What are the Gaps in Mobile Patient Portal? Mining Users Feedback Using Topic Modeling

Noteboom, Cherie; Al-Ramahi, Mohammad

doi:10.24251/hicss.2018.072

Cited by 10 publications

(3 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Topic models make it possible to summarize textual data at a scale that cannot possibly be tackled by human annotation. In this study, we chose the LDA algorithm [25] owing to its conceptual advantage over other latent topic models [35][36][37][38].…”

Section: Data Analysis Using the Lda Algorithm (Unsupervised Learning)mentioning

confidence: 99%

Public Discourse Against Masks in the COVID-19 Era: Infodemiology Study of Twitter Data

Al-Ramahi¹,

Elnoshokaty²,

El‐Gayar³

et al. 2021

JMIR Public Health Surveill

Self Cite

View full text Add to dashboard Cite

Background Despite scientific evidence supporting the importance of wearing masks to curtail the spread of COVID-19, wearing masks has stirred up a significant debate particularly on social media. Objective This study aimed to investigate the topics associated with the public discourse against wearing masks in the United States. We also studied the relationship between the anti-mask discourse on social media and the number of new COVID-19 cases. Methods We collected a total of 51,170 English tweets between January 1, 2020, and October 27, 2020, by searching for hashtags against wearing masks. We used machine learning techniques to analyze the data collected. We investigated the relationship between the volume of tweets against mask-wearing and the daily volume of new COVID-19 cases using a Pearson correlation analysis between the two-time series. Results The results and analysis showed that social media could help identify important insights related to wearing masks. The results of topic mining identified 10 categories or themes of user concerns dominated by (1) constitutional rights and freedom of choice; (2) conspiracy theory, population control, and big pharma; and (3) fake news, fake numbers, and fake pandemic. Altogether, these three categories represent almost 65% of the volume of tweets against wearing masks. The relationship between the volume of tweets against wearing masks and newly reported COVID-19 cases depicted a strong correlation wherein the rise in the volume of negative tweets led the rise in the number of new cases by 9 days. Conclusions These findings demonstrated the potential of mining social media for understanding the public discourse about public health issues such as wearing masks during the COVID-19 pandemic. The results emphasized the relationship between the discourse on social media and the potential impact on real events such as changing the course of the pandemic. Policy makers are advised to proactively address public perception and work on shaping this perception through raising awareness, debunking negative sentiments, and prioritizing early policy intervention toward the most prevalent topics.

show abstract

Section: Data Analysis Using the Lda Algorithm (Unsupervised Learning)mentioning

confidence: 99%

Public Discourse Against Masks in the COVID-19 Era: Infodemiology Study of Twitter Data

Al-Ramahi¹,

Elnoshokaty²,

El‐Gayar³

et al. 2021

JMIR Public Health Surveill

Self Cite

View full text Add to dashboard Cite

show abstract

“…Previous research had used the traditional qualitative content analysis to analyze these comments in order to retrieve the key themes in these OPRs [7,37]. More advanced techniques such as NLP have been used in recent articles [4,[33][34][35][36][37]112,114]. For example, topic models, such as Latent Dirichlet Allocation, point out the different themes involved in online reviews that may be linked to one of the topics.…”

Section: Study Design and Technological Roadmap Adoptedmentioning

confidence: 99%

Examining Different Factors in Web-Based Patients’ Decision-Making Process: Systematic Review on Digital Platforms for Clinical Decision Support System

Shah

Muhammad

Lee

2021

IJERPH

View full text Add to dashboard Cite

(1) Background: The appearance of physician rating websites (PRWs) has raised researchers’ interest in the online healthcare field, particularly how users consume information available on PRWs in terms of online physician reviews and providers’ information in their decision-making process. The aim of this study is to consistently review the early scientific literature related to digital healthcare platforms, summarize key findings and study features, identify literature deficiencies, and suggest digital solutions for future research. (2) Methods: A systematic literature review using key databases was conducted to search published articles between 2010 and 2020 and identified 52 papers that focused on PRWs, different signals in the form of PRWs’ features, the findings of these studies, and peer-reviewed articles. The research features and main findings are reported in tables and figures. (3) Results: The review of 52 papers identified 22 articles for online reputation, 15 for service popularity, 16 for linguistic features, 15 for doctor–patient concordance, 7 for offline reputation, and 11 for trustworthiness signals. Out of 52 studies, 75% used quantitative techniques, 12% employed qualitative techniques, and 13% were mixed-methods investigations. The majority of studies retrieved larger datasets using machine learning techniques (44/52). These studies were mostly conducted in China (38), the United States (9), and Europe (3). The majority of signals were positively related to the clinical outcomes. Few studies used conventional surveys of patient treatment experience (5, 9.61%), and few used panel data (9, 17%). These studies found a high degree of correlation between these signals with clinical outcomes. (4) Conclusions: PRWs contain valuable signals that provide insights into the service quality and patient treatment choice, yet it has not been extensively used for evaluating the quality of care. This study offers implications for researchers to consider digital solutions such as advanced machine learning and data mining techniques to test hypotheses regarding a variety of signals on PRWs for clinical decision-making.

show abstract

“…LDA is an unsupervised, generative statistical model for learning subgroups of observations within a data set based on similarities among observations [ 16 ]. LDA has been leveraged to derive insights into patient communication data in patient portals [ 17 , 18 ], but its usage to classify message intent, particularly in mobile text messaging technologies, has been largely unexplored. The purpose of this proof-of-concept study was to explore the extent to which LDA might be applied to patient-provider messages to accurately identify domains (ie, topics) of clinical relevance (ie, those topics that were deemed informative for clinical action that are indicative of intents).…”

Section: Introductionmentioning

confidence: 99%

Identifying Medication-Related Intents From a Bidirectional Text Messaging Platform for Hypertension Management Using an Unsupervised Learning Approach: Retrospective Observational Pilot Study

Davoudi¹,

Lee²,

Luong³

et al. 2022

J Med Internet Res

View full text Add to dashboard Cite

Background Free-text communication between patients and providers plays an increasing role in chronic disease management, through platforms varying from traditional health care portals to novel mobile messaging apps. These text data are rich resources for clinical purposes, but their sheer volume render them difficult to manage. Even automated approaches, such as natural language processing, require labor-intensive manual classification for developing training data sets. Automated approaches to organizing free-text data are necessary to facilitate use of free-text communication for clinical care. Objective The aim of this study was to apply unsupervised learning approaches to (1) understand the types of topics discussed and (2) learn medication-related intents from messages sent between patients and providers through a bidirectional text messaging system for managing participant blood pressure (BP). Methods This study was a secondary analysis of deidentified messages from a remote, mobile, text-based employee hypertension management program at an academic institution. We trained a latent Dirichlet allocation (LDA) model for each message type (ie, inbound patient messages and outbound provider messages) and identified the distribution of major topics and significant topics (probability >.20) across message types. Next, we annotated all medication-related messages with a single medication intent. Then, we trained a second medication-specific LDA (medLDA) model to assess how well the unsupervised method could identify more fine-grained medication intents. We encoded each medication message with n-grams (n=1-3 words) using spaCy, clinical named entities using Stanza, and medication categories using MedEx; we then applied chi-square feature selection to learn the most informative features associated with each medication intent. Results In total, 253 participants and 5 providers engaged in the program, generating 12,131 total messages: 46.90% (n=5689) patient messages and 53.10% (n=6442) provider messages. Most patient messages corresponded to BP reporting, BP encouragement, and appointment scheduling; most provider messages corresponded to BP reporting, medication adherence, and confirmatory statements. Most patient and provider messages contained 1 topic and few contained more than 3 topics identified using LDA. In total, 534 medication messages were annotated with a single medication intent. Of these, 282 (52.8%) were patient medication messages: most referred to the medication request intent (n=134, 47.5%). Most of the 252 (47.2%) provider medication messages referred to the medication question intent (n=173, 68.7%). Although the medLDA model could identify a majority intent within each topic, it could not distinguish medication intents with low prevalence within patient or provider messages. Richer feature engineering identified informative lexical-semantic patterns associated with each medication intent class. Conclusions LDA can be an effective method for generating subgroups of messages with similar term usage and facilitating the review of topics to inform annotations. However, few training cases and shared vocabulary between intents precludes the use of LDA for fully automated, deep, medication intent classification. International Registered Report Identifier (IRRID) RR2-10.1101/2021.12.23.21268061

show abstract

What are the Gaps in Mobile Patient Portal? Mining Users Feedback Using Topic Modeling

Cited by 10 publications

References 24 publications

Public Discourse Against Masks in the COVID-19 Era: Infodemiology Study of Twitter Data

Public Discourse Against Masks in the COVID-19 Era: Infodemiology Study of Twitter Data

Examining Different Factors in Web-Based Patients’ Decision-Making Process: Systematic Review on Digital Platforms for Clinical Decision Support System

Identifying Medication-Related Intents From a Bidirectional Text Messaging Platform for Hypertension Management Using an Unsupervised Learning Approach: Retrospective Observational Pilot Study

Contact Info

Product

Resources

About