Skip to main content
main-content

Über dieses Buch

This book constitutes the refereed proceedings of the International Conference for Smart Health, ICSH 2014, held in Beijing, China, in July 2014. The 21 papers presented together with 4 extended abstracts were carefully reviewed and selected from numerous submissions. The papers are organized in topical sections on information sharing, integrating and extraction; health data analysis and management; clinical and medical data mining; and clinical practice and medical monitoring.

Inhaltsverzeichnis

Frontmatter

Information Sharing, Integrating and Extraction

Emoticon Analysis for Chinese Health and Fitness Topics

Abstract
An emoticon is a metacommunicative pictorial representation of facial expressions, which serves to convey information about the sender’s emotional state. To complement non-verbal communication, emoticons are frequently used in Chinese online social media, especially in discussions of health and fitness topics. However, limited research has been done to effectively analyze emoticons in a Chinese context. In this study, we developed an emoticon analysis system to extract emoticons from Chinese text and classify them into one of 7 affect categories. The system is based on a kinesics model which divides emoticons into semantic areas (eyes, mouths, etc.), with an improvement for adaption in the Chinese context. Empirical tests were conducted to evaluate the effectiveness of the proposed system in extracting and classifying emoticons, based on a corpus of more than one million sentences of Chinese health- and fitness-related online messages. Results showed the system to be effective in detecting and extracting emoticons from text, and in interpreting the emotion conveyed by emoticons.
Shuo Yu, Hongyi Zhu, Shan Jiang, Hsinchun Chen

Diabetes-Related Topic Detection in Chinese Health Websites Using Deep Learning

Abstract
With 98.4 million people diagnosed with diabetes in China, most of the Chinese health websites provide diabetes related news and articles in diabetes subsection for patients. However, most of the articles are uncategorized and without a clear topic or theme, resulting in time consuming information seeking experience. To address this issue, we propose an advanced deep learning approach to detect topics for diabetes related articles from health websites. Our research framework for topic detection on diabetes related articles in Chinese is the first one to incorporate deep learning in topic detection in Chinese. It can identify topics of diabetes articles with high performance and potentially assist health information seeking. To evaluate our framework, experiment is conducted on a test bed of 12,000 articles. The results showed the framework achieved an accuracy of 70% in detecting topics and significantly outperformed the SVM based approach.
Xinhuan Chen, Yong Zhang, Chunxiao Xing, Xiao Liu, Hsinchun Chen

Identifying Adverse Drug Events from Health Social Media: A Case Study on Heart Disease Discussion Forums

Abstract
Health social media sites have emerged as major platforms for discussions of treatments and drug side effects, making them a promising source for listening to patients’ voices in adverse drug event reporting. However, extracting patient adverse drug event reports from social media continues to be a challenge in health informatics research. To utilize the fertile health social media data for drug safety research, we develop advanced information extraction techniques for identifying adverse drug events in health social media. A case study is conducted on a heart disease discussion forum to evaluate the performance. Our approach achieves an f-measure of 82% in the recognition of medical events and treatments, an f-measure of 69% for identifying adverse drug events and an f-measure of 90% in patient report extraction. Analysis on the extracted adverse drug events suggests that health social media can provide supplemental information for adverse drug events and drug interactions. It provides a less biased insight into the distribution of adverse events among heart disease population compared to data from a drug regulatory agency.
Xiao Liu, Jing Liu, Hsinchun Chen

An Empirical Analysis on Communications about Electronic Nicotine Delivery Systems (ENDS) in Chinese Social Media

Abstract
China, with a smoking population of over 350 million, is the largest potential market for electronic nicotine delivery systems (ENDS). The importance of understanding how ENDS are promoted and discussed in China cannot be overstated. However, related research is sparse. This study aims to explore the nature and extent of discussions around ENDS in Chinese social media, which have the power to influence a massive audience. We collected the data from Sina Weibo, which is one of the most popular Chinese microblogging sites. The dataset, which consisted of 999 messages, was analyzed in terms of polarities, genres and discussion topics. Statistical test and regression analysis were used to explore whether those features of messages will affect the message popularity, which was measured by repost number and comment number. The results of our study showed that 1) The majority of the messages were Pro-ENDS; 2) The number of comments received by messages of different genes varies significantly; 3) Whether a message contains specific topics or not will affect the comment number.
Kainan Cui, Xiaolong Zheng, Daniel Zeng, Scott Leischow

