Associations between obstructive sleep apnea and sleep characteristics with chronic kidney disease in rural Pennsylvania.

STUDY OBJECTIVES: To examine the association between moderate-severe obstructive sleep apnea (msOSA) and sleep characteristics with chronic kidney disease (CKD) in a population of rural and urban adults in Pennsylvania. METHODS: A cross-sectional study of 23,643 adults who underwent polysomnography (PSG) at a rural healthcare system in Pennsylvania between 2009 and 2019. Serum creatinine was abstracted from electronic health records to calculate estimated glomerular filtration rate (eGFR). CKD was defined as an eGFR <60 mL/min/1.73 m2. msOSA was defined as an apnea-hypoxia index (AHI) ≥15 events/hour. Poisson regression was performed to estimate the prevalence ratio (PR) of CKD for various sleep measures while adjusting for age, sex, race, smoking (never, former, current), body mass index, diabetes, and hypertension at time of PSG. RESULTS: In this clinically-referred sample comprised of over one-third (35 %) rural individuals, the prevalence of CKD and msOSA was 9.4 % and 32.1 %, respectively. Patients with CKD had more severe OSA based on AHI and intermittent hypoxia profile and presented worse sleep quality across all studied measures. Having OSA was associated with a 13 % higher prevalence of CKD (95%CI: 1.04, 1.22). In addition, for every 5 % increment in sleep efficiency, the prevalence of CKD was 3 % lower (PR = 0.97, 95%CI: 0.96, 0.98). Significant associations that were in the expected direction were observed across most sleep characteristics in adjusted models. CONCLUSIONS: Moderate-severe OSA, nocturnal hypoxemia, and disruptions to normal sleep duration, continuity, and architecture are associated with increased CKD prevalence in Pennsylvania adults. Management of OSA and/or sleep disturbances may be an opportunity to improve CKD outcomes. The unique health disparities among vulnerable rural populations are deserving of future study.

IMPACT-FH Study for Implementing Innovative Family Communication and Cascade Testing Strategies for Familial Hypercholesterolemia.

Background: Relatives of probands diagnosed with familial hypercholesterolemia (FH) should undergo cascade testing for FH. Objectives: The purpose of this study was to evaluate probands' choices of innovative strategies to communicate their FH result with relatives and facilitate cascade testing uptake. Methods: Probands with an FH genetic result from the MyCode Community Health Initiative could choose to share their FH result with adult blood relatives via the Family and Healthcare Professional Packet (packet), family sharing and cascade chatbots (chatbot), and/or FH Outreach and Support Program (direct contact). Cascade testing uptake was measured as reported completion of genetic or cholesterol testing. Generalized estimating equations models were used to identify factors associated with testing. Results: One hundred seventy five probands received an FH result, median age was 58.9 (IQR: 44.9-69.3), and 58.9% were female. Probands shared information about 1,915 adult and 163 minor relatives (11.9 relatives per proband). Seventy percent of probands (121/175) selected at least one strategy for at least one adult relative. An average of 1.2 strategies was selected per adult relative. Cascade testing was completed for 26.6% (144/541) of adults with at least one strategy selected, 2.4% (33/1,374) of adults without a strategy selected, and 25.2% (41/163) of minor relatives. Factors associated with increased cascade testing uptake were selection of at least one strategy (6.32 higher odds), specifically, selection of direct contact (16.78 higher odds). Conclusions: Strategies implemented improved FH cascade testing uptake compared to previous estimates and in families where no strategy was selected. Overall uptake remains insufficient, which can be attributed to probands reluctance to select a strategy for many relatives.

"Faith and a sunny day": Association of patient frailty with strain experienced by informal caregivers of older adults with non-muscle-invasive bladder cancer.

INTRODUCTION: Few studies have evaluated the potential effects of aging-related conditions like frailty in older adults with cancer on informal caregivers. Our objective was to evaluate the association between the sum total of the aging-related conditions of older adults with non-muscle-invasive bladder cancer (NMIBC) and the strain reported by their informal caregivers. MATERIALS AND METHODS: We conducted an explanatory sequential mixed methods cross-sectional survey study that recruited 81 dyads of older adults with NMIBC (age ≥ 65 at diagnosis) and their informal caregivers. Our outcome was measured by the Caregiver Strain Index (CSI), a self-reported measure of informal caregivers. Our exposure was the patient's deficit accumulation index (DAI), a validated composite measure of frailty derived from a geriatric assessment. A multivariable negative binomial regression was conducted to model CSI. We conducted qualitative thematic content analysis of responses to open-ended survey questions to understand specific types of caregiver strain and to identify coping strategies. RESULTS: Mean ages of patients and caregivers were 79.4 years and 72.5 years, respectively. Most caregivers were spouses (75.3 %) and lived with the patient (80.2 %). Of patients, 54.3 % were robust, 29.6 % were pre-frail, and 16.1 % were frail. In the multivariable model, we found that patient DAI was significantly associated with CSI (adjusted incidence rate ratio 1.05, 95 % CI 1.02-1.09). The top three sources of strain identified by caregivers were emotional adjustments, medical management, and family adjustments. Coping strategies for each included self-management of emotions, self-education about bladder cancer, and social support, respectively. DISCUSSION: In this cross-sectional study, we found that worsening frailty in an older adult with NMIBC was associated with greater informal caregiver strain. Informal caregivers reported challenges with emotional management, family dynamics, and medical tasks. These findings may inform longitudinal research and interventions to support informal caregivers who provide care for older adults with NMIBC.

