Topic Modeling of Social Networking Service Data on Occupational Accidents in Korea: Latent Dirichlet Allocation Analysis

Harnessing Social Media Data to Understand Information Needs About Kidney Diseases and Emotional Experiences With Disease Management: Topic and Sentiment Analysis

BACKGROUND

Kidney diseases encompass a variety of conditions, including chronic kidney disease, acute kidney injury, glomerulonephritis, and polycystic kidney disease. These diseases significantly impact patients' quality of life and health care costs, often necessitating substantial lifestyle changes, especially regarding dietary management. However, patients frequently receive ambiguous or conflicting dietary advice from health care providers, leading them to seek information and support from online health communities.

OBJECTIVE

This study aimed to analyze social media data to better understand the experiences, challenges, and concerns of patients with kidney disease and their caregivers in South Korea. Specifically, it explored how online communities assist in disease management and examined the sentiment surrounding dietary management.

METHODS

Data were collected from KidneyCafe, a prominent South Korean online community for patients with kidney disease hosted on the Naver platform. A total of 124,211 posts from 10 disease-specific boards were analyzed using latent Dirichlet allocation for topic modeling and Bidirectional Encoder Representations From Transformers-based sentiment analysis. In addition, Efficiently Learning an Encoder That Classifies Token Replacements Accurately-based classification was used to further analyze posts related to disease management.

RESULTS

The analysis identified 6 main topics within the community: family health and support, medication and side effects, examination and diagnosis, disease management, surgery for dialysis, and costs and insurance. Sentiment analysis revealed that posts related to the medication and side effects and surgery for dialysis topics predominantly expressed negative sentiments. Both significant negative sentiments concerning worries about kidney transplantation among family members and positive sentiments regarding physical improvements after transplantation were expressed in posts about family health and support. For disease management, 7 key subtopics were identified, with inquiries about dietary management being the leading subtopic.

CONCLUSIONS

The findings highlight the critical role of online communities in providing support and information for patients with kidney disease and their caregivers. The insights gained from this study can inform health care providers, policy makers, and support organizations to better address the needs of patients with kidney disease, particularly in areas related to dietary management and emotional support.

Topic modeling on clinical social work notes for exploring social determinants of health factors

Objective

Existing research on social determinants of health (SDoH) predominantly focuses on physician notes and structured data within electronic medical records. This study posits that social work notes are an untapped, potentially rich source for SDoH information. We hypothesize that clinical notes recorded by social workers, whose role is to ameliorate social and economic factors, might provide a complementary information source of data on SDoH compared to physician notes, which primarily concentrate on medical diagnoses and treatments. We aimed to use word frequency analysis and topic modeling to identify prevalent terms and robust topics of discussion within a large cohort of social work notes including both outpatient and in-patient consultations.

Materials and methods

We retrieved a diverse, deidentified corpus of 0.95 million clinical social work notes from 181 644 patients at the University of California, San Francisco. We conducted word frequency analysis related to ICD-10 chapters to identify prevalent terms within the notes. We then applied Latent Dirichlet Allocation (LDA) topic modeling analysis to characterize this corpus and identify potential topics of discussion, which was further stratified by note types and disease groups.

Results

Word frequency analysis primarily identified medical-related terms associated with specific ICD10 chapters, though it also detected some subtle SDoH terms. In contrast, the LDA topic modeling analysis extracted 11 topics explicitly related to social determinants of health risk factors, such as financial status, abuse history, social support, risk of death, and mental health. The topic modeling approach effectively demonstrated variations between different types of social work notes and across patients with different types of diseases or conditions.

Discussion

Our findings highlight LDA topic modeling's effectiveness in extracting SDoH-related themes and capturing variations in social work notes, demonstrating its potential for informing targeted interventions for at-risk populations.

Conclusion

Social work notes offer a wealth of unique and valuable information on an individual's SDoH. These notes present consistent and meaningful topics of discussion that can be effectively analyzed and utilized to improve patient care and inform targeted interventions for at-risk populations.

Mining User Reviews From Hypertension Management Mobile Health Apps to Explore Factors Influencing User Satisfaction and Their Asymmetry: Comparative Study

BACKGROUND

Hypertension significantly impacts the well-being and health of individuals globally. Hypertension management apps (HMAs) have been shown to assist patients in controlling blood pressure (BP), with their efficacy validated in clinical trials. However, the utilization of HMAs continues to be suboptimal. Presently, there is a dearth of real-world research based on big data and exploratory mining that compares Chinese and American HMAs.

