Search

Type 2 diabetes disproportionately impacts ethnic minorities and individuals from low socioeconomic status. Diabetes self-management education and support has been shown to improve clinical outcomes in these populations, and mobile health (mHealth) interventions can reduce barriers to access. Dulce Digital-Me (DD-Me) was developed to integrate adaptive mHealth technologies to enhance self-management and reduce disparities in the high-risk, underserved Hispanic population. The objective of the present study was to evaluate reach, adoption, and implementation of an mHealth diabetes self-management education and support intervention in this underrepresented population. The present analysis is a multimethod process evaluation using the Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) framework. The study was effective in reaching a sample that was representative of the intended population; only modest but significant differences were observed in sex and age. The DD-Me health coach (HC) cited several important facilitators of intervention adoption, including outreach frequency and personalization, and the automated HC report. Implementation fidelity was high, with participants receiving >90% of intended interventions. Participants who received DD-Me with support from a HC were most engaged, suggesting utility and acceptability of integrating HCs with mHealth interventions. Perceptions of implementation among study participants were positive and consistent across study arms. This evaluation revealed the target population was successfully reached and engaged in the digital health interventions, which was implemented with high fidelity. Further studies should evaluate the efficacy and maintenance of the study following the RE-AIM model to determine whether this intervention warrants expansion to additional settings and populations.

BACKGROUND Older adults are the fastest growing digital technology users. IVAs embodied in smart speakers are popular and they are accessible to people with sensory, physical and cognitive impairments. The geriatrics 5Ms: Medications, Mind, Mobility, what Matters most and Multicomplexity, addresses the complex care needs of older adults. OBJECTIVE We examined data on older adults’ views of how an IVA may address their health-related needs and ascertained their fit into the geriatrics 5Ms framework. METHODS We analyzed qualitative data from interviews among 16 older adults. Two authors independently assigned a subset of the original codes into themes corresponding to the 5Ms. RESULTS We sorted 111 codes into the 5Ms. Representative themes for each of the 5Ms and potential IVA uses are: 1) Medication: difficulty remembering complex medication schedules. Use: medication reminders. 2) Mind: isolation, depression, anxiety and memory loss. Use: companionship, behavioral interventions and memory aids. 3) Mobility: barriers to exercise. Use: reminders, incentives, ideas for exercise. 4) Matters most: participant health concerns (e.g., nutrition). Use: suggesting nutritious foods, and ordering them from local stores. 5) Multicomplexity; managing multimorbidity. Use: symptom tracking and communicating with health care professionals and care partners. CONCLUSIONS Older adult participants’ needs and recommendations for a health-related IVA can be put into a 5Ms framework. Moreover, we identified numerous uses for an IVA utilizing this older adult needs-focused framework. Incorporating the 5Ms framework into IVA design will improve their ability to address health care priorities of older adults and their care partners.

Objectives. Remote monitoring (RM) of health-related outcomes may optimize cancer care and prevention outside of clinic settings. CYCORE is a software-based system for collection and analyses of sensor and mobile data. We evaluated CYCORE’s feasibility in studies assessing: (1) physical functioning in colorectal cancer (CRC) patients; (2) swallowing exercise adherence in head and neck cancer (HNC) patients during radiation therapy; and (3) tobacco use in cancer survivors post-tobacco treatment (TTP). Methods. Participants completed RM: for CRC, blood pressure, activity, GPS; for HNC, video of swallowing exercises; for TTP, expired carbon monoxide. Patient-reported outcomes were assessed daily. Results. For CRC, HNC and TTP, respectively, 50, 37, and 50 participants achieved 96%, 84%, 96% completion rates. Also, 91-100% rated ease and self-efficacy as highly favorable, 72-100% gave equivalent ratings for overall satisfaction, 72-93% had low/no data privacy concerns. Conclusion. RM was highly feasible and acceptable for patients across diverse use cases.

Voice based Intelligent Virtual Assistants (IVAs) promise to improve healthcare management and Quality of Life (QOL) by introducing the paradigm of hands free and eye free interactions. However, there has been little understanding regarding the challenges for designing such systems for older adults, especially when it comes to healthcare related tasks. To tackle this, we consider the processes of care delivery and QOL enhancements for older adults as a collaborative task between patients and providers. By interviewing 16 older adults living independently or semi independently and 5 providers, we identified 12 barriers that older adults might encounter during daily routine and while managing health. We ultimately highlighted key design challenges and opportunities that might be introduced when integrating voice based IVAs into the life of older adults. Our work will benefit practitioners who study and attempt to create full fledged IVA powered smart devices to deliver better care and support an increased QOL for aging populations., 16 pages, The 23rd International ACM SIGACCESS Conference on Computers and Accessibility

Understanding the intent of medical questions asked by patients, or Consumer Health Questions, is an essential skill for medical Conversational AI systems. We propose a novel data-augmented and simple joint learning approach combining question summarization and Recognizing Question Entailment (RQE) in the medical domain. Our data augmentation approach enables to use just one dataset for joint learning. We show improvements on both tasks across four biomedical datasets in accuracy (+8%), ROUGE-1 (+2.5%) and human evaluation scores. Human evaluation shows joint learning generates faithful and informative summaries. Finally, we release our code, the two question summarization datasets extracted from a large-scale medical dialogue dataset, as well as our augmented datasets.

Syntactic structure is an important component of natural language text. Recent top-performing models in Answer Sentence Selection (AS2) use self-attention and transfer learning, but not syntactic structure. Tree structures have shown strong performance in tasks with sentence pair input like semantic relatedness. We investigate whether tree structures can boost performance in AS2. We introduce the Tree Aggregation Transformer: a novel recursive, tree-structured self-attention model for AS2. The recursive nature of our model is able to represent all levels of syntactic parse trees with only one additional self-attention layer. Without transfer learning, we establish a new state of the art on the popular TrecQA and WikiQA benchmark datasets. Additionally, we evaluate our method on four Community Question Answering datasets, and find that tree-structured representations have limitations with noisy user-generated text. We conduct probing experiments to evaluate how our models leverage tree structures across datasets. Our findings show that the ability of tree-structured models to successfully absorb syntactic information is strongly correlated with a higher performance in AS2.

Users of medical question answering systems often submit long and detailed questions, making it hard to achieve high recall in answer retrieval. To alleviate this problem, we propose a novel Multi-Task Learning (MTL) method with data augmentation for medical question understanding. We first establish an equivalence between the tasks of question summarization and Recognizing Question Entailment (RQE) using their definitions in the medical domain. Based on this equivalence, we propose a data augmentation algorithm to use just one dataset to optimize for both tasks, with a weighted MTL loss. We introduce gradually soft parameter-sharing: a constraint for decoder parameters to be close, that is gradually loosened as we move to the highest layer. We show through ablation studies that our proposed novelties improve performance. Our method outperforms existing MTL methods across 4 datasets of medical question pairs, in ROUGE scores, RQE accuracy and human evaluation. Finally, we show that our method fares better than single-task learning under 4 low-resource settings.

6076 Background: Remote patient monitoring (RPM) may improve the early detection and mitigation of cancer treatment-related complications, health-related outcomes and quality of life. RPM’s success may depend, in part, on patients’ adherence to remote monitoring protocols. However, factors that influence adherence to RPM are largely unknown. Daily blood pressure/pulse (BP/P), weight, and electronic patient-reported outcomes (ePROs) were monitored remotely in head and neck cancer (HNC) patients undergoing radiation treatment (RT) to identify dehydration risk. We evaluated potential factors associated with RPM adherence. Methods: During RT (average 6 to 7 weeks), participants were asked to take daily (Monday-Friday) measures of BP/P and weight using Bluetooth-enabled devices and to complete daily ePROs using a mobile tablet application (app). Data were provided to their physicians for daily review. The MD Anderson Symptom Inventory-Head and Neck (MDASI-HN) was completed at baseline and end of RT, and 6-8 weeks post-RT completion. The Patient Activation Measure (PAM) was completed at baseline and 6-8 weeks post-RT completion. A device usability survey measuring perceived usefulness of RPM was completed at the end of RT. Adherence to daily monitoring was recorded objectively. Longitudinal analyses compared the relationship between demographic, clinical, and PRO data and monitoring adherence. Results: Participants (n = 169) were 80% male, 87% White, and 91% married. Overall adherence to monitoring BP/P, weight, and ePROs was 83%, 82% and 74%, respectively. Greater HN-specific symptom severity and interference was associated with decreased adherence to daily monitoring of BP/P, weight, and ePROs (P< 0.021). Higher PAM scores were associated with higher adherence to daily monitoring of BP/P only (p = 0.006). Participants reported modest levels of perceived usefulness of RPM across four categories: symptom management, early problem detection, illness monitoring by healthcare provider, and feeling of security during RT. Only a single item indicating perceived feeling of security was associated with greater adherence to daily monitoring of blood pressure/pulse (p = 0.032) and weight (p = 0.007). Conclusions: A benefit of frequent RPM may be early detection and mitigation of symptoms during RT for HNC, however, increasing symptom burden experienced during treatment may interfere with adherence to daily monitoring. Better adherence may be attributed to patients perceiving a sense of security from daily monitoring and may suggest a potentially important value that patients gain from RPM. Understanding factors that impact patient adherence to RPM may help improve acceptability and clinical utility of RPM in oncology. Clinical trial information: NCT02253238.

CyberInfrastructures (CIs) are complex socio-technical-economical systems that are difficult to describe, design, analyze, and evaluate. As one example, E-Health CIs are patient-centric community-serving systems that have particular regulations for information security, governance, resource management, scalability, and maintainability. We have built multiple successful E-Health interdisciplinary projects at the University of California, San Diego, including CY berinfrastructure for CO mparative effectiveness REsearch (CYCORE) for cancer clinical trials, P ersonal A ctivity L ocation M easurement S ystem (PALMS) for studying human activity patterns, CitiSense for sensing city pollution, and I ntegrating D ata for A nalysis, A nonymization, and SH aring (iDASH), which is a National Center for Biomedical Computing. These CIs have served their user communities well with respect to their projects’ goals and durations. Due to their shared requirements, these systems were all designed around the service bus-based Rich Services software architecture. However, there are also significant differences in how these CIs were designed and implemented, which might be informative to the community of CI developers. In this chapter, we detail our experience with developing E-Health CIs from the software quality perspective. We start from generic CI requirements along with domain and application-specific requirements, and we analyze the tradeoffs of development and deployment choices and propose a blueprint for next generation CIs. The main question that we answer is: How can we improve the quality of the next CIs by capitalizing on lessons learned, the best architectural and implementation choices, and reusing elements from previously existing systems? We present three contributions. First, we identify a set of activities and attributes that enable us to assess the quality of a CI with regard to important maintenance activities that are common to all CIs. These activities and attributes and the way we perform the evaluation are generic and can be applied to other projects and domains. Second, we analyze multiple projects and identify a set of common requirements for CIs and applications in the E-Health domain. We separate these requirements into common requirements that apply to all CIs, E-Health specific requirements that apply to CIs and most applications in the Health domain, and application-specific requirements. This separation makes many requirements reusable, and we expect that they will inform the development of future E-Health CIs and applications. Finally, we identify a set of architectural and implementation recommendations that will help developers create higher quality CIs for the E-Health domain.

6063Background: Assessment and intervention with mobile and sensor technology may improve early detection and mitigation of treatment-related symptoms, impact quality of life (QOL), reduce complications, and lower health care costs. The CYCORE (CYberinfrastructure for COmparative effectiveness Research) system utilizes sensor and mobile technology to remotely assess daily weight, blood pressure (BP)/pulse, and patient-reported outcomes in head and neck cancer (HNC) patients undergoing radiation treatment (RT). Clinicians reviewed data daily to identify early risk of dehydration and support early intervention to improve symptom management. We compared longitudinal symptom data in patients randomized to use CYCORE during RT versus those randomized to usual care. Methods: Methods: HNC patients (n = 357) completed the 28-item MD Anderson Symptom Inventory (MDASI) at RT initiation (baseline), completion of RT (6-7 weeks post-baseline), and 6-8 weeks post-RT completion. Symptom severity and interference were ra...