Innovative Implementation Strategies for Familial Hypercholesterolemia Cascade Testing: The Impact of Genetic Counseling.

The IMPACT-FH study implemented strategies (packet, chatbot, direct contact) to promote family member cascade testing for familial hypercholesterolemia (FH). We evaluated the impact of genetic counseling (GC) on medical outcomes, strategy selection, and cascade testing. Probands (i.e., patients with FH) were recommended to complete GC and select sharing strategies. Comparisons were performed for both medical outcomes and strategy selection between probands with or without GC. GEE models for Poisson regression were used to examine the relationship between proband GC completion and first-degree relative (FDR) cascade testing. Overall, 46.3% (81/175) of probands completed GC. Probands with GC had a median LDL-C reduction of -13.0 mg/dL (-61.0, 4.0) versus -1.0 mg/dL (-16.0, 17.0) in probands without GC (p = 0.0054). Probands with and without GC selected sharing strategies for 65.3% and 40.3% of FDRs, respectively (p < 0.0001). Similarly, 27.1% of FDRs of probands with GC completed cascade testing, while 12.0% of FDRs of probands without GC completed testing (p = 0.0043). Direct contact was selected for 47 relatives in total and completed for 39, leading to the detection of 18 relatives with FH. Proband GC was associated with improved medical outcomes and increased FDR cascade testing. Direct contact effectively identified FH cases for the subset who participated.

Predicting mortality within 1 year of ART initiation in children and adolescents living with HIV in sub-Saharan Africa: a retrospective observational cohort study.

BACKGROUND: Differentiated service delivery (DSD) for children and adolescents living with HIV can improve targeted resource use. We derived a mortality prediction score to guide clinical decision making for children and adolescents living with HIV. METHODS: Data for this retrospective observational cohort study were evaluated for all children and adolescents living with HIV and initiating antiretroviral therapy (ART); aged 0-19 years; and enrolled at Baylor clinics in Eswatini, Malawi, Lesotho, Tanzania, and Uganda between 2005 and 2020. Data for clinical prediction, including anthropometric values, physical examination, ART, WHO stage, and laboratory tests were captured at ART initiation. Backward stepwise variable selection and logistic regression were performed to develop predictive models for mortality within 1 year of ART initiation. Probabilities of mortality were generated, compared with true outcomes, internally validated, and evaluated against WHO advanced HIV criteria. FINDINGS: The study population included 16 958 children and adolescents living with HIV and initiated on ART between May 18, 2005, and Dec 18, 2020. Predictive variables for the most accurate model included: age, CD4 percentage, white blood cell count, haemoglobin concentration, platelet count, and BMI Z score as continuous variables, and WHO clinical stage and oedema, abnormal muscle tone and respiratory distress on examination as categorical variables. The area under the curve (AUC) of the predictive model was 0·851 (95% CI 0·839-0·863) in the training set and 0·822 (0·800-0·845) in the test set, compared with 0·606 (0·595-0·617) for the WHO advanced HIV criteria (p<0·0001). INTERPRETATION: This study evaluated a large, multinational population to derive a mortality prediction tool for children and adolescents living with HIV. The model more accurately predicted clinical outcomes than the WHO advanced HIV criteria and has the potential to improve DSD for children and adolescents living with HIV in high-burden settings. FUNDING: National Institute of Health Fogarty International Center.

Performance of a stool-based quantitative PCR assay for the diagnosis of tuberculosis in adolescents and adults: a multinational, prospective diagnostic accuracy study.