OBJECTIVE

This study aims to systematically gather HMAs and their user reviews from both China and the United States. Subsequently, using data mining techniques, the study aims to compare the user experience, satisfaction levels, influencing factors, and asymmetry between Chinese and American users of HMAs. In addition, the study seeks to assess the disparities in satisfaction and its determinants while delving into the asymmetry of these factors.

METHODS

The study sourced HMAs and user reviews from 10 prominent Chinese and American app stores globally. Using the latent Dirichlet allocation (LDA) topic model, the research identified various topics within user reviews. Subsequently, the Tobit model was used to investigate the impact and distinctions of each topic on user satisfaction. The Wald test was applied to analyze differences in effects across various factors.

RESULTS

We examined a total of 261 HMAs along with their associated user reviews, amounting to 116,686 reviews in total. In terms of quantity and overall satisfaction levels, Chinese HMAs (n=91) and corresponding reviews (n=16,561) were notably fewer compared with their American counterparts (n=220 HMAs and n=100,125 reviews). The overall satisfaction rate among HMA users was 75.22% (87,773/116,686), with Chinese HMAs demonstrating a higher satisfaction rate (13,866/16,561, 83.73%) compared with that for American HMAs (73,907/100,125, 73.81%). Chinese users primarily focus on reliability (2165/16,561, 13.07%) and measurement accuracy (2091/16,561, 12.63%) when considering HMAs, whereas American users prioritize BP tracking (17,285/100,125, 17.26%) and data synchronization (12,837/100,125, 12.82%). Seven factors (easy to use: P<.001; measurement accuracy: P<.001; compatibility: P<.001; cost: P<.001; heart rate detection function: P=.02; blood pressure tracking function: P<.001; and interface design: P=.01) significantly influenced the positive deviation (PD) of Chinese HMA user satisfaction, while 8 factors (easy to use: P<.001; reliability: P<.001; measurement accuracy: P<.001; compatibility: P<.001; cost: P<.001; interface design: P<.001; real-time: P<.001; and data privacy: P=.001) affected the negative deviation (ND). Notably, BP tracking had the greatest effect on PD (β=.354, P<.001), while cost had the most significant impact on ND (β=3.703, P<.001). All 12 factors (easy to use: P<.001; blood pressure tracking function: P<.001; data synchronization: P<.001; blood pressure management effect: P<.001; heart rate detection function: P<.001; data sharing: P<.001; reliability: P<.001; compatibility: P<.001; interface design: P<.001; advertisement distribution: P<.001; measurement accuracy: P<.001; and cost: P<.001) significantly influenced the PD and ND of American HMA user satisfaction. Notably, BP tracking had the greatest effect on PD (β=0.312, P<.001), while data synchronization had the most significant impact on ND (β=2.662, P<.001). In addition, the influencing factors of PD and ND in user satisfaction of HMA in China and the United States are different.

CONCLUSIONS

User satisfaction factors varied significantly between different countries, showing considerable asymmetry. For Chinese HMA users, ease of use and interface design emerged as motivational factors, while factors such as cost, measurement accuracy, and compatibility primarily contributed to user dissatisfaction. For American HMA users, motivational factors were ease of use, BP tracking, BP management effect, interface design, measurement accuracy, and cost. Moreover, users expect features such as data sharing, synchronization, software reliability, compatibility, heart rate detection, and nonintrusive advertisement distribution. Tailored experience plans should be devised for different user groups in various countries to address these diverse preferences and requirements.

Readability and topics of the German Health Web: Exploratory study and text analysis

BACKGROUND

The internet has become an increasingly important resource for health information, especially for lay people. However, the information found does not necessarily comply with the user's health literacy level. Therefore, it is vital to (1) identify prominent information providers, (2) quantify the readability of written health information, and (3) to analyze how different types of information sources are suited for people with differing health literacy levels.