Analyzing Spatio-temporal Patterns of Online Public Attentions in Emergency Events: A Case Study of 2009 H1N1 Influenza Outbreak in China

Abstract
Understanding the public attention and perception towards epidemics is critical for public health response. However, the research question concerning the spatio-temporal patterns of public attention and the interactions with media attention and severity of epidemic is still not well studied. Aim to fill this research gap, we chose the H1N1 influenza outbreak in the mainland of China in 2009 as case to study the spatio-temporal patterns of public attention, and their correlations with media attention and severity of epidemic. The results of this paper indicate that public attention and media attention had high correlation from both temporal and spatial perspectives, which can provide us significant insights to understand the collective behavior of massive online users during emergency events.
Kainan Cui, Xiaolong Zheng, Zhu Zhang, Daniel Zeng

HealthQA: A Chinese QA Summary System for Smart Health

Abstract
Although online health expert QA services can provide high quality information for health consumers, there is no Chinese question answering system built on knowledge from existing expert answers, leading to duplicated efforts of medical experts and reduced efficiency. To address this issue, we develop a Chinese QA system for smart health (HealthQA), which provides timely, automatic and valuable QA service. Our HealthQA collects diabetes expert question answer data from three major QA websites in China. We develop a hierarchical clustering method to group similar questions and answers, an extended similarity evaluation algorithm for retrieving relevant answers and a ranking based summarization for representing the answer. ROUGE and manual tests show that our system significantly outperforms the search engine.
Yanshen Yin, Yong Zhang, Xiao Liu, Yan Zhang, Chunxiao Xing, Hsinchun Chen

DiabeticLink: An Integrated and Intelligent Cyber-Enabled Health Social Platform for Diabetic Patients

Abstract
Given the demand of patient-centered care and limited healthcare resources, we believe that the community of diabetic patients is in need of an integrated cyber-enabled patient empowerment and decision support tool to promote diabetes prevention and self-management. Most existing tools are scattered and focused on solving a specific problem from a single angle. DiabeticLink offers an integrated and intelligent web-based platform that enables patient social connectivity and self-management, and offers behavior change aids using advanced health analytics techniques. DiabeticLink released a beta version in Taiwan in July 2013. The next versions of the DiabeticLink system are under active development and will be launched in the U.S., Denmark, and China in 2014. We describe the system functionalities and discuss the user testing and lessons learned from real-world experience. We also describe plans for future development.
Joshua Chuang, Owen Hsiao, Pei-Lin Wu, Jean Chen, Xiao Liu, Haily De La Cruz, Shu-Hsing Li, Hsinchun Chen

Health Data Analysis and Management

Collaborative Friendship Networks in Online Healthcare Communities: An Exponential Random Graph Model Analysis

Abstract
Health 2.0 provides patients an unprecedented way to connect with each other online. However, less attention has been paid to how patient collaborative friendships form in online healthcare communities. This study examines the relationship between collaborative friendship formation and patients’ characteristics. Results from Exponential Random Graph Model (ERGM) analysis indicate that gender homophily doesn’t appear in CFNs, while health homophily such as treatments homophily and health-status homophily increases the likelihood of collaborative friendship formation. This study provides insights for improving website design to help foster close relationship among patients and deepen levels of engagement.
Xiaolong Song, Shan Jiang, Xianbin Yan, Hsinchun Chen

The Impact of Alcohol Intake on Human Beings Health in China

Abstract
Drinking is one of the main causes of disease burden. In this paper, we tried to find the effect of alcohol intake on health state using the China Health and Nutrition Survey data in a model of regression. We initially used the self assessed health, which was widely used in related studies, in the database as the health variable but obtained controversial result. After that, to test the result we compared the drinking status between participants who are longevous and who are not and found the initial result is unreasonable. We then used the life span as our health variable which produced sensible result as expected. Linear regression and ordinal logistic regression were used to quantify the effect of alcohol intake on health. The results for our sample suggest two conclusions: 1) drinkers are overconfident in their health statuses; and 2) alcohol intake does shorten the life span on average.
Wenbin Wang, Kaiye Gao, Qing Wei