BACKGROUND: Despite increasing availability of rapid molecular tests for the diagnosis of tuberculosis in high-burden settings, many people with tuberculosis are undiagnosed. Reliance on sputum as the primary specimen for tuberculosis diagnostics contributes to this diagnostic gap. We evaluated the diagnostic accuracy and additive yield of a novel stool quantitative PCR (qPCR) assay for the diagnosis of tuberculosis in three countries in Africa with high tuberculosis burdens. METHODS: We undertook a prospective diagnostic accuracy study in Eswatini, Mozambique, and Tanzania from Sept 21, 2020, to Feb 2, 2023, to compare the diagnostic accuracy for tuberculosis of a novel stool qPCR test with the current diagnostic standard for Mycobacterium tuberculosis DNA detection from sputum and stool, Xpert-MTB/RIF Ultra (Xpert Ultra). Sputum, stool, and urine samples were provided by a cohort of participants, aged 10 years or older, diagnosed with tuberculosis. Participants with tuberculosis (cases) were enrolled within 72 h of treatment initiation for tuberculosis diagnosed clinically or following laboratory confirmation. Participants without tuberculosis (controls) consisted of household contacts of the cases who did not develop tuberculosis during a 6-month follow-up. The performance was compared with a robust composite microbiological reference standard (CMRS). FINDINGS: The cohort of adolescents and adults (n=408) included 268 participants with confirmed or clinical tuberculosis (cases), 147 (55%) of whom were living with HIV, and 140 participants (controls) without tuberculosis. The sensitivity of the novel stool qPCR was 93·7% (95% CI 87·4-97·4) compared with participants with detectable growth on M tuberculosis culture, and 88·1% (81·3-93·0) compared with sputum Xpert Ultra. The stool qPCR had an equivalent sensitivity as sputum Xpert Ultra (94·8%, 89·1-98·1) compared with culture. Compared with the CMRS, the sensitivity of the stool qPCR was higher than the current standard for tuberculosis diagnostics on stool, Xpert Ultra (80·4%, 73·4-86·2 vs 73·5%, 66·0-80·1; p=0·025 on paired comparison). The qPCR also identified 17-21% additional tuberculosis cases compared to sputum Xpert Ultra or sputum culture. In controls without tuberculosis, the specificity of the stool qPCR was 96·9% (92·2-99·1). INTERPRETATION: In this study, a novel qPCR for the diagnosis of tuberculosis from stool specimens had a higher accuracy in adolescents and adults than the current diagnostic PCR gold standard on stool, Xpert-MTB/RIF Ultra, and equivalent sensitivity to Xpert-MTB/RIF Ultra on sputum. FUNDING: National Institutes of Health (NIH) Allergy and Infectious Diseases, and NIH Fogarty International Center.

Using implementation science to develop a familial hypercholesterolemia screening program in primary care: The CARE-FH study.

BACKGROUND: We designed the Collaborative Approach to Reach Everyone with Familial Hypercholesterolemia (CARE-FH) clinical trial to improve FH screening in primary care and facilitate guideline-based care. OBJECTIVE: The goal was to incorporate perspectives from end-users (healthcare system leaders, primary care clinicians, cardiologists, genetic counselors, nurses, and clinic staff) and improve translation of screening guidance into practice. METHODS: We partnered with end-users to sequentially define the current state of FH screening, assess acceptability, feasibility, and appropriateness of implementing an FH screening program, and select clinically actionable strategies at the patient-, clinician-, and system-level to be deployed as a package in the CARE-FH clinical trial. Methods informed by implementation science and human centered design included: contextual inquiries, surveys, and deliberative engagement sessions. RESULTS: Screening for FH occurred rarely in primary care, and then only after a cardiovascular event or sometimes due to a family history of high cholesterol or early heart attack. Surveys suggested FH screening in primary care was acceptable, appropriate, and feasible. Reported and observed barriers to screening include insufficient time at patient encounters to screen, cost and convenience of testing for patients, and knowledge regarding causes of dyslipidemia. Facilitators included clear guidance on screening criteria and new therapies to treat FH. These results led to the development of multilevel strategies that were presented to end-users, modified, and then pilot tested in one primary care clinic. CONCLUSIONS: A refined implementation strategy package for FH screening was created with a goal of improving FH awareness, identification, and initiation of guideline-based care. CLINICAL TRIAL REGISTRATION: https://clinicaltrials.gov/study/NCT05284513?id=NCT05284513&rank=1 Unique Identifier: NCT05284513.

Development and multi-site external validation of a generalizable risk prediction model for bipolar disorder.

