Health data science最新文献

Background: There is a lot of fact-based information and misinformation in the online discourses and discussions about the COVID-19 vaccines.

Method: Using a sample of nearly four million geotagged English tweets and the data from the CDC COVID Data Tracker, we conducted the Fama-MacBeth regression with the Newey-West adjustment to understand the influence of both misinformation and fact-based news on Twitter on the COVID-19 vaccine uptake in the US from April 19 when US adults were vaccine eligible to June 30, 2021, after controlling state-level factors such as demographics, education, and the pandemic severity. We identified the tweets related to either misinformation or fact-based news by analyzing the URLs.

Results: One percent increase in fact-related Twitter users is associated with an approximately 0.87 decrease (B = -0.87, SE = 0.25, and p < .001) in the number of daily new vaccinated people per hundred. No significant relationship was found between the percentage of fake-news-related users and the vaccination rate.

Conclusion: The negative association between the percentage of fact-related users and the vaccination rate might be due to a combination of a larger user-level influence and the negative impact of online social endorsement on vaccination intent.

Background: During the COVID-19 pandemic, mental health concerns (such as fear and loneliness) have been actively discussed on social media. We aim to examine mental health discussions on Twitter during the COVID-19 pandemic in the US and infer the demographic composition of Twitter users who had mental health concerns.

Methods: COVID-19-related tweets from March 5^th, 2020, to January 31^st, 2021, were collected through Twitter streaming API using keywords (i.e., "corona," "covid19," and "covid"). By further filtering using keywords (i.e., "depress," "failure," and "hopeless"), we extracted mental health-related tweets from the US. Topic modeling using the Latent Dirichlet Allocation model was conducted to monitor users' discussions surrounding mental health concerns. Deep learning algorithms were performed to infer the demographic composition of Twitter users who had mental health concerns during the pandemic.

Results: We observed a positive correlation between mental health concerns on Twitter and the COVID-19 pandemic in the US. Topic modeling showed that "stay-at-home," "death poll," and "politics and policy" were the most popular topics in COVID-19 mental health tweets. Among Twitter users who had mental health concerns during the pandemic, Males, White, and 30-49 age group people were more likely to express mental health concerns. In addition, Twitter users from the east and west coast had more mental health concerns.

Conclusions: The COVID-19 pandemic has a significant impact on mental health concerns on Twitter in the US. Certain groups of people (such as Males and White) were more likely to have mental health concerns during the COVID-19 pandemic.

Background. Increasing evidence from human studies has revealed the adverse impact of ambient fine particles (PM _2.5) on health outcomes related to metabolic disorders and distant organs. Whether exposure to ambient PM _2.5 leads to kidney impairment remains unclear. The rapid air quality improvement driven by the clean air actions in China since 2013 provides an opportunity for a quasiexperiment to investigate the beneficial effect of PM _2.5 reduction on kidney function.Methods. Based on two repeated nationwide surveys of the same population of 5115 adults in 2011 and 2015, we conducted a difference-in-difference study. Variations in long-term exposure to ambient PM _2.5 were associated with changes in kidney function biomarkers, including estimated glomerular filtration rate by serum creatinine (GFR _scr) or cystatin C (GFR _cys), blood urea nitrogen (BUN), and uric acid (UA).Results. For a 10  μg/m ³ reduction in PM _2.5, a significant improvement was observed for multiple kidney functional biomarkers, including GFR _scr, BUN and UA, with a change of 0.42 (95% confidence interval [CI]: 0.06, 0.78) mL/min/1.73m ², -0.38 (-0.64, -0.12) mg/dL, and -0.06 (-0.12, -0.00) mg/dL, respectively. A lower socioeconomic status, indicated by rural residence or low educational level, enhanced the adverse effect of PM _2.5 on kidney function.Conclusions. These results support a significant nephrotoxicity of PM _2.5 based on multiple serum biomarkers and indicate a beneficial effect of improved air quality on kidney function.

Background: The behaviors and emotions associated with and reasons for nonmedical prescription drug use (NMPDU) are not well-captured through traditional instruments such as surveys and insurance claims. Publicly available NMPDU-related posts on social media can potentially be leveraged to study these aspects unobtrusively and at scale.

Methods: We applied a machine learning classifier to detect self-reports of NMPDU on Twitter and extracted all public posts of the associated users. We analyzed approximately 137 million posts from 87,718 Twitter users in terms of expressed emotions, sentiments, concerns, and possible reasons for NMPDU via natural language processing.

Results: Users in the NMPDU group express more negative emotions and less positive emotions, more concerns about family, the past, and body, and less concerns related to work, leisure, home, money, religion, health, and achievement compared to a control group (i.e., users who never reported NMPDU). NMPDU posts tend to be highly polarized, indicating potential emotional triggers. Gender-specific analyses show that female users in the NMPDU group express more content related to positive emotions, anticipation, sadness, joy, concerns about family, friends, home, health, and the past, and less about anger than males. The findings are consistent across distinct prescription drug categories (opioids, benzodiazepines, stimulants, and polysubstance).

Conclusion: Our analyses of large-scale data show that substantial differences exist between the texts of the posts from users who self-report NMPDU on Twitter and those who do not, and between males and females who report NMPDU. Our findings can enrich our understanding of NMPDU and the population involved.

