.
RATIONALE:Mycobacterium avium complex lung disease (MAC-LD) is clinically heterogeneous and carries diverse outcomes. OBJECTIVES:To describe predictors of clinical progression of MAC-LD in a state-wide cohort. METHODS:We enrolled adults with MAC-LD from across Virginia, USA starting in 2021. Every 6 months we performed respiratory quality of life questionnaire, scored CT scans, and recorded respiratory mycobacterial cultures including MAC speciation. Outcomes were classified using NTM-NET consensus definitions, factors predicting clinical progression analyzed by Poisson regression, and hierarchical clustering on principal components derived from Factorial Analysis of Mixed Data. MEASUREMENTS AND MAIN RESULTS:Of 105 participants the median follow-up was 917 days. Mean age was 69.8 years, 79 (75%) were women, and 70 (67%) had nodular bronchiectasis. M. intracellulare was the most common species, present in 48 (46%) participants at enrollment, followed by M. avium (29, 28%), and M. intracellulare subspecies chimaera (11, 10%). Only 2 (9%) of 22 evaluable participants met the NTM-NET definition of cure. In all participants after multivariable adjustment, older baseline age (incidence rate ratio 1.03 [1, 1.06], p = 0.04) and fibrocavitary CT scan pattern (2.57 [1.33, 4.96], p = 0.005), were associated with unfavorable 12-month clinical progression. Species type and species persistence contributed to characteristics of three distinct phenotypes of MAC-LD of varying severity and clinical progression. CONCLUSIONS:The majority of participants were not assessable for MAC-LD treatment outcomes using strict NTM-NET definitions. Species informed phenotypes of MAC lung disease are prognostically useful and can inform routine management and trial design.
BACKGROUND:Gaps in public health data include jurisdiction specific data systems that are used to measure progress on key regional and national indicators. The ATra Black Box (Box) is an electronic privacy-assuring system developed by Georgetown University which allows for the secure and streamlined exchange and analysis of sensitive data. An enhancement was added to the Box to calculate HIV Care Continuum measures in the Ryan White Part A DC Eligible Metropolitan Area (DC EMA). SETTING:The DC EMA includes DC, southern Maryland, northern Virginia, and 2 counties in West Virginia. METHODS:Georgetown implemented new functionality in the Box to create a DC EMA wide report of deduplicated care continuum data. SAS codes used the new Box functionality to select persons living in the DC EMA counties and produce the HIV Care Continuum for the DC EMA for calendar year 2024, including stratifications by race, sex, age, and transmission category, and place of care receipt. RESULTS:51,033 duplicated and 39,047 deduplicated persons were identified as alive and residing in the DC EMA in 2024. Overall, 68.4% of these persons received HIV care and 61.6% achieved viral suppression as of December 31, 2024. CONCLUSION:This analysis provides deduplicated estimates of persons living with HIV in the DC EMA and key indicators used to measure progress on ending HIV in the United States. The jurisdictions in the DC EMA are able to more accurately monitor engagement in care and viral suppression and more effectively guide public health efforts and resources.
Wastewater-based surveillance is an effective tool for disease monitoring and can provide early warning of outbreaks. Although wastewater viral loads (WVL) correlate with disease burden, their utility for improving real-time forecasting remains under investigation. During the early phases of an epidemic, many indicators can effectively monitor disease spread, but their reliability may decline because of reporting fatigue and low prevalence. Hospital burden can vary substantially even during low-prevalence periods, making accurate forecasting of burden indicators essential for minimizing disease impacts. In this paper, we present principled approaches for processing wastewater data, characterizing its relationship with burden indicators, and generating real-time forecasts. We assess the predictability of WVL using entropy measures. We analyze the relationship between WVL and burden indicators using causality tests that capture temporal dynamics and the leading-indicator behavior of WVL. We incorporate these insights into a time-varying forecasting model that accounts for the evolving relationship between the signals. We also evaluate the effects of delays in WVL reporting through simulations. We test the utility of our methods by forecasting COVID-19 hospital admissions across Virginia and its health regions during periods of varying disease prevalence. Incorporating WVL improves forecast accuracy relative to baseline models, particularly during critical epidemic phases, and results in a 20 percentage point improvement in forecast coverage. Our results demonstrate that WVL signals can improve infectious disease forecasting even under conditions of low prevalence or delayed reporting.
Health care coverage is key for health among people with HIV (PWH). We assessed differences in health care coverage and HIV outcomes and sexually transmitted infection (STI) testing between pre- and post-Medicaid expansion periods among PWH in Virginia. We analyzed Virginia's CDC Medical Monitoring Project data for 2015-2022 (N = 1169). Weighted percentages were assessed for characteristics. Differences in characteristics and outcomes between the pre-expansion (2015-2018) and post-expansion (2020-2022) were assessed using prevalence differences (PDs) and 95% CIs with predicted marginal means; data from 2019 were excluded as the transition year. The number of PWH covered by Medicaid was higher post- than pre-expansion (30.1%-33.9% vs. 18.9%-25.4%). Overall, the percentage of PWH missing ≥1 antiretroviral therapy (ART) dose in the past 30 days was 37.6% in pre- and 29.9% in post-expansion (PD: -7.7; 95% CI: -14.4-1.1). Rates of testing for gonorrhea, chlamydia, and syphilis during the past 12 months were higher in post- than pre-expansion (34.8% vs. 27.8%; PD: 6.9; 95% CI: 0.5-13.3). Medicaid coverage increases among PWH following expansion were associated with increases in ART adherence and STI testing. Expanding health care coverage options that decrease cost barriers may support ART adherence and STI screening among PWH.
Federated learning has become one of the most revolutionary methods of collaborative model training in distributed institutions of healthcare without access to raw patient data. In this paper, the researcher proposes a robust privacy-preserving federated learning architecture suitable for healthcare 5.0 settings to predict chronic diseases. The framework ensures a high degree of privacy alongside a high predictive accuracy by incorporating both the mechanisms of differential privacy and secure aggregation protocols into an edge intelligence architecture. The combination of edge device-encrypted model updates at four healthcare centers, also on the aggregation server, results in a global predictive model that achieves a root mean square error of 0.1892 and an overall classification accuracy of 95.3