Bipolar disorder is a leading contributor to disability, premature mortality, and suicide. Early identification of risk for bipolar disorder using generalizable predictive models trained on diverse cohorts around the United States could improve targeted assessment of high risk individuals, reduce misdiagnosis, and improve the allocation of limited mental health resources. This observational case-control study intended to develop and validate generalizable predictive models of bipolar disorder as part of the multisite, multinational PsycheMERGE Network across diverse and large biobanks with linked electronic health records (EHRs) from three academic medical centers: in the Northeast (Massachusetts General Brigham), the Mid-Atlantic (Geisinger) and the Mid-South (Vanderbilt University Medical Center). Predictive models were developed and valid with multiple algorithms at each study site: random forests, gradient boosting machines, penalized regression, including stacked ensemble learning algorithms combining them. Predictors were limited to widely available EHR-based features agnostic to a common data model including demographics, diagnostic codes, and medications. The main study outcome was bipolar disorder diagnosis as defined by the International Cohort Collection for Bipolar Disorder, 2015. In total, the study included records for 3,529,569 patients including 12,533 cases (0.3%) of bipolar disorder. After internal and external validation, algorithms demonstrated optimal performance in their respective development sites. The stacked ensemble achieved the best combination of overall discrimination (AUC = 0.82-0.87) and calibration performance with positive predictive values above 5% in the highest risk quantiles at all three study sites. In conclusion, generalizable predictive models of risk for bipolar disorder can be feasibly developed across diverse sites to enable precision medicine. Comparison of a range of machine learning methods indicated that an ensemble approach provides the best performance overall but required local retraining. These models will be disseminated via the PsycheMERGE Network website.

Limitations of Noninvasive Tests-Based Population-Level Risk Stratification Strategy for Nonalcoholic Fatty Liver Disease.

BACKGROUND: Nonalcoholic fatty liver disease (NAFLD) and nonalcoholic steatohepatitis (NASH) are highly prevalent but underdiagnosed. AIMS: We used an electronic health record data network to test a population-level risk stratification strategy using noninvasive tests (NITs) of liver fibrosis. METHODS: Data were obtained from PCORnet® sites in the East, Midwest, Southwest, and Southeast United States from patients aged [Formula: see text] 18 with or without ICD-10-CM diagnosis codes for NAFLD, NASH, and NASH-cirrhosis between 9/1/2017 and 8/31/2020. Average and standard deviations (SD) for Fibrosis-4 index (FIB-4), NAFLD fibrosis score (NFS), and Hepatic Steatosis Index (HSI) were estimated by site for each patient cohort. Sample-wide estimates were calculated as weighted averages across study sites. RESULTS: Of 11,875,959 patients, 0.8% and 0.1% were coded with NAFLD and NASH, respectively. NAFLD diagnosis rates in White, Black, and Hispanic patients were 0.93%, 0.50%, and 1.25%, respectively, and for NASH 0.19%, 0.04%, and 0.16%, respectively. Among undiagnosed patients, insufficient EHR data for estimating NITs ranged from 68% (FIB-4) to 76% (NFS). Predicted prevalence of NAFLD by HSI was 60%, with estimated prevalence of advanced fibrosis of 13% by NFS and 7% by FIB-4. Approximately, 15% and 23% of patients were classified in the intermediate range by FIB-4 and NFS, respectively. Among NAFLD-cirrhosis patients, a third had FIB-4 scores in the low or intermediate range. CONCLUSIONS: We identified several potential barriers to a population-level NIT-based screening strategy. HSI-based NAFLD screening appears unrealistic. Further research is needed to define merits of NFS- versus FIB-4-based strategies, which may identify different high-risk groups.

Geriatric assessment-derived deficit accumulation and patient-reported treatment burden in older adults with bladder cancer.

BACKGROUND: When a person's workload of healthcare exceeds their resources, they experience treatment burden. At the intersection of cancer and aging, little is known about treatment burden. We evaluated the association between a geriatric assessment-derived Deficit Accumulation Index (DAI) and patient-reported treatment burden in older adults with early-stage, non-muscle-invasive bladder cancer (NMIBC). METHODS: We conducted a cross-sectional survey of older adults with NMIBC (≥65 years). We calculated DAI using the Cancer and Aging Research Group's geriatric assessment and measured urinary symptoms using the Urogenital Distress Inventory-6 (UDI-6). The primary outcome was Treatment Burden Questionnaire (TBQ) score. A negative binomial regression with LASSO penalty was used to model TBQ. We further conducted qualitative thematic content analysis of responses to an open-ended survey question ("What has been your Greatest Challenge in managing medical care for your bladder cancer") and created a joint display with illustrative quotes by DAI category. RESULTS: Among 119 patients, mean age was 78.9 years (SD 7) of whom 56.3% were robust, 30.3% pre-frail, and 13.4% frail. In the multivariable model, DAI and UDI-6 were significantly associated with TBQ. Individuals with DAI above the median (>0.18) had TBQ scores 1.94 times greater than those below (adjusted IRR 1.94, 95% CI 1.33-2.82). Individuals with UDI-6 greater than the median (25) had TBQ scores 1.7 times greater than those below (adjusted IRR 1.70, 95% CI 1.16-2.49). The top 5 themes in the Greatest Challenge question responses were cancer treatments (22.2%), cancer worry (19.2%), urination bother (18.2%), self-management (18.2%), and appointment time (11.1%). CONCLUSIONS: DAI and worsening urinary symptoms were associated with higher treatment burden in older adults with NMIBC. These data highlight the need for a holistic approach that reconciles the burden from aging-related conditions with that resulting from cancer treatment.

Association Between Psychiatric Polygenic Scores, Healthcare Utilization and Comorbidity Burden.

Purpose: To estimate the association of psychiatric polygenic scores with healthcare utilization and comorbidity burden. Methods: Observational cohort study (N = 118,882) of adolescent and adult biobank participants with linked electronic health records (EHRs) from three diverse study sites; (Massachusetts General Brigham, Vanderbilt University Medical Center, Geisinger). Polygenic scores (PGS) were derived from the largest available GWAS of major depressive depression, bipolar disorder, and schizophrenia at the time of analysis. Negative binomial regression models were used to estimate the association between each psychiatric PGS and healthcare utilization and comorbidity burden. Healthcare utilization was measured as frequency of emergency department (ED), inpatient (IP), and outpatient (OP) visits. Comorbidity burden was defined by the Elixhauser Comorbidity Index and the Charlson Comorbidity Index. Results: Participants had a median follow-up duration of 12 years in the EHR. Individuals in the top decile of polygenic score for major depressive disorder had significantly more ED visits (RR=1.22, 95% CI; 1.17, 1.29) compared to those the lowest decile. Increases were also observed with IP and comorbidity burden. Among those diagnosed with depression and in the highest decile of the PGS, there was an increase in all utilization types (ED: RR=1.56, 95% CI 1.41, 1.72; OP: RR=1.16, 95% CI 1.08, 1.24; IP: RR=1.23, 95% CI 1.12, 1.36) post-diagnosis. No clinically significant results were observed with bipolar and schizophrenia polygenic scores. Conclusions: Polygenic score for depression is modestly associated with increased healthcare resource utilization and comorbidity burden, in the absence of diagnosis. Following a diagnosis of depression, the PGS was associated with further increases in healthcare utilization. These findings suggest that depression genetic risk is associated with utilization and burden of chronic disease in real-world settings.

Performance Characteristics of a Rule-Based Electronic Health Record Algorithm to Identify Patients with Gross and Microscopic Hematuria.

BACKGROUND: Two million patients per year are referred to urologists for hematuria, or blood in the urine. The American Urological Association recently adopted a risk-stratified hematuria evaluation guideline to limit multi-phase computed tomography to individuals at highest risk of occult malignancy. OBJECTIVES: To understand population-level hematuria evaluations, we developed an algorithm to accurately identify hematuria cases from electronic health records (EHRs). METHODS: We used International Classification of Diseases (ICD)-9/ICD-10 diagnosis codes, urine color, and urine microscopy values to identify hematuria cases and to differentiate between gross and microscopic hematuria. Using an iterative process, we refined the ICD-9 algorithm on a gold standard, chart-reviewed cohort of 3,094 hematuria cases, and the ICD-10 algorithm on a 300 patient cohort. We applied the algorithm to Geisinger patients ≥35 years (n = 539,516) and determined performance by conducting chart review (n = 500). RESULTS: After applying the hematuria algorithm, we identified 51,500 hematuria cases and 488,016 clean controls. Of the hematuria cases, 11,435 were categorized as gross, 26,658 as microscopic, 12,562 as indeterminate, and 845 were uncategorized. The positive predictive value (PPV) of identifying hematuria cases using the algorithm was 100% and the negative predictive value (NPV) was 99%. The gross hematuria algorithm had a PPV of 100% and NPV of 99%. The microscopic hematuria algorithm had lower PPV of 78% and NPV of 100%. CONCLUSION: We developed an algorithm utilizing diagnosis codes and urine laboratory values to accurately identify hematuria and categorize as gross or microscopic in EHRs. Applying the algorithm will help researchers to understand patterns of care for this common condition.

Yield of Familial Hypercholesterolemia Genetic and Phenotypic Diagnoses After Electronic Health Record and Genomic Data Screening.

Background Data mining of electronic health records to identify patients suspected of familial hypercholesterolemia (FH) has been limited by absence of both phenotypic and genomic data in the same cohort. Methods and Results Using the Geisinger MyCode Community Health Initiative cohort (n=130 257), we ran 2 screening algorithms (Mayo Clinic [Mayo] and flag, identify, network, deliver [FIND] FH) to determine FH genetic and phenotypic diagnostic yields. With 29 243 excluded by Mayo (for secondary causes of hypercholesterolemia, no lipid value in electronic health records), 52 034 excluded by FIND FH (insufficient data to run the model), and 187 excluded for prior FH diagnosis, a final cohort of 59 729 participants was created. Genetic diagnosis was based on presence of a pathogenic or likely pathogenic variant in FH genes. Charts from 180 variant-negative participants (60 controls, 120 identified by FIND FH and Mayo) were reviewed to calculate Dutch Lipid Clinic Network scores; a score ≥5 defined probable phenotypic FH. Mayo flagged 10 415 subjects; 194 (1.9%) had a pathogenic or likely pathogenic FH variant. FIND FH flagged 573; 34 (5.9%) had a pathogenic or likely pathogenic variant, giving a net yield from both of 197 out of 280 (70%). Confirmation of a phenotypic diagnosis was constrained by lack of electronic health record data on physical findings or family history. Phenotypic FH by chart review was present by Mayo and/or FIND FH in 13 out of 120 versus 2 out of 60 not flagged by either (P<0.09). Conclusions Applying 2 recognized FH screening algorithms to the Geisinger MyCode Community Health Initiative identified 70% of those with a pathogenic or likely pathogenic FH variant. Phenotypic diagnosis was rarely achievable due to missing data.

