scholarly journals On the Generation of Medical Dialogues for COVID-19

2020 ◽  
Author(s):  
Wenmian Yang ◽  
Guangtao Zeng ◽  
Bowen Tan ◽  
Zeqian Ju ◽  
Subrato Chakravorty ◽  
...  
Keyword(s):  
Risk Factors ◽  
High Risk ◽  
Transfer Learning ◽  
Large Scale ◽  
Medical Professionals ◽  
Dialogue System ◽  
Consulting Services ◽  
Medical Dialogue ◽  
Hospital Closure

Under the pandemic of COVID-19, people experiencing COVID19-related symptoms or exposed to risk factors have a pressing need to consult doctors. Due to hospital closure, a lot of consulting services have been moved online. Because of the shortage of medical professionals, many people cannot receive online consultations timely. To address this problem, we aim to develop a medical dialogue system that can provide COVID19-related consultations. We collected two dialogue datasets - CovidDialog - (in English and Chinese respectively) containing conversations between doctors and patients about COVID-19. On these two datasets, we train several dialogue generation models based on Transformer, GPT, and BERT-GPT. Since the two COVID-19 dialogue datasets are small in size, which bear high risk of overfitting, we leverage transfer learning to mitigate data deficiency. Specifically, we take the pretrained models of Transformer, GPT, and BERT-GPT on dialog datasets and other large-scale texts, then finetune them on our CovidDialog datasets. Experiments demonstrate that these approaches are promising in generating meaningful medical dialogue about COVID-19. But more advanced approaches are needed to build a fully useful dialogue system that can offer accurate COVID-related consultations. The data and code are available at https://github.com/UCSD-AI4H/COVID-Dialogue

scholarly journals On the Generation of Medical Dialogues for COVID19

2020 ◽  
Author(s):  
Wenmian Yang ◽  
Guangtao Zeng ◽  
Bowen Tan ◽  
Zeqian Ju ◽  
Subrato Chakravorty ◽  
...  
Keyword(s):  
Risk Factors ◽  
High Risk ◽  
Transfer Learning ◽  
Large Scale ◽  
Medical Professionals ◽  
Dialogue System ◽  
Consulting Services ◽  
Medical Dialogue ◽  
Hospital Closure

<div>Under the pandemic of COVID-19, people experiencing COVID19-related symptoms or exposed to risk factors have a pressing need to consult doctors. Due to hospital closure,</div><div>a lot of consulting services have been moved online. Because of the shortage of medical professionals, many people cannot receive online consultations timely. To address this problem, we aim to develop a medical dialogue system that can provide COVID19-related consultations. We collected two dialogue datasets - CovidDialog - (in English and Chinese respectively) containing conversations between doctors and patients about COVID-19. On these two datasets, we train several dialogue generation models based on Transformer, GPT, and BERT-GPT. Since the two COVID-19 dialogue datasets are small in size, which bear high risk of overftting, we leverage transfer learning to mitigate data deficiency. Specifically, we take the pretrained models of Transformer, GPT, and BERT-GPT on dialog datasets and other large-scale texts, then finetune them on our CovidDialog datasets. Experiments demonstrate that these approaches are promising in generating meaningful medical dialogues about COVID-19. But more advanced approaches are needed to build a fully useful dialogue system that can offer accurate COVID-related consultations. The data and code are available at https://github.com/UCSD-AI4H/COVID-Dialogue</div>

scholarly journals On the Generation of Medical Dialogues for COVID19

2020 ◽  
Author(s):  
Wenmian Yang ◽  
Guangtao Zeng ◽  
Bowen Tan ◽  
Zeqian Ju ◽  
Subrato Chakravorty ◽  
...  
Keyword(s):  
Risk Factors ◽  
High Risk ◽  
Transfer Learning ◽  
Large Scale ◽  
Medical Professionals ◽  
Dialogue System ◽  
Consulting Services ◽  
Medical Dialogue ◽  
Hospital Closure

<div>Under the pandemic of COVID-19, people experiencing COVID19-related symptoms or exposed to risk factors have a pressing need to consult doctors. Due to hospital closure,</div><div>a lot of consulting services have been moved online. Because of the shortage of medical professionals, many people cannot receive online consultations timely. To address this problem, we aim to develop a medical dialogue system that can provide COVID19-related consultations. We collected two dialogue datasets - CovidDialog - (in English and Chinese respectively) containing conversations between doctors and patients about COVID-19. On these two datasets, we train several dialogue generation models based on Transformer, GPT, and BERT-GPT. Since the two COVID-19 dialogue datasets are small in size, which bear high risk of overftting, we leverage transfer learning to mitigate data deficiency. Specifically, we take the pretrained models of Transformer, GPT, and BERT-GPT on dialog datasets and other large-scale texts, then finetune them on our CovidDialog datasets. Experiments demonstrate that these approaches are promising in generating meaningful medical dialogues about COVID-19. But more advanced approaches are needed to build a fully useful dialogue system that can offer accurate COVID-related consultations. The data and code are available at https://github.com/UCSD-AI4H/COVID-Dialogue</div>

scholarly journals Prevalence of diagnosed diabetes and associated risk factors: Evidence from the large-scale surveys in India

2017 ◽  
Vol 05 (01) ◽  
pp. 028-036 ◽  
Author(s):  
Saddaf Akhtar ◽  
Preeti Dhillon
Keyword(s):  
Risk Factors ◽  
High Risk ◽  
Large Scale ◽  
Multinomial Logistic Regression ◽  
Local Level ◽  
Diabetes Prevalence ◽  
World Health ◽  
Undiagnosed Diabetes ◽  
Newly Diagnosed ◽  
Associated Risk Factors

Abstract Context: India has observed the most devastating increases in the burden of diabetes in the contemporary era. However, so far, the comparable prevalence of diabetes is only available for limited geography. Aims: The present paper provides comparable estimates of diabetes prevalence in states and districts of India and examines the associated risk factors with newly diagnosed and self-reported diabetes. Setting and Design: The study uses clinical, anthropometric, and biochemical data from District Level Household and Facility Survey (2012–2013) and Annual Health Survey (2014). Subjects and Methods: The paper analyses the information on glucose level of the blood sample and defines diabetes as per the World Health Organization (1999) criteria. It applies multinomial logistic regression to identify the risk factors of diabetes. Results: The study estimates 7% adults with diabetes in India, with a higher level in urban (9.8%) than in the rural area (5.7%), a higher proportion of males (7.1%) than females (6.8%). Widowed, older persons, and persons with high blood pressure have very high risk of both diagnosed and self-reported diabetes. Comparing to Hindus, Muslims and Christians have higher, and Sikhs have less risk of diabetes. Further, corresponding to general caste, scheduled castes, and other backward classes have a high risk of newly diagnosed but the lower risk of self-reported diabetes. Conclusions: The list of districts and states with alarming diabetes prevalence is the valuable information for further programs and research. A significant population with undiagnosed diabetes reflects an urgent need to strengthen the diagnostics at the local level and for those who need them most.

scholarly journals Factors Determining Periprocedural and Long-term Complications of High Risk Carotid Artery Stenting

2015 ◽  
Vol 42 (1) ◽  
pp. 48-54
Author(s):  
Vincent Dinculescu ◽  
Anne C.M. Ritter ◽  
Marlise P. dos Santos ◽  
Ravi M. Mohan ◽  
Betty A. Schwarz ◽  
...  
Keyword(s):  
Risk Factors ◽  
High Risk ◽  
Carotid Artery ◽  
Carotid Artery Stenting ◽  
Large Scale ◽  
High Risk Patient ◽  
Patient Characteristics ◽  
High Risk Patients ◽  
Risk Patients

ABSTRACTBackground and Purpose: Carotid artery stenting (CAS) has been, historically, an alternative to open endarterectomy (CEA) for stroke prevention in high risk patients with carotid atherosclerosis. We sought to determine the rates of periprocedural and long-term stroke or death and the risk factors for complications after CAS in our high risk patient population. Methods: Clinical and treatment variables of consecutive CAS procedures performed between 2002 and 2011 were analyzed. Using univariate and multivariate logistic regression analyses we examined how patient characteristics influenced outcomes and changes in modified Rankin Score (mRS). Results: In 152 patients, the composite total of periprocedural death, stroke, transient ischemic attack (TIA) and myocardial infarction (MI) rate was 3.95% (6/152). Chronic kidney disease (CKD) was strongly associated with periprocedural complications (p<0.001). Coronary artery disease/peripheral vascular disease (CAD/PVD) (p=0.03), dyslipidemia (p=0.02), CKD (p=0.01), and contralateral internal carotid artery stenosis (p=0.02) were non-modifiable risk factors for mRS increase. There were 25 deaths, 8 strokes, 11 TIAs, and 1 MI (mean follow-up 38.4 months, range 0-116 months). The presence of CAD/PVD (p=0.009) and dyslipidemia (p=0.002) were significantly associated with long-term complications. Conclusion: CAS was performed with low periprocedural complications in high-risk patients. Our rates compare very favorably to large-scale trials that have ideal patients. This data encourages the consideration of CAS in patients considered high risk for CEA and provides possible patient characteristics (CKD) to help with periprocedural risk stratification.

scholarly journals Sleeping under the waves: a longitudinal study across the contagion peaks of COVID-19 pandemic in Italy

2021 ◽  
Author(s):  
Federico Salfi ◽  
Aurora D’Atri ◽  
Daniela Tempesta ◽  
Michele Ferrara
Keyword(s):  
Risk Factors ◽  
High Risk ◽  
Sleep Quality ◽  
Sleep Disturbances ◽  
Large Scale ◽  
Well Being ◽  
Poor Sleep ◽  
Second Wave ◽  
Insomnia Symptoms ◽  
Risk Categories

AbstractAfter the March–April 2020 COVID-19 outbreak, a second contagion wave afflicted Europe in autumn. This study aimed to evaluate sleep health/patterns of Italians during this further challenging situation.A total of 2013 Italians longitudinally participated in a web-based survey during the two contagion peaks of the COVID-19 outbreak. We investigated the risk factors for sleep disturbances during the second wave, and we compared sleep quality and psychological well-being between the two assessments (March–April and November–December 2020). Female gender, low education, evening chronotype, being at high-risk for COVID-19 infection, reporting negative social or economic impact, and evening smartphone overuse predicted a higher risk of poor sleep and insomnia symptoms during the second wave. Advanced age, living with high-risk subjects for COVID-19 infection, and having a relative/friend infected with COVID-19 before the prior two weeks were risk categories for poor sleep quality. Living with children, having contracted COVID-19 before the prior two weeks, being pessimistic on the vaccine, and working in healthcare were risk factors for insomnia symptoms. The follow-up assessment highlighted reduced insomnia symptoms and anxiety. Nevertheless, we showed reduced sleep duration, higher daytime dysfunction and sleep medication use, and advanced sleep phase, confirming the alarming prevalence of poor sleepers (∼60%) and severe depression (∼20%) in a context of increased perceived stress.This study demonstrated a persistent impact of the COVID-19 pandemic on sleep and mental health. Large-scale interventions to counteract the chronicity and exacerbation of sleep and psychological disturbances are necessary, especially for the risk categories.

scholarly journals A Comparison of Cut Points for Measuring Risk Factors for Adolescent Substance Use and Antisocial Behaviors in the U.S. and Colombia

2021 ◽  
Vol 18 (2) ◽  
pp. 470
Author(s):  
Eric C. Brown ◽  
Pablo Montero-Zamora ◽  
Francisco Cardozo-Macías ◽  
María Fernanda Reyes-Rodríguez ◽  
John S. Briney ◽  
...  
Keyword(s):  
Risk Factors ◽  
Substance Use ◽  
High Risk ◽  
Large Scale ◽  
The United States ◽  
Adolescent Substance Use ◽  
Adolescent Substance ◽  
Cut Points ◽  
Essential Question ◽  
Significant Difference

As the identification and targeting of salient risk factors for adolescent substance use become more widely used globally, an essential question arises as to whether U.S.-based cut points in the distributions of these risk factors that identify “high” risk can be used validly in other countries as well. This study examined proportions of youth at “high” risk using different empirically derived cut points in the distributions of 18 measured risk factors. Data were obtained from large-scale samples of adolescents in Colombia and the United States. Results indicated that significant (p < 0.05) differences in the proportions of “high” risk youth were found in 38.9% of risk factors for 6th graders, 61.1% for 8th graders, and 66.6% for 10th graders. Colombian-based cut points for determining the proportion of Colombian youth at “high” risk were preferable to U.S.-based cut points in almost all comparisons that exhibited a significant difference. Our findings suggest that observed differences were related to the type of risk factor (e.g., drug specific vs. non-drug specific). Findings from this study demonstrate the need for collecting large-scale national data on risk factors for adolescent substance use and developing country-specific cut points based on the distributions of these measures to avoid misidentification of youth at “high” risk.

2258-PUB: Impact of Cardiovascular (CV) Risk Factors during the Initial 6 Months of the First Large-Scale Longitudinal Nationwide Study on Management and Real-World Outcomes of Diabetes in India (LANDMARC)

Diabetes ◽  
2020 ◽  
Vol 69 (Supplement 1) ◽  
pp. 2258-PUB
Author(s):  
ROMIK GHOSH ◽  
ASHOK K. DAS ◽  
SHASHANK JOSHI ◽  
AMBRISH MITHAL ◽  
K.M. PRASANNA KUMAR ◽  
...  
Keyword(s):  
Risk Factors ◽  
Real World ◽  
Large Scale ◽  
Nationwide Study

Prevention of Preterm Birth

2018 ◽  
Vol 11 (2) ◽  
pp. 95-104
Author(s):  
Ivan D. Ivanov ◽  
Stefan A. Buzalov ◽  
Nadezhda H. Hinkova
Keyword(s):  
Risk Factors ◽  
Preterm Birth ◽  
High Risk ◽  
Medical Science ◽  
Secondary Prophylaxis ◽  
High Risk Group ◽  
Screening Methods ◽  
Perinatal Complications ◽  
Social Significance ◽  
Neonatal Deaths

Summary Preterm birth (PTB) is a worldwide problem with great social significance because it is a leading cause of perinatal complications and perinatal mortality. PTB is responsible for more than a half of neonatal deaths. The rate of preterm delivery varies between 5-18% worldwide and has not decreased in recent years, regardless of the development of medical science. One of the leading causes for that is the failure to identify the high-risk group in prenatal care. PTB is a heterogeneous syndrome in which many different factors interfere at different levels of the pathogenesis of the initiation of delivery, finally resulting in delivery before 37 weeks of gestation (wg). The various specificities of risk factors and the unclear mechanism of initiation of labour make it difficult to elaborate standard, unified and effective screening to diagnose pregnant women at high-risk for PTB correctly. Furthermore, they make primary and secondary prophylaxis less effective and render diagnostic and therapeutic measures ineffective and inappropriate. Reliable and accessible screening methods are necessary for antenatal care, and risk factors for PTB should be studied and clarified in search of useful tools to solve issues of risk pregnancies to decrease PTB rates and associated complications.

Sign in / Sign up

Export Citation Format

Share Document