RESUMEN

Most of the existing recommendation systems using deep learning are based on the method of RNN (Recurrent Neural Network). However, due to some inherent defects of RNN, recommendation systems based on RNN are not only very time consuming but also unable to capture the long-range dependencies between user comments. Through the sentiment analysis of user comments, we can better capture the characteristics of user interest. Information entropy can reduce the adverse impact of noise words on the construction of user interests. Information entropy is used to analyze the user information content and filter out users with low information entropy to achieve the purpose of filtering noise data. A self-attention recommendation model based on entropy regularization is proposed to analyze the emotional polarity of the data set. Specifically, to model the mixed interactions from user comments, a multi-head self-attention network is introduced. The loss function of the model is used to realize the interpretability of recommendation systems. The experiment results show that our model outperforms the baseline methods in terms of MAP (Mean Average Precision) and NDCG (Normalized Discounted Cumulative Gain) on several datasets, and it achieves good interpretability.

RESUMEN

BACKGROUND: With the rapid development of information technology and web-based communities, a growing number of patients choose to consult physicians in online health communities (OHCs) for information and treatment. Although extant research has primarily discussed factors that influence the consulting choices of OHC patients, there is still a lack of research on the effects of log-in behaviors and web reviews on patient consultation. OBJECTIVE: This study aims to explore the impact of physicians' log-in behavior and web reviews on patient consultation. METHODS: We conducted a longitudinal study to examine the effects of physicians' log-in behaviors and web reviews on patient consultation by analyzing short-panel data from 911 physicians over five periods in a Chinese OHC. RESULTS: The results showed that the physician's log-in behavior had a positive effect on patient consultation. The maximum number of days with no log-ins for a physician should be 20. The two web signals (log-in behavior and web reviews) had no complementary relationship. Moreover, the offline signal (ie, offline status) has different moderating effects on the two web signals, positively moderating the relationship between web reviews and patient consultation. CONCLUSIONS: Our study contributes to the eHealth literature and advances the understanding of physicians' web-based behaviors. This study also provides practical implications, showing that physicians' log-in behavior alone can affect patient consultation rather than complementing web reviews.

Asunto(s)

RESUMEN

OBJECTIVES: Users share valuable information through online smoking cessation communities (OSCCs), which help people maintain and improve smoking cessation behavior. Although OSCC utilization is common among smokers, limitations exist in identifying the smoking status of OSCC users ("quit" vs. "not quit"). Thus, the current study implicitly analyzed user-generated content (UGC) to identify individual users' smoking status through advanced computational methods and real data from an OSCC. METHODS: Secondary data analysis was conducted using data from 3,833 users of BcomeAnEX.org. Domain experts reviewed posts and comments to determine the authors' smoking status when they wrote them. Seven types of feature sets were extracted from UGC (textual, Doc2Vec, social influence, domain-specific, author-based, and thread-based features, as well as adjacent posts). RESULTS: Introducing novel features boosted smoking status recognition (quit vs. not quit) by 9.3% relative to the use of text-only post features. Furthermore, advanced computational methods outperformed baseline algorithms across all models and increased the smoking status prediction performance by up to 12%. CONCLUSIONS: The results of this study suggest that the current research method provides a valuable platform for researchers involved in online cessation interventions and furnishes a framework for on-going machine learning applications. The results may help practitioners design a sustainable real-time intervention via personalized post recommendations in OSCCs. A major limitation is that only users' smoking status was detected. Future research might involve programming machine learning classification methods to identify abstinence duration using larger datasets.

RESUMEN

INTRODUCTION: An increasing number of patients are voicing their opinions and expectations about the quality of care in online forums and on physician rating websites (PRWs). This paper analyzes patient online reviews (PORs) to identify emerging and fading topics and sentiment trends in PRWs during the early stage of the COVID-19 outbreak. METHODS: Text data were collected, including 55,612 PORs of 3430 doctors from three popular PRWs in the United States (RateMDs, HealthGrades, and Vitals) from March 01 to June 27, 2020. An improved latent Dirichlet allocation (LDA)-based topic modeling (topic coherence-based LDA [TCLDA]), manual annotation, and sentiment analysis tool were applied to extract a suitable number of topics, generate corresponding keywords, assign topic names, and determine trends in the extracted topics and specific emotions. RESULTS: According to the coherence value and manual annotation, the identified taxonomy includes 30 topics across high-rank and low-rank disease categories. The emerging topics in PRWs focus mainly on themes such as treatment experience, policy implementation regarding epidemic control measures, individuals' attitudes toward the pandemic, and mental health across high-rank diseases. In contrast, the treatment process and experience during COVID-19, awareness and COVID-19 control measures, and COVID-19 deaths, fear, and stress were the most popular themes for low-rank diseases. Panic buying and daily life impact, treatment processes, and bedside manner were the fading themes across high-rank diseases. In contrast, provider attitude toward patients during the pandemic, detection at public transportation, passenger, travel bans and warnings, and materials supplies and society support during COVID-19 were the most fading themes across low-rank diseases. Regarding sentiment analysis, negative emotions (fear, anger, and sadness) prevail during the early wave of the COVID-19. CONCLUSION: Mining topic dynamics and sentiment trends in PRWs may provide valuable knowledge of patients' opinions during the COVID-19 crisis. Policymakers should consider these PORs and develop global healthcare policies and surveillance systems through monitoring PRWs. The findings of this study identify research gaps in the areas of e-health and text mining and offer future research directions.

Asunto(s)

RESUMEN

The recommendation system in the online medical consultation website is a system to assist patients to find appropriate doctors. Based on the analysis of the current situation of the development of an online medical community (Haodf.com) in China, this paper puts forward recommendation suggestions of finding the right hospital and doctor to promote the rapid integration of Internet technology and traditional medical services. A new recommendation model called Probabilistic Matrix Factorization integrated with Convolutional Neural Network (PMF-CNN) is proposed in the paper. Doctors' data in Haodf.com were used to evaluate the performance of our system. The model improves the performance of medical consultation recommendations by fusing review text and doctor information based on CNN (Convolutional Neural Network). Specifically, CNN is used to learn the feature representation of the review text and the doctors' information. Furthermore, the extended matrix factorization model is exploited to fuse the review information feature and the initial value of the doctors' information for recommendation. As is shown in the experimental results on Haodf.com datasets, the proposed PMF-CNN achieves better recommendation performances than the other state-of-the-art recommendation algorithms. And the recommendation system in an online medical website improves the utilization efficiency of doctors and the balance of public health resources allocation.

Asunto(s)

RESUMEN

BACKGROUND: With the dramatic development of Web 2.0, increasing numbers of patients and physicians are actively involved in online health communities. Despite extensive research on online health communities, the conversion rate from visitor to customer and its driving factors have not been discussed. OBJECTIVE: The aim of this study was to analyze the conversion rate of online health communities and to explore the effects of multisource online health community information, including physician-generated information, patient-generated information, and system-generated information. METHODS: An empirical study was conducted to examine the effects of physician-generated, patient-generated, and system-generated information on the conversion rate of physicians' personal websites by analyzing short panel data from 2112 physicians over five time periods in a Chinese online health community. RESULTS: Multisource online health community information (ie, physician-generated, patient-generated, and system-generated information) positively affected the conversion rate. Physician-generated and patient-generated information showed a substitute relationship rather than a complementary relationship. In addition, the usage time of a personal website positively moderated patient-generated information, but negatively moderated physician-generated information. CONCLUSIONS: This study contributes to the electronic health literature by investigating the conversion rate of online health communities and the effect of multisource online health community information. This study also contributes to understanding the drivers of conversion rate on service websites, which can help to successfully improve the efficiency of online health communities.

Asunto(s)

RESUMEN

Human behavior is the largest source of variance in health-related outcomes, and the increasingly popular online health communities (OHC) can be used to promote healthy behavior and outcomes. We explored how the social influence (social integration, descriptive norms and social support) exerted by online social relationships does affect the health behavior of users. Based on an OHC, we considered the effect of three types of social relationships (friendship, mutual support group and competing group) in the OHC. We found that social integration, descriptive norms and social support (information and emotional support) from the OHC had a positive effect on dietary and exercise behavior. Comparing the effects of different social relationships, we found that the stronger social relationship-friendship-had a stronger effect on health behavior than the mutual support group and competing group. Emotional support had a stronger effect on health behavior than informational support. We also found that the effects of social integration and informational support became stronger as membership duration increased, but the effects of the descriptive norms and emotional support became smaller. This study extended the research on health behavior to the online social environment and explored how the social influence exerted by various social relationships in an OHC affected health behavior. The results could be used for guiding users to make use of online social relationships for changing and maintaining healthy behavior, and helping healthcare websites improve their services.

Asunto(s)

RESUMEN

Most of the supply chain literature assumes that product substitution is an effective method to mitigate supply chain disruptions and that all production lines either survive or are disrupted together. Such assumptions, however, may not hold in the real world: (1) when there is a shortfall of all products, product substitution may be inadequate unless it is paired with other strategies such as dual sourcing; and (2) production lines do not survive forever and may fail. To relax such assumptions, this paper therefore investigates the situations that the manufacturer may optimize substitution policy and dual sourcing policy to cope with supply chain disruptions. The paper obtains and compares the optimal policies for both deterministic and stochastic demands. A real-world case is also studied to verify the effectiveness of the proposed model.

RESUMEN

INTRODUCTION: With the growth in Internet technology, online rating websites encourage patients to contribute actively in rating their physicians. These rating sites provide more information for patients, such as electronic word of mouth (eWOM) and physician trustworthiness. Although several studies in e-commerce have investigated the role of eWOM and seller trustworthiness in the consumer purchase decision-making process and the price premium for products or services, studies on the role of different information sources that reflect the service quality and delivery process in choosing a competent physician remain scarce. This research develops a two-equation model to examine the effect of different signals, i.e., patient-generated signals (PGSs) and system-generated signals (SGSs), on patient choice, which is an important predictor of physicians' economic returns. METHODS: A secondary data econometric analysis and structural modeling using 2896 physicians' real data from a publicly available online physician rating site, i.e., Healthgrades.com, were conducted using a mixed-methods approach. A hybrid text mining approach was adopted to calculate the sentiment of each review. RESULTS: We find that both PGSs and SGSs have a significant impact on patient choice at different stages of health consultation. Furthermore, disease risk negatively moderates the association between PGSs and information search, while the impact of both signals on patient willingness to pay a price premium is positively moderated by the disease risk. CONCLUSION: Our study contributes to the unified framework of signaling theory and Maslow's hierarchy of needs theory by making a clear distinction between PGSs or SGSs and their influence on patient decision-making across different disease risks. Moreover, PGSs and SGSs are two essential factors for physicians to increase their income.

Asunto(s)

RESUMEN

BACKGROUND: Web-based medical consultation, which has been adopted by patients in many countries, requires a large number of doctors to provide services. However, no study has provided an overall picture of the doctors who provide online services. OBJECTIVE: This study sought to examine doctors' participation in medical consultation practice in an online consultation platform. This paper aimed to address the following questions: (1) which doctors provide Web-based consultation services, (2) how many patients do the doctors get online, and (3) what price do they charge. We further explored the development of market segments in various departments and various provinces. METHODS: This study explored the dataset including all doctors providing consultation services in their spare time on a Chinese Web-based consultation platform. We also brought in statistics for doctors providing offline consultations in China. We made use of Bonferroni multiple comparison procedures and z test to compare doctors in each group. RESULTS: There are 88,308 doctors providing Web-based consultation in their spare time on Haodf, accounting for 5.25% (88,308/1,680,062) of all doctors in China as of September 23, 2017. Of these online doctors, 49.9% (44,066/88,308) are high-quality doctors having a title of chief physician or associate chief physician, and 84.8% (74,899/88,308) come from the top, level 3, hospitals. Online doctors had an average workload of 0.38 patients per doctor per day, with an online and offline ratio of 1:14. The average price of online consultation is ¥32.3. The online ratios for the cancer, internal medicine, ophthalmology-otorhinolaryngology, psychiatry, gynecology-obstetrics-pediatrics, dermatology, surgery, and traditional Chinese medicine departments are 16.1% (2,983/18,481), 4.4% (16,231/372,974), 6.3% (8,389/132,725), 9.5% (1,600/16,801), 5.8% (12,955/225,128), 18.0% (3,334/18,481), 10.8% (24,030/223,448), and 3.8% (8,393/22,3448), respectively. Most provinces located in eastern China have more than 4000 doctors online. The online workloads for doctors from most provinces range from 0.3 to 0.4 patients per doctor per day. In most provinces, doctors' charges range from ¥20 to ¥30. CONCLUSIONS: Quality doctors are more likely to provide Web-based consultation services, get more patients, and charge higher service fees in an online consultation platform. Policies and promotions could attract more doctors to provide Web-based consultation. The online submarket for the departments of dermatology, psychiatry, and gynecology-obstetrics-pediatrics developed better than other departments such as internal medicine and traditional Chinese medicine. The result could be a reference for policy making to improve the medical system both online and offline. As all the data used for analysis were from a single website, the data might be biased and might not be a representative national sample of China.

Asunto(s)

RESUMEN

BACKGROUND: Suicide is a leading cause of death in China, and so suicide intervention on social media is an important issue in the field of public health. Sina Weibo (Weibo) is an emerging surveillance tool that may provide online assistance for users at the risk of suicide. MATERIALS AND METHODS: Keyword-based methods and supervised classifiers were employed to conduct this research. A control group was established to explore the differences between Weibo users with suicidal ideation (USI) and the general population. RESULTS: A total of 114 USI were detected from 1 million active Weibo users. By studying the negative postings of these USI, disclosure of the reasons for their bad moods was the most common theme. The emotions of USI tend to be particularly down between 05:00 pm and midnight. Use of the first-person pronoun by Weibo USI is significantly frequent. CONCLUSIONS: Our findings may help to identify individuals with suicidal ideation who are not identified by the traditional clinical approach. Consequently, detecting and helping individuals who may be at risk of committing suicide may become more efficient.

Asunto(s)

RESUMEN

OBJECTIVES: The emergence of social media technology has led to the creation of many online healthcare communities, where patients can easily share and look for healthcare-related information from peers who have experienced a similar problem. However, with increased user-generated content, there is a need to constantly analyse which content should be trusted as one sifts through enormous amounts of healthcare information. This study aims to explore patients' healthcare information seeking behavior in online communities. METHODS: Based on dual-process theory and the knowledge adoption model, we proposed a healthcare information adoption model for online communities. This model highlights that information quality, emotional support, and source credibility are antecedent variables of adoption likelihood of healthcare information, and competition among repliers and involvement of recipients moderate the relationship between the antecedent variables and adoption likelihood. Empirical data were collected from the healthcare module of China's biggest Q&A community-Baidu Knows. Text mining techniques were adopted to calculate the information quality and emotional support contained in each reply text. A binary logistics regression model and hierarchical regression approach were employed to test the proposed conceptual model. RESULTS: Information quality, emotional support, and source credibility have significant and positive impact on healthcare information adoption likelihood, and among these factors, information quality has the biggest impact on a patient's adoption decision. In addition, competition among repliers and involvement of recipients were tested as moderating effects between these antecedent factors and the adoption likelihood. Results indicate competition among repliers positively moderates the relationship between source credibility and adoption likelihood, and recipients' involvement positively moderates the relationship between information quality, source credibility, and adoption decision. CONCLUSIONS: In addition to information quality and source credibility, emotional support has significant positive impact on individuals' healthcare information adoption decisions. Moreover, the relationships between information quality, source credibility, emotional support, and adoption decision are moderated by competition among repliers and involvement of recipients.

Asunto(s)