Development and Multi-Site External Validation of a Generalizable Risk Prediction Model for Bipolar Disorder.

Bipolar disorder is a leading contributor to disability, premature mortality, and suicide. Early identification of risk for bipolar disorder using generalizable predictive models trained on diverse cohorts around the United States could improve targeted assessment of high risk individuals, reduce misdiagnosis, and improve the allocation of limited mental health resources. This observational case-control study intended to develop and validate generalizable predictive models of bipolar disorder as part of the multisite, multinational PsycheMERGE Consortium across diverse and large biobanks with linked electronic health records (EHRs) from three academic medical centers: in the Northeast (Massachusetts General Brigham), the Mid-Atlantic (Geisinger) and the Mid-South (Vanderbilt University Medical Center). Predictive models were developed and validated with multiple algorithms at each study site: random forests, gradient boosting machines, penalized regression, including stacked ensemble learning algorithms combining them. Predictors were limited to widely available EHR-based features agnostic to a common data model including demographics, diagnostic codes, and medications. The main study outcome was bipolar disorder diagnosis as defined by the International Cohort Collection for Bipolar Disorder, 2015. In total, the study included records for 3,529,569 patients including 12,533 cases (0.3%) of bipolar disorder. After internal and external validation, algorithms demonstrated optimal performance in their respective development sites. The stacked ensemble achieved the best combination of overall discrimination (AUC = 0.82 - 0.87) and calibration performance with positive predictive values above 5% in the highest risk quantiles at all three study sites. In conclusion, generalizable predictive models of risk for bipolar disorder can be feasibly developed across diverse sites to enable precision medicine. Comparison of a range of machine learning methods indicated that an ensemble approach provides the best performance overall but required local retraining. These models will be disseminated via the PsycheMERGE Consortium website.

The Kidney Failure Risk Equation: Evaluation of Novel Input Variables including eGFR Estimated Using the CKD-EPI 2021 Equation in 59 Cohorts.

SIGNIFICANCE STATEMENT: The kidney failure risk equation (KFRE) uses age, sex, GFR, and urine albumin-to-creatinine ratio (ACR) to predict 2- and 5-year risk of kidney failure in populations with eGFR <60 ml/min per 1.73 m 2 . However, the CKD-EPI 2021 creatinine equation for eGFR is now recommended for use but has not been fully tested in the context of KFRE. In 59 cohorts comprising 312,424 patients with CKD, the authors assessed the predictive performance and calibration associated with the use of the CKD-EPI 2021 equation and whether additional variables and accounting for the competing risk of death improves the KFRE's performance. The KFRE generally performed well using the CKD-EPI 2021 eGFR in populations with eGFR <45 ml/min per 1.73 m 2 and was not improved by adding the 2-year prior eGFR slope and cardiovascular comorbidities. BACKGROUND: The kidney failure risk equation (KFRE) uses age, sex, GFR, and urine albumin-to-creatinine ratio (ACR) to predict kidney failure risk in people with GFR <60 ml/min per 1.73 m 2 . METHODS: Using 59 cohorts with 312,424 patients with CKD, we tested several modifications to the KFRE for their potential to improve the KFRE: using the CKD-EPI 2021 creatinine equation for eGFR, substituting 1-year average ACR for single-measure ACR and 1-year average eGFR in participants with high eGFR variability, and adding 2-year prior eGFR slope and cardiovascular comorbidities. We also assessed calibration of the KFRE in subgroups of eGFR and age before and after accounting for the competing risk of death. RESULTS: The KFRE remained accurate and well calibrated overall using the CKD-EPI 2021 eGFR equation. The other modifications did not improve KFRE performance. In subgroups of eGFR 45-59 ml/min per 1.73 m 2 and in older adults using the 5-year time horizon, the KFRE demonstrated systematic underprediction and overprediction, respectively. We developed and tested a new model with a spline term in eGFR and incorporating the competing risk of mortality, resulting in more accurate calibration in those specific subgroups but not overall. CONCLUSIONS: The original KFRE is generally accurate for eGFR <45 ml/min per 1.73 m 2 when using the CKD-EPI 2021 equation. Incorporating competing risk methodology and splines for eGFR may improve calibration in low-risk settings with longer time horizons. Including historical averages, eGFR slopes, or a competing risk design did not meaningfully alter KFRE performance in most circumstances.