OBJECTIVE

In previous work, we showed the use of a focused crawler to "capture" and describe a large sample of the "German Health Web", which we call the "Sampled German Health Web" (sGHW). It includes health-related web content of the three mostly German speaking countries Germany, Austria, and Switzerland, i.e. country-code top-level domains (ccTLDs) ".de", ".at" and ".ch". Based on the crawled data, we now provide a fully automated readability and vocabulary analysis of a subsample of the sGHW, an analysis of the sGHW's graph structure covering its size, its content providers and a ratio of public to private stakeholders. In addition, we apply Latent Dirichlet Allocation (LDA) to identify topics and themes within the sGHW.

METHODS

Important web sites were identified by applying PageRank on the sGHW's graph representation. LDA was used to discover topics within the top-ranked web sites. Next, a computer-based readability and vocabulary analysis was performed on each health-related web page. Flesch Reading Ease (FRE) and the 4th Vienna formula (WSTF) were used to assess the readability. Vocabulary was assessed by a specifically trained Support Vector Machine classifier.

RESULTS

In total, n = 14,193,743 health-related web pages were collected during the study period of 370 days. The resulting host-aggregated web graph comprises 231,733 nodes connected via 429,530 edges (network diameter = 25; average path length = 6.804; average degree = 1.854; modularity = 0.723). Among 3000 top-ranked pages (1000 per ccTLD according to PageRank), 18.50%(555/3000) belong to web sites from governmental or public institutions, 18.03% (541/3000) from nonprofit organizations, 54.03% (1621/3000) from private organizations, 4.07% (122/3000) from news agencies, 3.87% (116/3000) from pharmaceutical companies, 0.90% (27/3000) from private bloggers, and 0.60% (18/3000) are from others. LDA identified 50 topics, which we grouped into 11 themes: "Research & Science", "Illness & Injury", "The State", "Healthcare structures", "Diet & Food", "Medical Specialities", "Economy", "Food production", "Health communication", "Family" and "Other". The most prevalent themes were "Research & Science" and "Illness & Injury" accounting for 21.04% and 17.92% of all topics across all ccTLDs and provider types, respectively. Our readability analysis reveals that the majority of the collected web sites is structurally difficult or very difficult to read: 84.63% (2539/3000) scored a WSTF ≥ 12, 89.70% (2691/3000) scored a FRE ≤ 49. Moreover, our vocabulary analysis shows that 44.00% (1320/3000) web sites use vocabulary that is well suited for a lay audience.

CONCLUSIONS

We were able to identify major information hubs as well as topics and themes within the sGHW. Results indicate that the readability within the sGHW is low. As a consequence, patients may face barriers, even though the vocabulary used seems appropriate from a medical perspective. In future work, the authors intend to extend their analyses to identify trustworthy health information web sites.

Mining the Influencing Factors and Their Asymmetrical Effects of mHealth Sleep App User Satisfaction From Real-world User-Generated Reviews: Content Analysis and Topic Modeling

BACKGROUND

Sleep disorders are a global challenge, affecting a quarter of the global population. Mobile health (mHealth) sleep apps are a potential solution, but 25% of users stop using them after a single use. User satisfaction had a significant impact on continued use intention.

OBJECTIVE

This China-US comparison study aimed to mine the topics discussed in user-generated reviews of mHealth sleep apps, assess the effects of the topics on user satisfaction and dissatisfaction with these apps, and provide suggestions for improving users' intentions to continue using mHealth sleep apps.

METHODS

An unsupervised clustering technique was used to identify the topics discussed in user reviews of mHealth sleep apps. On the basis of the two-factor theory, the Tobit model was used to explore the effect of each topic on user satisfaction and dissatisfaction, and differences in the effects were analyzed using the Wald test.