Background: There is an increasing trend to represent domain knowledge in structured graphs, which provide efficient knowledge representations for many downstream tasks. Knowledge graphs are widely used to model prior knowledge in the form of nodes and edges to represent semantically connected knowledge entities, which several works have adopted into different medical imaging applications.

Methods: We systematically searched over five databases to find relevant articles that applied knowledge graphs to medical imaging analysis. After screening, evaluating, and reviewing the selected articles, we performed a systematic analysis.

Results: We looked at four applications in medical imaging analysis, including disease classification, disease localization and segmentation, report generation, and image retrieval. We also identified limitations of current work, such as the limited amount of available annotated data and weak generalizability to other tasks. We further identified the potential future directions according to the identified limitations, including employing semisupervised frameworks to alleviate the need for annotated data and exploring task-agnostic models to provide better generalizability.

Conclusions: We hope that our article will provide the readers with aggregated documentation of the state-of-the-art knowledge graph applications for medical imaging to encourage future research.

Background: Angiotensin-converting enzyme inhibitors (ACEi) and angiotensin-II receptor blockers (ARB), the most commonly prescribed antihypertensive medications, counter renin-angiotensin-aldosterone system (RAAS) activation via induction of angiotensin-converting enzyme 2 (ACE2) expression. Considering that ACE2 is the functional receptor for SARS-CoV-2 entry into host cells, the association of ACEi and ARB with COVID-19 outcomes needs thorough evaluation.

Methods: We conducted retrospective analyses using both unmatched and propensity score (PS)-matched cohorts on electronic health records (EHRs) to assess the impact of RAAS inhibitors on the risk of receiving invasive mechanical ventilation (IMV) and 30-day mortality among hospitalized COVID-19 patients. Additionally, we investigated the immune cell gene expression profiles of hospitalized COVID-19 patients with prior use of antihypertensive treatments from an observational prospective cohort.

Results: The retrospective analysis revealed that there was no increased risk associated with either ACEi or ARB use. In fact, the use of ACEi showed decreased risk for mortality. Survival analyses using PS-matched cohorts suggested no significant relationship between RAAS inhibitors with a hospital stay and in-hospital mortality compared to non-RAAS medications and patients not on antihypertensive medications. From the analysis of gene expression profiles, we observed a noticeable up-regulation in the expression of 1L1R2 (an anti-inflammatory receptor) and RETN (an immunosuppressive marker) genes in monocytes among prior users of ACE inhibitors.

Conclusion: Overall, the findings do not support the discontinuation of ACEi or ARB treatment and suggest that ACEi may moderate the COVID-19 hyperinflammatory response.

Background. In critical care, intensivists are required to continuously monitor high-dimensional vital signs and lab measurements to detect and diagnose acute patient conditions, which has always been a challenging task. Recently, deep learning models such as recurrent neural networks (RNNs) have demonstrated their strong potential on predicting such events. However, in real deployment, the patient data are continuously coming and there is no effective adaptation mechanism for RNN to incorporate those new data and become more accurate.Methods. In this study, we propose a novel self-correcting mechanism for RNN to fill in this gap. Our mechanism feeds prediction errors from the predictions of previous timestamps into the prediction of the current timestamp, so that the model can "learn" from previous predictions. We also proposed a regularization method that takes into account not only the model's prediction errors on the labels but also its estimation errors on the input data.Results. We compared the performance of our proposed method with the conventional deep learning models on two real-world clinical datasets for the task of acute kidney injury (AKI) prediction and demonstrated that the proposed model achieved an area under ROC curve at 0.893 on the MIMIC-III dataset and 0.871 on the Philips eICU dataset.Conclusions. The proposed self-correcting RNNs demonstrated effectiveness in AKI prediction and have the potential to be applied to clinical applications.

Objective. The strongest locus which associated with type 2 diabetes (T2D) by the common variant rs7903146 is the transcription factor 7-like 2 gene (TCF7L2). We aimed to quantify the interaction of diet/lifestyle interventions and the genetic effect of TCF7L2 rs7903146 on glycemic traits, body weight, or waist circumference in overweight or obese adults in several randomized controlled trials (RCTs).Methods. From October 2016 to May 2018, a large collaborative analysis was performed by pooling individual-participant data from 7 RCTs. These RCTs reported changes in glycemic control and adiposity of the variant rs7903146 after dietary/lifestyle-related interventions in overweight or obese adults. Gene treatment interaction models which used the genetic effect encoded by the allele dose and common covariates were applicable to individual participant data in all studies.Results. In the joint analysis, a total of 7 eligible RCTs were included ($n=\mathrm{4,114}$). Importantly, we observed a significant effect modification of diet/lifestyle-related interventions on the TCF7L2 variant rs7903146 and changes in fasting glucose. Compared with the control group, diet/lifestyle interventions were related to lower fasting glucose by -3.06 (95% CI, -5.77 to -0.36) mg/dL (test for heterogeneity and overall effect: $I^2=45.1\%$, $p<0.05$; $z=2.20$, $p=0.028$) per one copy of the TCF7L2 T risk allele. Furthermore, regardless of genetic risk, diet/lifestyle interventions were associated with lower waist circumference. However, there was no significant change for diet/lifestyle interventions in other glycemic control and adiposity traits per one copy of TCF7L2 risk allele.Conclusions. Our findings suggest that carrying the TCF7L2 T risk allele may have a modestly greater benefit for specific diet/lifestyle interventions to improve the control of fasting glucose in overweight or obese adults.