Association of Supernumerary Sex Chromosome Aneuploidies With Venous Thromboembolism.

Importance: An increased risk of venous thromboembolism (VTE) has been reported in men with an additional sex chromosome. The association between other sex chromosome aneuploidies and VTE is not well characterized. Objective: To determine if sex chromosome aneuploidy is associated with VTE. Design, Setting, and Participants: Retrospective cohort study of sex chromosome aneuploidy and VTE, performed by analyzing X- and Y-chromosome dosage and VTE incidence in 642 544 individuals from 2 population-scale biobanks: the US Geisinger MyCode Community Health Initiative (N = 154 519) and the UK Biobank (N = 488 025); analysis was limited to participants self-identified as White because of inadequate sample sizes for other race and ethnicity groups. A total of 108 461 unrelated MyCode participants with electronic health record follow-up ranging from September 1996 to December 2020 and 418 725 unrelated British and Irish UK Biobank participants who attended the baseline assessment between March 2006 and October 2010, with follow-up extending to November 2020, were included in analyses of VTE. Exposures: Sex chromosome aneuploidies. Main Outcomes and Measures: Individuals with 1 primary inpatient VTE diagnosis, 2 primary outpatient VTE diagnoses, or a self-reported VTE diagnosis were defined as VTE cases. P values were adjusted for multiple comparisons. Results: Identification of sex chromosome aneuploidy was undertaken among 642 544 individuals aged 18 to 90 years. Identification of a diagnosis of VTE was undertaken among 108 461 unrelated MyCode participants (65 565 [60.5%] female; mean age at last visit, 58.0 [SD, 17.6] years; median follow-up, 15.3 [IQR, 9.7] years) and among 418 725 unrelated UK Biobank participants (224 695 [53.7%] female; mean age at baseline interview, 56.9 [SD, 8.0] years; median follow-up, 12.0 [IQR, 1.6] years). Among MyCode participants, during 10 years of follow-up, 17 incident VTE events per 1353 person-years were detected among those with supernumerary sex chromosome aneuploidy (1.3% per person-year) compared with 2060 per 816 682 person-years among those with 46,XX or 46,XY (0.25% per person-year) (hazard ratio, 5.4 [95% CI, 3.4-8.7]; 10-year risk difference, 8.8% [95% CI, 4.2%-14.0%]; P < .001). Among UK Biobank participants, during 10 years of follow-up, 16 incident VTE events per 3803 person-years were detected among those with supernumerary sex chromosome aneuploidy (0.42% per person-year) compared with 4491 per 3 970 467 person-years among those with 46,XX or 46,XY (0.11% per person-year) (hazard ratio, 4.1 [95% CI, 2.5-6.7]; 10-year risk difference, 3.7% [95% CI, 1.4%-5.9%]; P < .001). Conclusions and Relevance: Adults with supernumerary sex chromosome aneuploidies compared with 2 sex chromosomes had a small but statistically significant increased risk of VTE. Further research is needed to understand the clinical implications of this association.

Exploring Social Cognition Tests to Differentiate Frontotemporal Dementia from Depression: A Two-Step Pilot Study.

Behavioral-variant frontotemporal dementia (bvFTD) is challenging to recognize, and often misdiagnosed as depression (DEP). Evidence suggests changes in social cognition (SoCog) precede general cognitive decline in bvFTD. Currently, there are no screening measures of social cognition. 17 bvFTD, 16 DEP, and 18 control participants underwent 6 SoCog tests measuring: emotion recognition; theory of mind; empathy; insight. We used χ 2 , Wilcoxon rank sum, Kruskal-Wallis tests to compare groups, with decision tree analysis to identify items that best differentiated bvFTD from DEP. bvFTD performed significantly worse on all SoCog tasks compared with other groups. Decision tree analysis yielded a 5-item test with ROC area under the curve of 0.973 (95% CI: 0.928, 1.0) for differentiating bvFTD versus depression. These results suggest that it may be feasible to develop a screening measure of social cognition.

Optimal Timing of Antiretroviral Therapy Initiation in Children and Adolescents With Human Immunodeficiency Virus-Associated Pulmonary Tuberculosis.

BACKGROUND: There is insufficient evidence in children and adolescents with human immunodeficiency virus (CAHIV) to guide the timing of antiretroviral treatment (ART) initiation after starting treatment for pulmonary tuberculosis (pTB). To address this knowledge gap, we evaluated the risk of mortality associated with timing of ART initiation in ART-naive CAHIV treated for pTB. METHODS: Data were extracted from electronic medical records of ART-naive patients, aged 0-19 years, who were treated for HIV-associated pTB at Baylor Centers of Excellence in Botswana, Eswatini, Malawi, Lesotho, Tanzania, or Uganda between 2013 and 2020. Data were analyzed against a primary outcome of all-cause mortality with unadjusted Kaplan-Meier curves and Cox proportional hazard models. RESULTS: The study population included 774 CAHIV with variable intervals to ART initiation after starting TB treatment: <2 weeks (n = 266), 2 weeks to 2 months (n = 398), >2 months (n = 66), and no ART initiated (n = 44). Adjusted Cox proportional hazards models demonstrated increased mortality 1 year from TB treatment initiation in children never starting ART (adjusted HR [aHR]: 2.67; 95% CI: 1.03, 6.94) versus children initiating ART between 2 weeks and 2 months from TB treatment initiation. Mortality risk did not differ for the <2-weeks group (aHR: 1.02; 95% CI: .55, 1.89) versus the group initiating ART between 2 weeks and 2 months. CONCLUSIONS: This retrospective study demonstrated no increase in mortality among CAHIV initiating ART <2 weeks from TB treatment initiation. Given the broad health benefits of ART, this evidence supports the recent WHO recommendation for CAHIV to initiate ART within 2 weeks of initiating TB treatment.

Comparing enhancements to well-child visits in the prevention of obesity: ENCIRCLE cluster-randomized controlled trial.

BACKGROUND: Obesity disproportionally impacts rural, lower-income children in the United States. Primary care providers are well-positioned to engage parents in early obesity prevention, yet there is a lack of evidence regarding the most effective care delivery models. The ENCIRCLE study, a pragmatic cluster-randomized controlled trial, will respond to this gap by testing the comparative effectiveness of standard care well-child visits (WCV) versus two enhancements: adding a patient-reported outcome (PRO) measure (PRO WCV) and PRO WCV plus Food Care (telehealth coaching and a grocery store tour). METHODS: A total of 2,025 parents and their preschool-aged children (20-60 months of age) will be recruited from 24 Geisinger primary care clinics, where providers are randomized to the standard WCV, PRO WCV, or PRO WCV plus Food Care intervention arms. The PRO WCV includes the standard WCV plus collection of the PRO-the Family Nutrition and Physical Activity (FNPA) risk assessment-from parents. Parents complete the PRO in the patient-portal or in the clinic (own device, tablet, or kiosk), receive real-time feedback, and select priority topics to discuss with the provider. These results are integrated into the child's electronic health record to inform personalized preventive counseling by providers. PRO WCV plus Food Care includes referrals to community health professionals who deliver evidence-based obesity prevention and food resource management interventions via telehealth following the WCV. The primary study outcome is change in child body mass index z-score (BMIz), based on the World Health Organization growth standards, 12 months post-baseline WCV. Additional outcomes include percent of children with overweight and obesity, raw BMI, BMI50, BMIz extended, parent involvement in counseling, health behaviors, food resource management, and implementation process measures. DISCUSSION: Study findings will inform health care systems' choices about effective care delivery models to prevent childhood obesity among a high-risk population. Additionally, dissemination will be informed by an evaluation of mediating, moderating, and implementation factors. TRIAL REGISTRATION: ClinicalTrials.gov identifier (NCT04406441); Registered May 28, 2020.

A Simpler Machine Learning Model for Acute Kidney Injury Risk Stratification in Hospitalized Patients.

Background: Hospitalization-associated acute kidney injury (AKI), affecting one-in-five inpatients, is associated with increased mortality and major adverse cardiac/kidney endpoints. Early AKI risk stratification may enable closer monitoring and prevention. Given the complexity and resource utilization of existing machine learning models, we aimed to develop a simpler prediction model. Methods: Models were trained and validated to predict risk of AKI using electronic health record (EHR) data available at 24 h of inpatient admission. Input variables included demographics, laboratory values, medications, and comorbidities. Missing values were imputed using multiple imputation by chained equations. Results: 26,410 of 209,300 (12.6%) inpatients developed AKI during admission between 13 July 2012 and 11 July 2018. The area under the receiver operating characteristic curve (AUROC) was 0.86 for Random Forest and 0.85 for LASSO. Based on Youden's Index, a probability cutoff of >0.15 provided sensitivity and specificity of 0.80 and 0.79, respectively. AKI risk could be successfully predicted in 91% patients who required dialysis. The model predicted AKI an average of 2.3 days before it developed. Conclusions: The proposed simpler machine learning model utilizing data available at 24 h of admission is promising for early AKI risk stratification. It requires external validation and evaluation of effects of risk prediction on clinician behavior and patient outcomes.