Impact of COVID-19 Pandemic on Sleep Including HRV and Physical Activity as Mediators: A Causal ML Approach (preprint)

Sleep quality is crucial to both mental and physical well-being. The COVID-19 pandemic, which has notably affected the population's health worldwide, has been shown to deteriorate people's sleep quality. Numerous studies have been conducted to evaluate the impact of the COVID-19 pandemic on sleep efficiency, investigating their relationships using correlation based methods. These methods merely rely on learning spurious correlation rather than the causal relations among variables. Furthermore, they fail to pinpoint potential sources of bias and mediators and envision counterfactual scenarios, leading to a poor estimation. In this paper, we develop a Causal Machine Learning method, which encompasses causal discovery and causal inference components, to extract the causal relations between the COVID-19 pandemic (treatment variable) and sleep quality (outcome) and estimate the causal treatment effect, respectively. We conducted a wearable-based health monitoring study to collect data, including sleep quality, physical activity, and Heart Rate Variability (HRV) from college students before and after the COVID-19 lockdown in March 2020. Our causal discovery component generates a causal graph and pinpoints mediators in the causal model. We incorporate the strongly contributing mediators (i.e., HRV and physical activity) into our causal inference component to estimate the robust, accurate, and explainable causal effect of the pandemic on sleep quality. Finally, we validate our estimation via three refutation analysis techniques. Our experimental results indicate that the pandemic exacerbates college students' sleep scores by 8%. Our validation results show significant p-values confirming our estimation.

Estimating the Impact of Pre-Exposure Prophylaxis (PrEP) on Mortality in COVID-19 Patients: A Causal Inference Approach (preprint)

Traditional machine learning (ML) approaches learn to recognize patterns in the data but fail to go beyond observing associations. Such data-driven methods can lack generalizability when the data is outside the independent and identically distributed (i.i.d) setting. Using causal inference can aid data-driven techniques to go beyond learning spurious associations and frame the data-generating process in a causal lens. We can combine domain expertise and traditional ML techniques to answer causal questions on the data. Hypothetical questions on alternate realities can also be answered with such a framework. In this paper, we estimate the causal effect of Pre-Exposure Prophylaxis (PrEP) on mortality in COVID-19 patients from an observational dataset of over 120,000 patients. With the help of medical experts, we hypothesize a causal graph that identifies the causal and non-causal associations, including the list of potential confounding variables. We use estimation techniques such as linear regression, matching, and machine learning (meta-learners) to estimate the causal effect. On average, our estimates show that taking PrEP can result in a 2.1% decrease in the death rate or a total of around 2,540 patients lives saved in the studied population.

The Effect of SARS-COV-2 Variant on Respiratory Features and Mortality Among Vaccinated and Non-Fully Vaccinated Patients (preprint)

Background SARS-CoV-2 (COVID-19) has caused over 80 million infections and 973,000 deaths in the United States, and mutations are linked to increased transmissibility. This study aimed to determine the effect of SARS-CoV-2 variants on respiratory features and mortality and to determine the effect of vaccination status. Method A retrospective review of medical records (n=63,454 unique patients) using The University of California Health COvid Research Data Set (UC CORDS) was performed to identify respiratory features, vaccination status, and mortality. Variants were identified using the CDC data tracker. Results Increased odds of death were observed among those not fully vaccinated (Delta OR: 1.64, p = 0.052; Omicron OR: 1.96, p < 0.01). Later variants (i.e., Delta and Omicron) demonstrated a reduction in the frequency of lower respiratory tract features with a concomitant increase in upper respiratory tract features. Vaccination status was associated with survival and a decrease in the frequency of many upper and lower respiratory tract features. Discussion SARS-CoV-2 variants show a reduction in lower respiratory tract features with an increase in upper respiratory tract features. Being fully vaccinated results in fewer respiratory features and higher odds of survival, supporting vaccination in preventing morbidity and mortality from COVID-19.

A distinct symptom pattern emerges for COVID-19 Long-Haul: A nationwide study (preprint)

Long-haul COVID-19, also called Post-Acute Sequelae of SARS-CoV-2 (PASC), is a new illness caused by SARS-CoV-2 infection and characterized by the persistence of symptoms. The purpose of this cross-sectional study was to identify a distinct and significant temporal pattern of PASC symptoms (symptom type and onset) among a nationwide sample of PASC survivors (n= 5,652). The sample was randomly sorted into two independent samples for exploratory (EFA) and confirmatory factor analyses (CFA). Five factors emerged from the EFA: (1) cold & flu-like symptoms, (2) change in smell and/or taste, (3) dyspnea and chest pain, (4) cognitive & visual problems, and (5) cardiac symptoms. The CFA had excellent model fit (x2 = 513.721, df= 207, p<0.01, TLI= 0.952, CFI= 0.964, RMSEA= 0.024). These findings demonstrate a novel symptom pattern for PASC. These findings can enable nurses in the identification of at-risk patients and facilitate early, systematic symptom management strategies for PASC.

Nine Months in the Life of a College Student During 2020: A Case Study Using Multi-Modal Assessments to Capture Personalized Contexts of Well-being (preprint)

Emerging adulthood is a period marked by tremendous growth and opportunity, but also of significant risk. It is during this phase that the majority of major disease processes have their onset and also the time during which individuals develop habits for adaptive ways of managing stress and help-seeking when necessary. Recent work has demonstrated the feasibility of estimating stress levels through monitoring of physiological signs of sympathetic nervous system activity (e.g., heart rate, heart rate variability, respiration rate, galvanic skin response, etc.) using wearable devices. However, this method of objective monitoring only captures changes in internal states but not the contextual factors – such as mental activity and social interactions – that are critical for diagnosis, treatment and prevention of mental health problems that may follow from the long-term effects of stress. Thus, there is a fundamental need to capture higher-level life events and contextual information to enable root cause analysis for treatment and prevention. The proposed case study is part of a larger pilot study in which we intensely follow a small sample of college students over the course of 9-months during an historic and tumultuous year in which COVID-19 disrupted daily living to understand contextual factors relevant to their mental health while capturing mood, sleep, physical activity, and physiology. In this case study, we showcase a single participant as support for the feasibility and potential of this approach for understanding more personalized models for mental health and treatment.