HTN, or both, with added emphasis on women, particularly those with HTN.