The health-care sector has stringent needs for CyberInfrastructures (CIs) to improve both its patient-care capabilities and its potential for advanced clinical research. Large-scale projects reveal that economics, technology, science, and people need a careful balance to be successful. In this chapter, we detail our value-based methodology for architecting and implementing CIs for multiple projects in this field. We highlight the requirements and challenges presented by this field, the process for value-cost analysis, proper stakeholder involvement, and architecture evaluation to match a system’s timeline and budget. We also discuss lessons learned and typical CI aspects, including data ingestion and sharing, privacy and security, life-cycle management, rapid prototyping and user adoption, performance, and scalability, and their impact on project economics. We conclude with strategic factors for overall project success.

The potential of communication and health information technologies to transform research and improve clinical care in cancer has only begun to be realized (1,2). There is a need for innovative strategies and technologies that enable the aggregation and analysis of clinical and research data (including real-time health-related outcomes) and that foster the translation of knowledge gained into improvements in patient care (2–4). The use of health information technologies may alter how health-care providers and patients interact with personal health information by facilitating the collection, analysis, and dissemination of data to enhance clinical decision making (3). In turn, patients may report symptoms more readily and be reassured knowing that their providers are better informed about all aspects of their treatment and care. The American Society of Clinical Oncology has emphasized the need to use communication and information technology to better integrate clinical and translational research with patient care so that patients’ experiences can inform research and improve care (1). Accomplishing this goal requires a systems approach that enables greater and more rapid access to patient clinical information to better understand therapeutic responses, side effects, quality of life, and general health status (1). This approach may improve oncology care by synthesizing and distilling large amounts of data to identify earlier opportunities to manage treatment side effects and complications, as well as long-term survivorship issues (1,3). Systems that enable remote and real-time monitoring of patients’ symptoms and other health-related outcomes may offer cost-effective strategies to optimize cancer care outside of the clinic setting (5,6). Remote collection and transmission of patient health–related data to facilitate clinical decision making have been used in the management of chronic diseases other than cancer (7–12), and this has been shown to improve patient quality of life and symptom control, decrease emergency room visits and unplanned hospitalizations, and decrease overall health-care costs (12–15). Less is known about which patients and under which circumstances remote monitoring is most appropriate and effective, particularly in oncology (16,17). Studies evaluating remote real-time monitoring of cancer patients have reported positive outcomes (5); however, empiric evidence includes the collection of patient-reported outcomes largely using currently available telecommunications technologies (17,18). Systems that include biometric and self-reported feedback for cancer patients have been conceptualized; however, there is limited evidence regarding their feasibility or efficacy (19,20). This paper describes the results of an initial application of a novel system that incorporates data on environmental, psychological, treatment adherence, and lifestyle behaviors and the patient-reported outcomes into cancer treatment and research. The system, called CYCORE (CYberinfrastructure for COmparative effectiveness REsearch), is a software-based prototype of a user-friendly cyberinfrastructure that supports the acquisition, storage, visualization, analysis, and communication of data across the cancer continuum, using a suite of home-based and mobile sensors (19). Health-related data collected include symptoms, quality of life, performance status, and physiological parameters that can signal how patients are self-managing their treatment. The CYCORE prototype was rigorously tested in several diverse use cases that were grounded in challenging, real-world clinical problems in cancer prevention and treatment (19). One use case, which is the focus of this report, engaged head and neck cancer (HNC) patients in the use of home-based sensors for collecting and communicating blood pressure, pulse, weight, and patient-reported symptoms to their HNC radiation therapy (RT) clinicians. Clinicians reviewed these data outside of clinic visits for earlier identification of dehydration risk secondary to outpatient RT. HNCs are highly curable, yet primary RT is physically and emotionally challenging for patients (21). RT involves a 6- to 7-week course of 5 day/week treatments and causes serious locoregional side effects. In its severe form, mucositis, which occurs in up to 80% of irradiated HNC patients, makes it very difficult for patients to drink adequate levels of fluids (22). As a result, dehydration and its accompanying physiological changes can develop rapidly, often between weekly physician visits that are standard of care during RT. When unresolved, dehydration can lead to acute episodes requiring hospitalization and/or emergency care visits that are costly and potentially preventable. The number of patients affected is substantial, as up to 27% of HNC patients may be admitted to inpatient or emergency care units during RT (23–25) (personal communication, E. Shinn). Dehydration may represent one of the largest incremental health-care expenses in treating mucositis and related side effects for HNC (23). The development of systems to closely monitor hydration status has been suggested as a potentially valuable intervention that may reduce hospitalization and health-care costs, in addition to improving quality of life (26–28). Technology that supports remote monitoring of HNC patients during RT may enable greater access to clinical and patient-reported data that are currently unavailable, may provide greater insight into patients’ therapeutic response, and may help to identify opportunities for earlier and more effective intervention. In this context, CYCORE may serve as a new model of a communication architecture for improving clinical care and patient support outside of the clinic setting during critical periods of cancer treatment. This study evaluated the feasibility of using CYCORE to address early at-home identification of dehydration risk in HNC patients undergoing RT.

152 Background: Early identification and mitigation of cancer therapy-induced complications can improve quality of life and reduce complications and health care costs. We are evaluating the efficacy of a system, called CYCORE (CYberinfrastructure for Comparative effectiveness Research), to systematically and accurately collect daily weight, blood pressure, pulse and symptom data from head and neck cancer (HNC) survivors during their 6 to 7-week radiation treatment (RT) regimen using home-based biometric and other sensors integrated with a cyber-physical system. Aims include: 1) evaluate CYCORE’s efficacy in reducing hospitalization and emergency room visits related to dehydration; and, 2) evaluate related impact on costs related to treating dehydration. Methods: Remote, sensor-based monitoring via CYCORE identifies HNC survivors at increased risk for dehydration (i.e., reduced blood pressure and weight, increased pulse, and patient-reported outcomes (PROs) including decreased food and fluid intake). Upon initation of RT, HNC survivors are randomized to either: 1) standard care plus use of the CYCORE system, wherein biometric sensor data and PROs (e.g., pain, nausea, medications, food/fluid intake) are collected ; or, 2) standard care. Outcomes include hospital and emergency room admissions, and related costs. Results: To date, 113 survivors are enrolled (88.9% recruitment and 90.3% retention rates). Those randomized to CYCORE take daily home readings of weight, blood pressure and pulse via Bluetooth-enabled devices, and complete daily PRO assessments on a mobile application. Survivors’ data are available for viewing and clinical decision-making via a Web-based interface by their physicians and other providers. Previously, 59% of HNC survivors on RT showed physiologic signs of dehydration while using the CYCORE system. Worse ratings of nausea and vomiting predicted dehydration risk. Conclusions: CYCORE may serve as one model for integrating mobile and sensor technologies during critical periods of cancer care that occur outside of the clinic setting, and may improve management of treatment side effects, quality of life, and greater satisfaction with care. Clinical trial information: NCT02253238.

The Ocean Observatories Initiative (OOI) through its Cyberinfrastructure (CI) Implementing Organization is developing a next generation platform for ocean sciences that will integrate a wide variety of information resources at scales unattainable before in the earth and ocean sciences. We introduce a novel scientific data model that enables distributed, large-scale storage and query of science data. Our model is built on multiple levels of abstraction ranging from domain-specific at the top down to encodings for message-oriented transport and persistence at the base. The key is exposing the properties of scientific feature types separately from the underlying structure of the data, which in turn is separated from their representation. The data representation is further isolated from the serialization and encoding used for transport and persistence. Our model greatly simplifies expressions of provenance and versioning of various data entities. It is robust, scalable and reliable. We implemented it for the first release of the OOI Integrated Observatory Network (ION), with rollout to operations currently underway.