Community health centers (CHCs) pivoted to using telehealth to deliver chronic care during the coronavirus COVID-19 pandemic. While care continuity can improve care quality and patients' experiences, it is unclear whether telehealth supported this relationship.We examine the association of care continuity with diabetes and hypertension care quality in CHCs before and during COVID-19 and the mediating effect of telehealth.This was a cohort study.Electronic health record data from 166 CHCs with n=20,792 patients with diabetes and/or hypertension with ≥2 encounters/year during 2019 and 2020.Multivariable logistic regression models estimated the association of care continuity (Modified Modified Continuity Index; MMCI) with telehealth use and care processes. Generalized linear regression models estimated the association of MMCI and intermediate outcomes. Formal mediation analyses assessed whether telehealth mediated the association of MMCI with A1c testing during 2020.MMCI [2019: odds ratio (OR)=1.98, marginal effect=0.69, z=165.50, P<0.001; 2020: OR=1.50, marginal effect=0.63, z=147.73, P<0.001] and telehealth use (2019: OR=1.50, marginal effect=0.85, z=122.87, P<0.001; 2020: OR=10.00, marginal effect=0.90, z=155.57, P<0.001) were associated with higher odds of A1c testing. MMCI was associated with lower systolic (β=-2.90, P<0.001) and diastolic blood pressure (β=-1.44, P<0.001) in 2020, and lower A1c values (2019: β=-0.57, P=0.007; 2020: β=-0.45, P=0.008) in both years. In 2020, telehealth use mediated 38.7% of the relationship between MMCI and A1c testing.Higher care continuity is associated with telehealth use and A1c testing, and lower A1c and blood pressure. Telehealth use mediates the association of care continuity and A1c testing. Care continuity may facilitate telehealth use and resilient performance on process measures.
We introduce an annotated corpus of 600 ophthalmology notes labeled with detailed spatial and contextual information of ophthalmic entities. We extend our previously proposed frame semantics-based spatial representation schema, Rad-SpatialNet, to represent spatial language in ophthalmology text, resulting in the Eye-SpatialNet schema. The spatially-grounded entities are findings, procedures, and drugs. To accurately capture all spatial details, we add some domain-specific elements in Eye-SpatialNet. The annotated corpus contains 1715 spatial triggers, 7308 findings, 2424 anatomies, and 9914 descriptors. To automatically extract the spatial information, we employ a two-turn question answering approach based on the transformer language model BERT. The results are promising, with F1 scores of 89.31, 74.86, and 88.47 for spatial triggers, Figure, and Ground frame elements, respectively. This is the first work to represent and extract a wide variety of clinical information in ophthalmology. Extracting detailed information can benefit ophthalmology applications and research targeted toward disease progression and screening.
This paper applies multiple machine learning (ML) algorithms to a dataset of de-identified COVID-19 patients provided by the COVID-19 Research Database. The dataset consists of 20,878 COVID-positive patients, among which 9,177 patients died in the year 2020. This paper aims to understand and interpret the association of socio-economic characteristics of patients with their mortality instead of maximizing prediction accuracy. According to our analysis, a patient's household's annual and disposable income, age, education, and employment status significantly impacts a machine learning model's prediction. We also observe several individual patient data, which gives us insight into how the feature values impact the prediction for that data point. This paper analyzes the global and local interpretation of machine learning models on socio-economic data of COVID patients.
Liver transplant is an essential therapy performed for severe liver diseases. The fact of scarce liver resources makes the organ assigning crucial. Model for End-stage Liver Disease (MELD) score is a widely adopted criterion when making organ distribution decisions. However, it ignores post-transplant outcomes and organ/donor features. These limitations motivate the emergence of machine learning (ML) models. Unfortunately, ML models could be unfair and trigger bias against certain groups of people. To tackle this problem, this work proposes a fair machine learning framework targeting graft failure prediction in liver transplant. Specifically, knowledge distillation is employed to handle dense and sparse features by combining the advantages of tree models and neural networks. A two-step debiasing method is tailored for this framework to enhance fairness. Experiments are conducted to analyze unfairness issues in existing models and demonstrate the superiority of our method in both prediction and fairness performance.
Siru Liu, PhD1, Allison B. McCoy, PhD1, Bryan Steitz, PhD1, Adam Wright, PhD1 1Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN Introduction Lung cancer is the leading cause of cancer mortality in the US, accounting for 25% of all cancer deaths.1 Lung cancer screening (LCS) has been shown to reduce mortality rates by 20%2, but only 5% of eligible adults have received LCS nationally.3 Several gaps exist in the implementation of LCS, particularly in the critical first step of identifying eligible patients. The eligibility criteria rely on smoking history, which is inconsistently documented in electronic health records (EHRs).4 Structured documentation tends to be incomplete, but additional information may be present in clinical notes. These shortcoming make it hard for clinicians, and especially for clinical decision support (CDS) tools, to recognize patients’ full smoking history.5 Moreover, the 2021 United States Preventive Services Task Force (USPSTF) LCS guideline nearly doubles the number of individuals eligible for LCS to 14.5 million Americans,6 so efficient strategies for identifying and referring these patients are needed. In this study, we sought to develop a natural language processing (NLP) pipeline to systematically extract smoking information from clinical notes to identify patients eligible for LCS.
