Author: huhjina

Paper Presents empirical evaluation of using lightweight signal features from a set of complex activities during the meal (e.g., clattering sound, arm gestures of eating, human voice, TV sound) and fusing built-in sensor data of multiple mobile devices available in a family with a CRFs-based classifier.

Abstract. Monitoring the family mealtime activities enables the analysis of the previous daily routine, hence the positive changes can be made towards better relationships among family members and better physical/mental health. Moreover, the details of family mealtime activities provide important information for study in sociology and culture. This paper presents FamilyLog — a practical system to log family mealtime activities using smartphones and smartwatches. FamilyLog automatically detects and logs details of activities during the mealtime, including occurrence and duration of meal, conversations, participants, TV viewing etc., in an unobtrusive manner. Based on the sensor data collected from real families, we carefully design robust yet lightweight signal features from a set of complex activities during the meal, including clattering sound, arm gestures of eating, human voice, TV sound, etc. Moreover, FamilyLog opportunistically fuses data from built-in sensors of multiple mobile devices available in a family with a CRFs-based classifier. To evaluate the real-world performance of FamilyLog, we perform extensive experiments that consist of 77 days of sensor data from 37 subjects in 8 families with children. FamilyLog can detect those events with high accuracy across different families and home environments.

Bi C, Xing G, Hao T, Huh J, Peng W, Ma M, Chang X. FamilyLog: Monitoring Family Mealtime Activities by Mobile Devices. IEEE Transactions on Mobile Computing. 2019 May 14.

Paper Systematic review of facilitators and barriers to underserved populations’ consumer health technology adoption

Objectives: Underserved populations can benefit from consumer health informatics (CHI) that promotes self-management at a lower cost. However, prior literature suggested that the digital divide and low motivation constituted barriers to CHI adoption. Despite increased Internet use, underserved populations continue to show slow CHI uptake. The aim of the paper is to revisit barriers and facilitators that may impact CHI adoption among underserved populations.

Methods: We surveyed the past five years of literature. We searched PubMed for articles published between 2012 and 2017 that describe empirical evaluations involving CHI use by under- served populations. We abstracted and summarized data about facilitators and barriers impacting CHI adoption.

Results:From 645 search results, after abstract and full-text screening, 13 publications met the inclusion criteria of identify- ing barriers to and facilitators of underserved populations’ CHI adoption. Contrary to earlier literature, the studies suggested that the motivation to improve health literacy and adopt technology was high among studied populations. Beyond the digital divide, barriers included: low health and computer literacy, challenges in accepting the presented information, poor usability, and unclear content. Factors associated with increased use were: user needs for information, user-access mediated by a proxy person, and early user engagement in system design.

Conclusions: While the digital divide remains a barrier, newer studies show that high motivation for CHI use exists. However, simply gaining access to technology is not sufficient to improve adoption unless CHI technology is tailored to address user needs. Future interventions should consider building larger empirical evidence on identifying CHI barriers and facilitators.

Keywords Medical informatics applications, consumer health information, ethnic groups, socioeconomic factors, minority groups, health disparities

Identified the features critical for predicting social support needs in online health communities.

Background: While online health social networks (OHSNs) serve as an effective platform for patients to fulfill their various social support needs, predicting the needs of users and providing tailored information remains a challenge.

Objective: The objective of this study is to discriminate important features for identifying users’ social support needs based on knowledge gathered from survey data. This study also provides guidelines for a technical framework which can be used to predict users’ social support needs based on raw data collected from OHSNs.

Methods: We initially conducted an online survey with 184 OHSN users. From this survey data, we extracted 34 features based on five categories: (1) demographics, (2) reading behavior, (3) posting behavior, (4) perceived roles in OHSNs, and (5) values sought in OHSNs. Features from the first four categories were used as variables for binary classification. For the prediction outcomes, we used features from the last category: the needs for emotional support, experience-based information, unconventional information, and medical facts. We compared 5 binary classifier algorithms: Gradient Boosting Tree, Random Forest, Decision Tree, Support Vector Machines, and Logistic Regression. We then calculated the scores of the area under the ROC curve (AUC) to understand the comparative effectiveness of the used features.

Results: The best performance was AUC scores of 0.89 for predicting users seeking emotional support, 0.86 for experience-based information, 0.80 for unconventional information, and 0.83 for medical facts. With the gradient boosting tree as our best performing model, we analyzed the strength of individual features in predicting one’s social support need. Among other discoveries, we discovered that users seeking emotional support tend to post more in OHSNs compared to others.

Conclusions: We developed an initial framework for automatically predicting social support needs in OHSNs using survey data. Future work should involve non-survey data to evaluate the feasibility of the framework. Our study contributes to providing personalized social support in OHSNs.

Min-Je Choi, Sung-Hee Kim, Sukwon Lee, et al. 2017. Toward Predicting Social Support Needs in Online Health Social Networks. Journal of Medical Internet Research (forthcoming). http://doi.org/10.2196/jmir.7660

The paper investigated whether conversations on online breast cancer community contribute to people choosing an aggressive, higher risk surgery.

Abstract
Background: The increased uptake of contralateral prophylactic mastectomy (CPM) among breast cancer patients remains poorly understood. We hypothesized that the increased rate of CPM is represented in conversations on an online breast cancer community and may contribute to patients choosing this operation.

Methods: We downloaded 328,763 posts and their dates of creation from an online breast cancer community from August 1, 2000 to May 22, 2016. We then performed a keyword search to identify posts which mentioned breast cancer surgeries: contralateral prophylactic mastectomy (n=7,095), mastectomy (n=10,889) and lumpectomy (n=9,694). We graphed the percentage of CPM-related, lumpectomy-related and mastectomy-related conversations over time. We also graphed the frequency of posts which mentioned multiple operations over time. Finally, we performed a qualitative study to identify factors influencing the observed trends.

Results: Surgically-related posts (e.g., mentioning at least one operation) made up a small percentage (n=27,678; 8.4%) of all posts on this community. The percentage of surgically related posts mentioning CPM was found to increase over time, whereas the percentage of *Revision version w/ Markings surgically-related posts mentioning mastectomy decreased over time. Among posts that mentioned more than one operation, mastectomy and lumpectomy were the procedures most commonly mentioned together, followed by mastectomy and CPM. There was no change over time in the frequency of posts that mentioned more than one operation. Our qualitative review found that the majority of posts mentioning a single operation were unrelated to surgical decision-making; rather the operation was mentioned only in the context of the patient’s cancer history. Conversely, the majority of posts mentioning multiple operations centered around the patients’ surgical decision-making process.

Conclusions: CPM-related conversation is increasing on this online breast cancer community, while mastectomy-related conversation is decreasing. These results appear to be primarily informed by patients reporting the types of operations they have undergone, and thus appear to correspond to the known increased uptake of CPM.

R.A. Marmor*, W. Dai, X. Jiang, S. Wang, S.L. Blair, J. Huh, Increase In Contralateral Prophylactic Mastectomy Conversation Online Unrelated To Decision-Making, Journal of Surgical Research. 218 (2017).

This paper presents evaluation of the FamilyLog system with 37 subjects from 8 families. The system automatically detects activities during the mealtime, including occurrence and duration of meal, conversations, participants, and TV viewing using acoustic data from phones and smartwatches. [paper]

Title
FamilyLog: A Mobile System for Monitoring Family Mealtime Activities

Abstract
Research has shown that family mealtime plays a critical role in establishing good relationships among family members and maintaining their physical and mental health. In particular, regularly eating dinner as a family significantly reduces prevalence of obesity. However, American families with children spend only 1 hour on family meals while three hours watching TV on an average work day. Fine-grained activity logging is proven effective for increasing self-awareness and motivating people to modify their life styles for improved wellness. This paper presents FamilyLog – a practical system to log family mealtime activities using smartphones and smartwatches. FamilyLog automatically detects and logs details of activities during the mealtime, including occurrence and duration of meal, conversations, participants, TV viewing etc., in an unobtrusive manner. Based on the sensor data collected from real families, we carefully design robust yet lightweight signal features from a set of complex activities during the meal, including clattering sound, arm gestures of eating, human voice, TV sound, etc. Moreover, FamilyLog opportunistically fuses data from built-in sensors of multiple mobile devices available in a family through an HMM-based classifier. To evaluate the real-world performance of FamilyLog, we perform extensive experiments that consist of 77 days of sensor data from 37 subjects in 8 families with children. Our results show that FamilyLog can detect those events with high accuracy across different families and home environments.

C. Bi, G. Xing, T. Hao, J. Huh, W. Peng, M. Ma, FamilyLog: A Mobile System for Monitoring Family Mealtime Activities, in: IEEE International Conference on Pervasive Computing and Communications 2017, Institute of Electrical and Electronics Engineers Inc., 2017: pp. 21–30.