RESULTS

A total of 488,071 user reviews of 10 mainstream sleep apps were collected, including 267,589 (54.8%) American user reviews and 220,482 (45.2%) Chinese user reviews. The user satisfaction rates of sleep apps were poor (China: 56.58% vs the United States: 45.87%). We identified 14 topics in the user-generated reviews for each country. In the Chinese data, 13 topics had a significant effect on the positive deviation (PD) and negative deviation (ND) of user satisfaction. The 2 variables (PD and ND) were defined by the difference between the user rating and the overall rating of the app in the app store. Among these topics, the app's sound recording function (β=1.026; P=.004) had the largest positive effect on the PD of user satisfaction, and the topic with the largest positive effect on the ND of user satisfaction was the sleep improvement effect of the app (β=1.185; P<.001). In the American data, all 14 topics had a significant effect on the PD and ND of user satisfaction. Among these, the topic with the largest positive effect on the ND of user satisfaction was the app's sleep promotion effect (β=1.389; P<.001), whereas the app's sleep improvement effect (β=1.168; P<.001) had the largest positive effect on the PD of user satisfaction. The Wald test showed that there were significant differences in the PD and ND models of user satisfaction in both countries (all P<.05), indicating that the influencing factors of user satisfaction with mHealth sleep apps were asymmetrical. Using the China-US comparison, hygiene factors (ie, stability, compatibility, cost, and sleep monitoring function) and 2 motivation factors (ie, sleep suggestion function and sleep promotion effects) of sleep apps were identified.

CONCLUSIONS

By distinguishing between the hygiene and motivation factors, the use of sleep apps in the real world can be effectively promoted.

Questionnaires based on natural language processing elicit immersive ruminative thinking in ruminators: Evidence from behavioral responses and EEG data

Rumination is closely related to mental disorders and can thus be used as a marker of their presence or a predictor of their development. The presence of masking and fabrication in psychological selection can lead to inaccurate detection of psychological disorders. Human language is considered crucial in eliciting specific conscious activities, and the use of natural language processing (NLP) in the development of questionnaires for psychological tests has the potential to elicit immersive ruminative thinking, leading to changes in neural activity. Electroencephalography (EEG) is commonly used to detect and record neural activity in the human brain and is sensitive to changes in brain activity. In this study, we used NLP to develop a questionnaire to induce ruminative thinking and then recorded the EEG signals in response to the questionnaire. The behavioral results revealed that ruminators exhibited higher arousal rates and longer reaction times, specifically in response to the ruminative items of the questionnaire. The EEG results showed no significant difference between the ruminators and the control group during the resting state; however, a significant alteration in the coherence of the entire brain of the ruminators existed while they were answering the ruminative items. No differences were found in the control participants while answering the two items. These behavioral and EEG results indicate that the questionnaire elicited immersive ruminative thinking, specifically in the ruminators. Therefore, the questionnaire designed using NLP is capable of eliciting ruminative thinking in ruminators, offering a promising approach for the early detection of mental disorders in psychological selection.

Identifying Topics and Evolutionary Trends of Literature on Brain Metastases Using Latent Dirichlet Allocation

Research on brain metastases kept innovating. We aimed to illustrate what topics the research focused on and how it varied in different periods of all the studies on brain metastases with topic modelling. We used the latent Dirichlet allocation model to analyse the titles and abstracts of 50,176 articles on brain metastases retrieved from Web of Science, Embase and MEDLINE. We further stratified the articles to find out the topic trends of different periods. Our study identified that a rising number of studies on brain metastases were published in recent decades at a higher rate than all cancer articles. Overall, the major themes focused on treatment and histopathology. Radiotherapy took over the first and third places in the top 20 topics. Since the 2010's, increasing attention concerned about gene mutations. Targeted therapy was a popular topic of brain metastases research after 2020.

Use of Machine Learning to Mine User-Generated Content From Mobile Health Apps for Weight Loss to Assess Factors Correlated With User Satisfaction

IMPORTANCE

The effectiveness of mobile health (mHealth) apps for reducing obesity is not ideal in daily life. Therefore, it would be useful to explore factors associated with user satisfaction with weight loss apps. Currently, research on these factors from the perspective of user-generated content is lacking.

OBJECTIVE

To mine the themes and topics frequently discussed in user-generated content in mHealth apps for weight loss, explore correlations of the topics with user satisfaction and dissatisfaction, and assess whether these correlations were asymmetric.

DESIGN, SETTING, AND PARTICIPANTS

In this population-based cross-sectional study, unsupervised machine learning was used to identify themes and topics in online discussions generated between January 1, 2019, and May 20, 2021, by Chinese users of mHealth apps for weight loss.

MAIN OUTCOMES AND MEASURES

Based on the 2-factor theory, a tobit regression model was used to explore the correlation of various app discussion topics with user satisfaction and dissatisfaction. Differences of the coefficients in models of positive rating deviation (PD) and negative rating deviation (ND), defined as the difference between the users' rating of the app and the app's comprehensive rating in the app stores, were analyzed by the Wald test.

RESULTS

In total, 191 619 reviews and ratings from unique usernames were collected for 2139 weight loss apps; 86 423 reviews (45.1%) from 339 apps (15.8%) were included in the study. Most users (65 249 [75.5%]) were satisfied with the mHealth app. Eighteen topics were identified and summarized into 9 themes. Nine topics had significant positive correlations with the PD of user satisfaction, and 6 had significant negative correlations. The factor with the strongest positive correlation with the PD was celebrity effect (β = 0.307; 95% CI, 0.290-0.323), and the factor with the weakest correlation was economic cost (β = -0.426; 95% CI, -0.447 to -0.406). Nine topics had significant positive correlations with the ND of user satisfaction, whereas 7 topics had significant negative correlations. The factor with the strongest positive correlation with the ND was fitness effect (β = 1.369; 95% CI, 1.283-1.455), and the factor with the strongest negative correlation was economic cost (β = -2.813; 95% CI, -2.875 to -2.751). There were significant differences in the PD and ND of user satisfaction. Nine motivation factors (ie, value-added attributes) and 7 hygiene factors (ie, user-expected attributes) for mHealth apps were identified.

CONCLUSIONS AND RELEVANCE

In this cross-sectional study, 16 factors had asymmetric correlations with user satisfaction and dissatisfaction with weight loss apps; 7 were related to basic expected attributes of the apps and 9 to value-added attributes. By distinguishing between expected and value-added factors, the use of weight loss apps may be improved.

Evaluating the Consistency Between Conceptual Frameworks and Factors Influencing the Safe Behavior of Iranian Workers in the Petrochemical Industry: Mixed Methods Study

BACKGROUND

Unsafe worker behavior is often identified as a major cause of dangerous incidents in the petrochemical industry. Behavioral safety models provide frameworks that may help to prevent such incidents by identifying factors promoting safe or unsafe behavior. We recently conducted a qualitative study to identify factors affecting workers' unsafe behaviors in an Iranian petrochemical company.

OBJECTIVE

The aims of this study were to (1) conduct a review of the relevant research literature between the years 2000 and 2019 to identify theoretical models proposed to explain and predict safe behavior in the workplace and (2) to select the model that best reflects our qualitative findings and other evidence about the factors influencing safe behaviors among petrochemical workers.

METHODS

This research used mixed methods. Initially, we conducted a qualitative study of factors that Iranian petrochemical workers believed affected their safety behavior. Four themes emerged from the semistructured interviews: (1) poor direct safety management and supervision; (2) unsafe workplace conditions; (3) workers' perceptions, skills, and training; and (4) broader organizational factors. Electronic databases, including PubMed, Embase, Scopus, Google Scholar, EBSCOhost, and Science Direct, were then searched for eligible studies on models to explain and predict safe behaviors, which were published between the years 2000 and 2019. Medical subject headings were used as the primary analytical element. Medical subject headings and subheadings were then extracted from the literature. One researcher conducted the search and 3 researchers performed screening and data extraction. Then, constructs described in each study were assessed to determine which were the most consistent with themes derived from our qualitative analysis.

RESULTS

A total of 2032 publications were found using the search strategy. Of these, 142 studies were assessed and 28 studies met the inclusion criteria and were included in the review. The themes identified in the qualitative study most closely matched 3 scales included in Wu et al's model that measured safety behavior and performance, safety leadership, and safety climate in petrochemical industries. Poor direct safety management and supervision matched with safety leadership and its subscales; unsafe workplace conditions matched with safety climate and its subscales; workers' perceptions, skills, and training matched with safety performance and its subscales; and broader organizational factors matched with some subscales of the model.

CONCLUSIONS

This is the first literature review to identify models intended to explain and predict safe behavior and select the model most consistent with themes elicited from a qualitative study. Our results showed that effective safety leadership and management and safety climate and culture systems are the most frequently identified factors affecting safe behaviors in the petrochemical industry. These results can further help safety researchers and professionals design effective behavior-based safety interventions, which can have a more sustainable and persistent impact on workers' safety behaviors.

TRIAL REGISTRATION

Iranian Registry of Clinical Trials IRCT20170515033981N2; https://www.irct.ir/trial/26107.

INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID)

RR2-10.1186/s12889-019-7126-1.