Social Support and User Engagement in Online Health Communities

Abstract
Online health communities (OHCs) have become a major source of social support for people with health problems. Members of OHCs interact online with those who face similar problems and are involved in different types of social supports, such as informational support, emotional support and companionship. Using a case study of an OHC among breast cancer survivors, we first use machine learning techniques to reveal the types of social support embedded in each post from an OHC. Then we generate each user’s contribution profile by aggregating the user’s involvement in various types of social support and reveal that users play different roles in the OHC. By comparing online activities for users with different roles and conducting survival analysis on users’ time span of online activities, we illustrate that users’ levels of engagement in an OHC are related to various types of social support in different ways.
Xi Wang, Kang Zhao, Nick Street

Doctor’s Effort Influence on Online Reputation and Popularity

Abstract
This research examines why doctors participate in online activities and how they fully use online effort. We identify two dimensions of effort, i.e., online healthcare community function breadth (OHCFB), the number of different online functions used, and online healthcare community function depth (OHCFD), the degree of involvement with use of special online functions. We examine the effect of these two dimensions of online effort on doctor’s online reputation and popularity and the impact of interaction between OHCFB and OHCFD. The study has three major findings: a) doctor’s effort can increase their reputation and popularity, b) doctors with a low title have the potential to overcome shortcomings to increase popularity if they give special attention to a few select functions; however, doctors with a high title should try to expand the number of functions, and c) the interaction effect between OHCFB and OHCFD manifests differently across reputation and popularity.
Xiaoxiao Liu, Xitong Guo, Hong Wu, Doug Vogel

Extended Abstract: Study on the Design for Consumer Health Knowledge Organization System in China

Abstract
Objective: In order to bridge the “knowledge gap” between the technical terms used by healthcare professionals and the health terms by consumers, this paper designs a knowledge representation method to organize various health concepts and terms which can be understanding by general public in China.
Hou Li

Trend and Network Analysis of Common Eligibility Features for Cancer Trials in ClinicalTrials.gov

Abstract
ClinicalTrials.gov has been archiving clinical trials since 1999, with > 165,000 trials at present. It is a valuable but relatively untapped resource for understanding trial design patterns and acquiring reusable trial design knowledge. We extracted common eligibility features using an unsupervised tag-mining method and mined their temporal usage patterns in clinical trials on various cancers. We then employed trend and network analysis to investigate two questions: (1) what eligibility features are frequently used to select patients for clinical trials within one cancer or across multiple cancers; and (2) what are the trends in eligibility feature adoption or discontinuation across cancer research domains? Our results showed that each cancer domain reuses a small set of eligibility features frequently for selecting cancer trial patients and some features are shared across different cancers, with value range adjustments for numerical measures. We discuss the implications for facilitating community-based clinical research knowledge sharing and reuse.
Chunhua Weng, Anil Yaman, Kuo Lin, Zhe He

A Conceptual Framework of Information Analysis and Modelling for E-health

Abstract
E-health (or e-healthcare) aims at applying modern information and telecommunication technologies in the healthcare sector. Recently, more and more research attention has been paid to this area, from clinic data analysis to patient record management. If we say the e-health has been mainly dealing with patient data analysis and various disease diagnosis at its early time, nowadays various data sources, such as social network, physician and patients’ blogs, and various monitoring data (like sensors attached to PD patients at home) are widely used for aid of better disease diagnosis through gathering richer experts’ knowledge and expertise, richer data (from patients, diseases, diagnosis, medical experiments, etc.) for analysis, and quicker (or better timely) access to various resources (e.g. telemedicine) pervasively.
William W. Song

Clinical and Medical Data Mining

Extended Abstract: Combining Statistical Analysis and Markov Models with Public Health Data to Infer Age-Specific Background Mortality Rates for Hepatitis C Infection in the U.S.

Abstract
Chronic hepatitis C (HCV) is a significant public health problem affecting 2.7-3.9 million Americans. Quantifying mortality rates of HCV-infected individuals permits more accurate estimates of the potential benefits of HCV screening and treatment. With 5% of older Americans infected with HCV, cost-effectiveness analyses of expanded HCV screening and treatment require methods to appropriately quantify differential mortality risks. No single study contains data needed to estimate subgroup-specific prevalence of HCV, risk factor status, and mortality risks. We developed a combined modeling approach to infer risk-group-specific mortality rates for chronically HCV-infected U.S. adults. We incorporated estimates from public health data into a Markov model to infer the age-, sex-, race-, risk-, and HCV infection status-specific mortality rates that best fit the overall age-specific population mortality rates.
Shan Liu, Lauren E. Cipriano, Jeremy D. Goldhaber-Fiebert

Apply Autocorrelation and Forward Difference to Measure Vital Signs Using Ordinary Camera

Abstract
Measuring heart rate by portable equipments becomes more and more popular. Current methods such as wavelet, fast fourier transform, peak detection, have been used to analyze heart rate. However, in some cases these methods are ineffective. For example, as a denoising tool, wavelet is not necessary in a few cases. One of the main challenges is determining an effective size of sliding window for heart rate detection when using peak detection. In addition, the time complexity of fast fourier transform is large which can increase the processing time that is not desirable for real-time heart rate detection systems. In this paper, we introduce autocorrelation and forward difference to count heart rate based on the features of cardiac cycle. The results show that our method is good enough so that it can be applied to non-invasive health state detection. And the time complexity of our method is satisfactory.
Letian Sun, Li Liu, Ye Wei, Jun Zhong, Dashi Luo, Ming Liu, Hamed Monkaresi

Visual Analysis for Type 2 Diabetes Mellitus – Based on Electronic Medical Records

Abstract
A multidimensional-scaling approach is proposed to analyze the main symptoms of T2DM. Based on 200 Type 2 diabetes patients’ electronic medical records, the terms which were used to described symptoms in the records and their co-occurring query terms were analyzed. A distanced-based similarity measure was used to calculate the proximity of terms to one and another based on their co-occurrences in the 200 medical records. After the calculation of the distance between each two keywords, a visual clustering of groups of terms was conducted. Each terms distribution within each visual configuration showed the most common symptoms of Type 2 diabetes such as Foam in Urine, Intermittent Dizziness, Hyperlipemia, Feeble, Diuresis etc; however it also showed some hidden relations behind our cognition.
Xi Meng, Ji-Jiang Yang

A Preliminary Variable Selection Based Regression Analysis for Predicting Patient Satisfaction on Physician-Patient Cancer Prognosis Communication

Abstract
We explore the use of variable selection methods to deal with high correlations among predicative variables (e.g., physician’s voice tone and language certainty) for examining physician communication associated with prognosis discussion with cancer patients. Our main method is principal component analysis. The comparative results show its benefit in predicting patient satisfaction on the prognosis communication. This preliminary regression analysis is expected to offer insights into patient-centered communication strategy design, especially for cancer prognosis communication with end-stage patients.
Shuai Fang, Wenting Shi, Nan Kong, Cleveland G. Shields

Clinical Practice and Medical Monitoring

Intelli-food: Cyberinfrastructure for Real-Time Outbreak Source Detection and Rapid Response

Abstract
Foodborne diseases cause an estimated 48 million illnesses each year in the United States, including 9.4 million caused by known pathogens. Real time detection of cases and outbreak sources are important epidemic intelligence services that can decrease morbidity and mortality of foodborne illnesses, and allow optimal response to identify the causal pathways leading to contamination. For most outbreaks associated with fresh produce items, outbreak source detection typically occurs after the contaminated produce items have been consumed and are no longer in the marketplace. We developed a probabilistic model for real time outbreak source detection, prediction of outbreaks, and contamination-prone area mapping with the aim of developing a cyber-infrastructure to support this activity. The models inputs include environmental, trade and epidemiological dynamics. Because effective distance reliably predicts disease arrival times we estimate the distance of outbreak sources from spatio-temporal patterns of foodborne outbreaks. As a case study we consider the 2013 Cyclospora outbreaks in the USA that were related to contaminated fresh produce (cilantro and fresh salad mix) from Mexico. We are able to match case distributions related to both food commodities and determine their outbreak sources with an average accuracy of 0.93. Assuming a similar pattern of contamination for 2014, outbreak patterns can be similar or worse with an unchanged food trade that is likely. The study aims to provide a methodological framework to evaluate environmentally sensitive food contamination and assess interdependencies of socio-environmental factors causing contamination. We emphasize the linkage of patterns and processes, the positive role of uncertainty, and challenge the belief that information about the whole food supply chain is needed for traceback analysis to be useful for identifying likely sources. Our specific prediction strongly emphasizes the need for real-time surveillance to identify and respond to this pending outbreak.
Matteo Convertino, Craig Hedberg

A Case Study in Healthcare Informatics: A Telemedicine Framework for Automated Parkinson’s Disease Symptom Assessment

Abstract
This paper reports the development and evaluation of a mobile-based telemedicine framework for enabling remote monitoring of Parkinson’s disease (PD) symptoms. The system consists of different measurement devices for remote collection, processing and presentation of symptom data of advanced PD patients. Different numerical analysis techniques were applied on the raw symptom data to extract clinically symptom information which in turn were then used in a machine learning process to be mapped to the standard clinician-based measures. The methods for quantitative and automatic assessment of symptoms were then evaluated for their clinimetric properties such as validity, reliability and sensitivity to change. Results from several studies indicate that the methods had good metrics suggesting that they are appropriate to quantitatively and objectively assess the severity of motor impairments of PD patients.
Taha Khan, Mevludin Memedi, William Song, Jerker Westin

A Prototype Mobile Virtual Assistant for Semantic-Based Vaccine Information Retrieval

Abstract
Since the early 90s, healthcare providers have been mandated to provide VIS (Vaccine Information Statement) from the Centers for Disease Control and Prevention (CDC) to parents and patients before their children or themselves receive any vaccination uptake. Despite the initiative, there exist issues of patients not receiving a comprehensive understanding about the vaccines and some evidence of doubt of the safety of vaccines. In addition, a significant number of patients find vaccine information on the Internet, which may inevitably influence perceptions of vaccines. This paper introduces Vaccine Helmsman, an initial prototype of a mobile client that allows for natural language querying of semantically-driven knowledge-base of vaccine information for patients. With the ubiquitous impact of the web and mobile devices in the hands of many patients, this would allow for contextual and instant access to vaccine information, and therefore enhance vaccine literacy.
Muhammad Amith, Cui Tao

Supportive Glucose Sensing Mobile Application to Improve the Accuracy of Continuous Glucose Monitors

Abstract
An insulin pump can be programmed to continuously deliver accurate amounts of insulin to diabetic patients. A continuous glucose monitor (CGM) which provides continuous patient glucose levels, needs to be calibrated at least every 6 hours. This paper provides an overview of the software and hardware requirements to increase the calibration duration with a high level of accuracy in an open source Artificial Pancreas platform. On the software level, it uses a smartphone camera to capture the food intake, and a smartphone sensor and positioning system to capture the patient movements.  The system maps three months’ worth of data points to the actual glucose level generated by the CGM. It then generates the probability of the estimated insulin needed based on the recorded movements and food intake activities for the patient. The logged data is used as the training data set. Using Bayes’ analysis, the generated probability that is based on the patient activities is used as posterior probabilities to the CGM results, which generates a more accurate estimation of the glucose level. On the hardware level, the paper presents a Universal Remote Control and its associated protocol to connect the smartphone with the CGM for information retrieval and with the insulin pump for information dissemination. The information is sent to the insulin pump using a Field Programmable Gate Array (FPGA). For communication, there are two kinds of message frames: Dosage Delivery Frame (DDF) and Acknowledgement frame (ACKF) with a secure layer of encryption.
Ahmed Gomaa, Chaogui Zhang, Muhammad Hasan, Mary Beth Roche, Shaun Hynes

A Control Study on the Effects of HRV Biofeedback Therapy in Patients with Post-Stroke Depression

Abstract
The post-stroke is often associated with emotional disorders, among which post-stroke depression (PSD) has a high incidence. We applied Heart Rate Variability (HRV) biofeedback to train PSD patients by a prospective randomized control study. The purpose of this study was to investigate the effectiveness of the HRV biofeedback on stroke patients’ emotional improvement, autonomic nerve function and prognostic implications. In the feedback group, the patients had learned to breathe at the resonant frequency to increase their low frequency (LF) as well as adjust their respiration to synchronize with heart rate fluctuations. Our findings suggest that the HRV biofeedback may be a valid treatment especially on the improvement of depression levels and sleep disturbance in PSD patients.
Xin Li, Tong Zhang, Luping Song, Yong Zhang, Chunxiao Xing, Hsinchun Chen

Optimal ST-Elevation Myocardial Infarction System by Regional Cooperative Emergency Care Based on the Internet of Things

Abstract
Prompt reperfusion treatment significantly reduces mortality and morbidity in patients with ST-elevation myocardial infarction (STEMI). In this article, the current status and problems of STEMI patient emergency care were studied and key influence factors were found. A regional cooperative emergency care system for STEMI patients were established based on the internet of things. An expedited pre-hospital diagnosis and transfer pathway was developed, as a result of which, a shorter time from symptom onset to reperfusion was achieved and the outcome for patients was improved.
Hao Chen, Dingcheng Xiang, Weiyi Qin, Minwei Zhou, Ji-Jiang Yang, Qiang Gao, Jian Liu

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise