Dataset Information

Classifying Chinese Questions Related to Health Care Posted by Consumers Via the Internet.

ABSTRACT:

Background

In question answering (QA) system development, question classification is crucial for identifying information needs and improving the accuracy of returned answers. Although the questions are domain-specific, they are asked by non-professionals, making the question classification task more challenging.

Objective

This study aimed to classify health care-related questions posted by the general public (Chinese speakers) on the Internet.

Methods

A topic-based classification schema for health-related questions was built by manually annotating randomly selected questions. The Kappa statistic was used to measure the interrater reliability of multiple annotation results. Using the above corpus, we developed a machine-learning method to automatically classify these questions into one of the following six classes: Condition Management, Healthy Lifestyle, Diagnosis, Health Provider Choice, Treatment, and Epidemiology.

Results

The consumer health question schema was developed with a four-hierarchical-level of specificity, comprising 48 quaternary categories and 35 annotation rules. The 2000 sample questions were coded with 2000 major codes and 607 minor codes. Using natural language processing techniques, we expressed the Chinese questions as a set of lexical, grammatical, and semantic features. Furthermore, the effective features were selected to improve the question classification performance. From the 6-category classification results, we achieved an average precision of 91.41%, recall of 89.62%, and F₁ score of 90.24%.

Conclusions

In this study, we developed an automatic method to classify questions related to Chinese health care posted by the general public. It enables Artificial Intelligence (AI) agents to understand Internet users' information needs on health care.

SUBMITTER: Guo H

PROVIDER: S-EPMC5497072 | biostudies-literature | 2017 Jun

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Classifying Chinese Questions Related to Health Care Posted by Consumers Via the Internet.

Guo Haihong H Na Xu X Hou Li L Li Jiao J

Journal of medical Internet research 20170620 6

<h4>Background</h4>In question answering (QA) system development, question classification is crucial for identifying information needs and improving the accuracy of returned answers. Although the questions are domain-specific, they are asked by non-professionals, making the question classification task more challenging.<h4>Objective</h4>This study aimed to classify health care-related questions posted by the general public (Chinese speakers) on the Internet.<h4>Methods</h4>A topic-based classifi ...[more]

PMID: 28634156

Similar Datasets

Project description:BackgroundThe advancement of wearable devices and growing demand of consumers to monitor their own health have influenced the medical industry. Health care providers, insurers, and global technology companies intend to develop more wearable devices incorporating medical technology and to target consumers worldwide. However, acceptance of these devices varies considerably among consumers of different cultural backgrounds. Consumer willingness to use health care wearables is influenced by multiple factors that are of varying importance in various cultures. However, there is insufficient knowledge of the extent to which social and cultural factors affect wearable technology acceptance in health care.ObjectiveThe aims of this study were to examine the influential factors on the intention to adopt health care wearables, and the differences in the underlying motives and usage barriers between Chinese and Swiss consumers.MethodsA new model for acceptance of health care wearables was conceptualized by incorporating predictors of different theories such as technology acceptance, health behavior, and privacy calculus based on an existing framework. To verify the model, a web-based survey in both the Chinese and German languages was conducted in China and Switzerland, resulting in 201 valid Chinese and 110 valid Swiss respondents. A multigroup partial least squares path analysis was applied to the survey data.ResultsPerformance expectancy (β=.361, P<.001), social influence (β=.475, P<.001), and hedonic motivation (β=.111, P=.01) all positively affected the behavioral intention of consumers to adopt wearables, whereas effort expectancy, functional congruence, health consciousness, and perceived privacy risk did not demonstrate a significant impact on behavioral intention. The group-specific path coefficients indicated health consciousness (β=.150, P=.01) as a factor positively affecting only the behavior intention of the Chinese respondents, whereas the factors affecting only the behavioral intention of the Swiss respondents proved to be effort expectancy (β=.165, P=.02) and hedonic motivation (β=.212, P=.02). Performance expectancy asserted more of an influence on the behavioral intention of the Swiss (β=.426, P<.001) than the Chinese (β=.271, P<.001) respondents, whereas social influence had a greater influence on the behavioral intention of the Chinese (β=.321, P<.001) than the Swiss (β=.217, P=.004) respondents. Overall, the Chinese consumers displayed considerably higher behavioral intention (P<.001) than the Swiss. These discrepancies are explained by differences in national culture.ConclusionsThis is one of the first studies to investigate consumers' intention to adopt wearables from a cross-cultural perspective. This provides a theoretical and methodological foundation for future research, as well as practical implications for global vendors and insurers developing and promoting health care wearables with appropriate features in different countries. The testimonials and support by physicians, evidence of measurement accuracy, and easy handling of health care wearables would be useful in promoting the acceptance of wearables in Switzerland. The opinions of in-group members, involvement of employers, and multifunctional apps providing credible health care advice and solutions in cooperation with health care institutions would increase acceptance among the Chinese.

Project description:BackgroundThe global COVID-19 pandemic has been widely regarded as a catalyst for adopting internet health care technology (IHT) in China. IHT consists of new health care technologies that are shaping health services and medical consultations. Health care professionals play a substantial role in the adoption of any IHT, but the consequences of doing so can often be challenging, particularly when employee burnout is prevalent. Few studies have explored whether employee burnout influences the adoption intention of IHT in health care professionals.ObjectiveThis study aims to explain the determinants influencing the adoption of IHT from the perspective of health care professionals. To do so, the study extends the value-based adoption model (VAM) with consideration for employee burnout as a determining factor.MethodsA cross-sectional web-based survey using a sample of 12,031 health care professionals selected through multistage cluster sampling from 3 provinces in mainland China was conducted. The hypotheses of our research model were developed based on the VAM and employee burnout theory. Structural equation modeling was then used to test the research hypotheses.ResultsThe results indicate that perceived usefulness, perceived enjoyment, and perceived complexity positively correlate with perceived value (β=.131, P=.01; β=.638, P<.001; β=.198, P<.001, respectively). Perceived value had a positive direct effect on adoption intention (β=.725, P<.001), perceived risk negatively correlated with perceived value (β=-.083, P<.001), and perceived value negatively correlated with employee burnout (β=-.308, P<.001). In addition, employee burnout was negatively related to adoption intention (β=-.170, P<.001) and mediated the relationship between perceived value and adoption intention (β=.052, P<.001).ConclusionsPerceived value, perceived enjoyment, and employee burnout were the most important determinants of IHT adoption intention by health care professionals. In addition, while employee burnout was negatively related to adoption intention, perceived value inhibited employee burnout. Therefore, this study finds that it is necessary to develop strategies to improve the perceived value and reduce employee burnout, which will benefit the promotion of the adoption intention of IHT in health care professionals. This study supports the use of the VAM and employee burnout in explaining health care professionals' adoption intention regarding IHT.

Project description:BACKGROUND:As the US health care system is embracing data-driven care, personal health information (PHI) has become a valuable resource for various health care stakeholders. In particularly, health consumers are expected to autonomously manage and share PHI with their health care partners. To date, there have been mixed views on the factors influencing individuals' health data-sharing behaviors. OBJECTIVE:This study aimed to identify a key factor to better understand health information sharing behavior from a health consumer's perspective. We focused on daily settings, wherein health data-sharing behavior becomes a part of individuals' daily information management activities. Considering the similarity between health and finance information management, we explicitly examined whether health consumers' daily habit of similar data sharing from the financial domain affects their PHI-sharing behaviors in various scenarios. METHODS:A Web-based survey was administered to US health consumers who have access to and experience in using the internet. We collected individual health consumers' intention to share PHI under varying contexts, habit of financial information management (operationalized as internet banking [IB] use in this paper), and the demographic information from the cross-sectional Web-based survey. To isolate the effect of daily IB on PHI-sharing behaviors in everyday contexts, propensity score matching was used to estimate the average treatment effect (ATE) and average treatment effect on the treated (ATET) regarding IB use. We balanced the treatment and control groups using caliper matching based on the observed confounding variables (ie, gender, income, health status, and access to primary care provider), all of which resulted in a minimal level of bias between unmatched and matched samples (bias <5%). RESULTS:A total of 339 responses were obtained from a cross-sectional Web-based survey. The ATET results showed that in terms of sharing contents, those who used IB daily were more likely to share general information (P=.01), current information (P=.003), and entire data (P=.04). Regarding occasions for sharing occasions, IB users were prone to share their information in all cases (P=.02). With regard to sharing recipients, daily IB users were more willing to share their personal health data with stakeholders who were not directly involved in their care, such as health administrators (P=.05). These results were qualitatively similar to the ATE results. CONCLUSIONS:This study examined whether daily management of similar information (ie, personal financial information) changes health consumers' PHI-sharing behavior under varying sharing conditions. We demonstrated that daily financial information management can encourage health information sharing to a much broader extent, in several instances, and with many stakeholders. We call for more attention to this unobserved daily habit driven by the use of various nonhealth technologies, all of which can implicitly affect patterns and the extent of individuals' PHI-sharing behaviors.

Project description:ImportanceLittle is known about how discrimination in health care relates to inequities in hospital-based care because of limitations in the ability to measure discrimination. Consumer reviews offer a novel source of data to capture experiences of discrimination in health care settings.ObjectiveTo examine how health care consumers perceive and report discrimination through public consumer reviews.Design, setting, and participantsThis qualitative study assessed Yelp online reviews from January 1, 2011, to December 31, 2020, of 100 randomly selected acute care hospitals in the US. Word filtering was used to identify reviews potentially related to discrimination by using keywords abstracted from the Everyday Discrimination Scale, a commonly used questionnaire to measure discrimination. A codebook was developed through a modified grounded theory and qualitative content analysis approach to categorize recurrent themes of discrimination, which was then applied to the hospital reviews.ExposuresReported experiences of discrimination within a health care setting.Main outcomes and measuresPerceptions of how discrimination in health care is experienced and reported by consumers.ResultsA total of 10 535 reviews were collected. Reviews were filtered by words commonly associated with discriminatory experiences, which identified 2986 reviews potentially related to discrimination. Using the codebook, the team manually identified 182 reviews that described at least 1 instance of discrimination. Acts of discrimination were categorized by actors of discrimination (individual vs institution), setting (clinical vs nonclinical), and directionality (whether consumers expressed discriminatory beliefs toward health care staff). A total of 53 reviews (29.1%) were coded as examples of institutional racism; 89 reviews (48.9%) mentioned acts of discrimination that occurred in clinical spaces as consumers were waiting for or actively receiving care; 25 reviews (13.7%) mentioned acts of discrimination that occurred in nonclinical spaces, such as lobbies; and 66 reviews (36.3%) documented discrimination by the consumer directed at the health care workforce. Acts of discrimination are described through 6 recurrent themes, including acts of commission, omission, unprofessionalism, disrespect, stereotyping, and dehumanizing.Conclusions and relevanceIn this qualitative study, consumer reviews were found to highlight recurrent patterns of discrimination within health care settings. Applying quality improvement tools, such as the Plan-Do-Study-Act cycle, to this source of data and this study's findings may help inform assessments and initiatives directed at reducing discrimination within the health care setting.

Project description:BACKGROUND:Snapchat has seen one of the most rapid, and unprecedented, growths in the history of social networking sites and social media with 3 billion Snapchats sent daily. In 2015, Snapchat introduced a new feature, Snapchat Discover, providing a unique way for publishers, such as magazines, to connect their content to Snapchat users. OBJECTIVE:This study aimed to evaluate qualitatively the health-related content distributed among male-focused and female-focused Discover channels and to determine whether differences exist between the content posted to these channels. METHODS:Magazine Discover channels with male and female target audiences were identified based on the magazine's claimed audience and a search of Snapchat Discover's magazine publishers, resulting in the selection of two male-focused and two female-focused channels. Stories were collected daily from each of the selected channels during a 4-week period. Using the constant comparative method, 406 Discover stories were collected and analyzed. RESULTS:Differences in health content coverage existed between male- and female-focused channels. General health stories from male channels comprised 7.5% (10/134) of total stories compared with 22.8% (62/272) for female channels. Sexual health stories from male channels comprised 3.0% (4/134) of total stories compared with 18.8% (51/272) for female channels. Moreover, female-focused channels' content was more comprehensive. Female audiences were portrayed as being health information seekers, concerned with sexual health and male satisfaction, primarily responsible for contraception and pregnancy prevention, and less informed about sex. Male audiences were portrayed as being less likely to seek health information, obsessed with and driven by sex, and less concerned with sexual health. CONCLUSIONS:Understanding the content shared to social media is important, especially when considering the implications content may have for behavior. In terms of content, these findings suggest Discover channels appear to promote gender stereotypes and norms for health and sexual health through the information posted.

Project description:BackgroundWestern accredited medical universities can offer graduate-level academic courses to health care workers (HCWs) in resource-limited settings through the Internet. It is not known whether HCWs are interested in these online courses, whether they can perform as well as matriculated students, or whether such courses are educationally or practically relevant.Methods and findingsIn 2011, the University of Washington (UW) Schools of Medicine and Nursing offered the graduate course, "Clinical Management of HIV", to HCWs that included a demographic survey, knowledge assessment, and course evaluation. UW faculty delivered HIV clinical topics through ten 2-hour weekly sessions from the perspectives of practicing HIV medicine in developed and developing settings. HCWs viewed lectures through Adobe Acrobat Connect Pro (Adobe Systems, San Jose, CA), and completed online homework on HIV Web Study (http://depts.washington.edu/hivaids/) and online quizzes. HCWs, who met the same passing requirements as UW students by attending 80% lectures, completing ?90% homework, and achieving a cumulative ?70% grade on quizzes, were awarded a certificate. 369 HCWs at 33 sites in 21 countries joined the course in 2011, a >15-fold increase since the course was first offered in 2007. The majority of HCWs came from Africa (72%), and most were physicians (41%), nurses (22%), or midlevel practitioners (20%). 298 HCWs (81%) passed all requirements and earned a certificate. In a paired analysis of pre- and post-course HIV knowledge assessments, 56% of HCWs improved their post-course score (p<0.0001) with 27% improving by at least 30%. In the course evaluation, most HCWs rated the course as excellent (53%) or very good (39%).ConclusionsThis online HIV course demonstrated that opening a Western graduate medical and nursing curriculum to HCWs in resource-limited settings is feasible, popular, and valuable, and may address logistic and economic barriers to the provision of high quality education in these settings.

Dataset Information

Classifying Chinese Questions Related to Health Care Posted by Consumers Via the Internet.

Background

Objective

Methods

Results

Conclusions

Publications

Classifying Chinese Questions Related to Health Care Posted by Consumers Via the Internet.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets