Mitigating Patient and Consumer Safety Risks When Using Conversational Assistants for Medical Information: Exploratory Mixed Methods Experiment.

BACKGROUND: Prior studies have demonstrated the safety risks when patients and consumers use conversational assistants such as Apple's Siri and Amazon's Alexa for obtaining medical information. OBJECTIVE: The aim of this study is to evaluate two approaches to reducing the likelihood that patients or consumers will act on the potentially harmful medical information they receive from conversational assistants. METHODS: Participants were given medical problems to pose to conversational assistants that had been previously demonstrated to result in potentially harmful recommendations. Each conversational assistant's response was randomly varied to include either a correct or incorrect paraphrase of the query or a disclaimer message-or not-telling the participants that they should not act on the advice without first talking to a physician. The participants were then asked what actions they would take based on their interaction, along with the likelihood of taking the action. The reported actions were recorded and analyzed, and the participants were interviewed at the end of each interaction. RESULTS: A total of 32 participants completed the study, each interacting with 4 conversational assistants. The participants were on average aged 42.44 (SD 14.08) years, 53% (17/32) were women, and 66% (21/32) were college educated. Those participants who heard a correct paraphrase of their query were significantly more likely to state that they would follow the medical advice provided by the conversational assistant (χ21=3.1; P=.04). Those participants who heard a disclaimer message were significantly more likely to say that they would contact a physician or health professional before acting on the medical advice received (χ21=43.5; P=.001). CONCLUSIONS: Designers of conversational systems should consider incorporating both disclaimers and feedback on query understanding in response to user queries for medical advice. Unconstrained natural language input should not be used in systems designed specifically to provide medical advice.

The Atrial Fibrillation Health Literacy Information Technology Trial: Pilot Trial of a Mobile Health App for Atrial Fibrillation.

BACKGROUND: Atrial fibrillation (AF) is a common arrhythmia that adversely affects health-related quality of life (HRQoL). We conducted a pilot trial of individuals with AF using a smartphone to provide a relational agent as well as rhythm monitoring. We employed our pilot to measure acceptability and adherence and to assess its effectiveness in improving HRQoL and adherence. OBJECTIVE: This study aims to measure acceptability and adherence and to assess its effectiveness to improve HRQoL and adherence. METHODS: Participants were recruited from ambulatory clinics and randomized to a 30-day intervention or usual care. We collected baseline characteristics and conducted baseline and 30-day assessments of HRQoL using the Atrial Fibrillation Effect on Quality of Life (AFEQT) measure and self-reported adherence to anticoagulation. The intervention consisted of a smartphone-based relational agent, which simulates face-to-face counseling and delivered content on AF education, adherence, and symptom monitoring with prompted rhythm monitoring. We compared differences in AFEQT and adherence at 30 days, adjusted for baseline values. We quantified participants' use and acceptability of the intervention. RESULTS: A total of 120 participants were recruited and randomized (59 to control and 61 to intervention) to the pilot trial (mean age 72.1 years, SD 9.10; 62/120, 51.7% women). The control group had a 95% follow-up, and the intervention group had a 93% follow-up. The intervention group demonstrated significantly higher improvement in total AFEQT scores (adjusted mean difference 4.5; 95% CI 0.6-8.3; P=.03) and in daily activity (adjusted mean difference 7.1; 95% CI 1.8-12.4; P=.009) compared with the control between baseline and 30 days. The intervention group showed significantly improved self-reported adherence to anticoagulation therapy at 30 days (intervention 3.5%; control 23.2%; adjusted difference 16.6%; 95% CI 2.8%-30.4%; P<.001). Qualitative assessments of acceptability identified that participants found the relational agent useful, informative, and trustworthy. CONCLUSIONS: Individuals randomized to a 30-day smartphone intervention with a relational agent and rhythm monitoring showed significant improvement in HRQoL and adherence. Participants had favorable acceptability of the intervention with both objective use and qualitative assessments of acceptability.

Patient and Consumer Safety Risks When Using Conversational Assistants for Medical Information: An Observational Study of Siri, Alexa, and Google Assistant.

BACKGROUND: Conversational assistants, such as Siri, Alexa, and Google Assistant, are ubiquitous and are beginning to be used as portals for medical services. However, the potential safety issues of using conversational assistants for medical information by patients and consumers are not understood. OBJECTIVE: To determine the prevalence and nature of the harm that could result from patients or consumers using conversational assistants for medical information. METHODS: Participants were given medical problems to pose to Siri, Alexa, or Google Assistant, and asked to determine an action to take based on information from the system. Assignment of tasks and systems were randomized across participants, and participants queried the conversational assistants in their own words, making as many attempts as needed until they either reported an action to take or gave up. Participant-reported actions for each medical task were rated for patient harm using an Agency for Healthcare Research and Quality harm scale. RESULTS: Fifty-four subjects completed the study with a mean age of 42 years (SD 18). Twenty-nine (54%) were female, 31 (57%) Caucasian, and 26 (50%) were college educated. Only 8 (15%) reported using a conversational assistant regularly, while 22 (41%) had never used one, and 24 (44%) had tried one "a few times." Forty-four (82%) used computers regularly. Subjects were only able to complete 168 (43%) of their 394 tasks. Of these, 49 (29%) reported actions that could have resulted in some degree of patient harm, including 27 (16%) that could have resulted in death. CONCLUSIONS: Reliance on conversational assistants for actionable medical information represents a safety risk for patients and consumers. Patients should be cautioned to not use these technologies for answers to medical questions they intend to act on without further consultation from a health care provider.

Rationale and design of the Atrial Fibrillation health Literacy Information Technology Trial: (AF-LITT).

BACKGROUND: Atrial Fibrillation (AF) is a common cardiac arrhythmia that is challenging for patients and adversely impacts health-related quality of life (HRQoL). Long-term management of AF requires that patients adhere to complex therapies, understand difficult terminology, navigate subspecialty care, and have continued symptom monitoring with the goal of preventing adverse outcomes. Continued interventions to ameliorate the patient experience of AF are essential. DESIGN: The Atrial Fibrillation health Literacy Information Technology Trial (AF-LITT; NCT03093558) is an investigator-initiated, 2-arm randomized clinical trial (RCT). This RCT is a pilot in order to implement a novel, smartphone-based intervention to address the patient experience of AF. This pilot RCT will compare a combination of the Embodied Conversational Agent (ECA) and the Alive Cor Kardia Mobile heart rhythm monitor to the current standard of care. The study will enroll 180 adults with non-valvular AF who are receiving anticoagulation for stroke prevention and randomize them to receive a 30-day intervention (smartphone-based ECA/Kardia) or standard of care, which will include a symptom and adherence journal. The primary end-points are improvement in HRQoL and self-reported adherence to anticoagulation. The secondary end-points are the acceptability of the intervention to participants, its use by participants, and acceptability to referring physicians. CONCLUSIONS: The AF-LITT pilot aims to evaluate the efficacy of the ECA/Kardia to improve HRQoL and anticoagulant adherence, and to guide its implementation in a larger, multicenter clinical trial. The intervention has potential to improve HRQoL, adherence, and health care utilization in individuals with chronic AF.

Engaging women with an embodied conversational agent to deliver mindfulness and lifestyle recommendations: A feasibility randomized control trial.

OBJECTIVE: This randomized controlled trial evaluates the feasibility of using an Embodied Conversational Agent (ECA) to teach lifestyle modifications to urban women. METHODS: Women were randomized to either 1) an ECA (content included: mindfulness, stress management, physical activity, and healthy eating) or 2) patient education sheets mirroring same content plus a meditation CD/MP3 once a day for one month. General outcome measures included: number of stress management techniques used, physical activity levels, and eating patterns. RESULTS: Sixty-one women ages 18 to 50 were enrolled. On average, 51% identified as white, 26% as black, 23% as other races; and 20% as Hispanic. The major stress management techniques reported at baseline were: exercise (69%), listening to music (70%), and social support (66%). After one month, women randomized to the ECA significantly decreased alcohol consumption to reduce stress (p=0.03) and increased daily fruit consumption by an average of 2 servings compared to the control (p=0.04). CONCLUSION: It is feasible to use an ECA to promote health behaviors on stress management and healthy eating among diverse urban women. PRACTICE IMPLICATIONS: Compared to patient information sheets, ECAs provide promise as a way to teach healthy lifestyle behaviors to diverse urban women.

Perceived organizational affiliation and its effects on patient trust: Role modeling with embodied conversational agents.

OBJECTIVE: Verbal and non-verbal behaviors, which are known as "relational contextualization cues", relay information about relationships and how they are structured. We developed a computer-simulated provider conducting an informed consent process for clinical research to investigate the effects of a provider's alignment of interests with a patient, the research team, or a neutral party on patient trust in the provider. METHODS: Participants (N=43) interacted with a simulated provider for a research informed consent process in a three-arm, counterbalanced, within-subjects experiment. Participants reported their trust in the simulated provider after each treatment. RESULTS: Participants successfully recognized the alignment manipulation, and perceived the patient-aligned provider as more trustworthy than the other providers. Participants were also more satisfied with the patient-aligned provider, liked this provider more, expressed more desire to continue working with this provider, and stated that they were significantly more likely to sign the consent form after interacting with this provider compared to the other two. CONCLUSION: Relational contextualization that aligns with the patient increases trust, satisfaction, and willingness to enroll in the context of research informed consent. PRACTICE IMPLICATIONS: Health providers should align themselves with patients' interests.

The Atrial Fibrillation Health Literacy Information Technology System: Pilot Assessment.

BACKGROUND: Atrial fibrillation (AF) is a highly prevalent heart rhythm condition that has significant associated morbidity and requires chronic treatment. Mobile health (mHealth) technologies have the potential to enhance multiple aspects of AF care, including education, monitoring of symptoms, and encouraging and tracking medication adherence. We have previously implemented and tested relational agents to improve outcomes in chronic disease and sought to develop a smartphone-based relational agent for improving patient-centered outcomes in AF. OBJECTIVE: The objective of this study was to pilot a smartphone-based relational agent as preparation for a randomized clinical trial, the Atrial Fibrillation Health Literacy Information Technology Trial (AF-LITT). METHODS: We developed the relational agent for use by a smartphone consistent with our prior approaches. We programmed the relational agent as a computer-animated agent to simulate a face-to-face conversation and to serve as a health counselor or coach specific to AF. Relational agent's dialogue content, informed by a review of literature, focused on patient-centered domains and qualitative interviews with patients with AF, encompassed AF education, common symptoms, adherence challenges, and patient activation. We established that the content was accessible to individuals with limited health or computer literacy. Relational agent content coordinated with use of the smartphone AliveCor Kardia heart rate and rhythm monitor. Participants (N=31) were recruited as a convenience cohort from ambulatory clinical sites and instructed to use the relational agent and Kardia for 30 days. We collected demographic, social, and clinical characteristics and conducted baseline and 30-day assessments of health-related quality of life (HRQoL) with the Atrial Fibrillation Effect on Quality of life (AFEQT) measure; self-reported medication adherence with the Morisky 8-item Medication Adherence Scale (MMAS-8); and patient activation with the Patient Activation Measure (PAM). RESULTS: Participants (mean age 68 [SD 11]; 39% [12/31] women) used the relational agent for an average 17.8 (SD 10.0) days. The mean number of independent log-ins was 19.6 (SD 10.7), with a median of 20 times over 30 days. The mean number of Kardia uses was 26.5 (SD 5.9), and participants using Kardia were in AF for 14.3 (SD 11.0) days. AFEQT scores improved significantly from 64.5 (SD 22.9) at baseline to 76.3 (SD 19.4) units at 30 days (P<.01). We observed marginal but statistically significant improvement in self-reported medication adherence (baseline: 7.3 [SD 0.9], 30 days: 7.7 [SD 0.5]; P=.01). Assessments of acceptability identified that most of the participants found the relational agent useful, informative, and trustworthy. CONCLUSIONS: We piloted a 30-day smartphone-based intervention that combined a relational agent with dedicated content for AF alongside Kardia heart rate and rhythm monitoring. Pilot participants had favorable improvements in HRQoL and self-reported medication adherence, as well as positive responses to the intervention. These data will guide a larger, enhanced randomized trial implementing the smartphone relational agent and the Kardia monitor system.

Disparities in Age-Associated Cognitive Decline Between African-American and Caucasian Populations: The Roles of Health Literacy and Education.

OBJECTIVES: To examine health literacy as a mediator of racial disparities in cognitive decline as measured by executive function in elderly adults. DESIGN: Prospective cohort study. SETTING: Secondary analysis of ElderWalk trial in Boston, Massachusetts. PARTICIPANTS: English-speaking African-American and Caucasian individuals in a walking intervention for community-dwelling adults aged 65 and older without dementia at baseline who completed baseline and 12-month evaluations (N = 198). MEASUREMENTS: Health literacy was measured using the Short Test of Functional Health Literacy in Adults. Fluid and crystallized cognitive functions were measured at baseline and 12 months using the Trail-Making Test Part B minus Part B (TMT B-A) and the Controlled Oral Word Association Test (COWAT). Associations between health literacy and 12-month cognitive decline were modeled using multivariate linear regression. RESULTS: Participants with higher health literacy and education experienced less cognitive decline than those with limited health literacy according to the TMT B-A (P = .01). After adjusting for covariates, Caucasian participants (n = 63) experienced less decline than African-American participants (n = 135) on TMT B-A (P = .001) and COWAT (P = .001). Adjusting for health literacy led to a 25.3% decrease in the point estimate for racial difference in TMT B-A and a 19.5% decrease in COWAT. Although independently related to cognitive decline, educational attainment did not mediate racial differences. CONCLUSION: Health literacy is a partial mediator of racial disparities in cognitive decline. These results indicate the need to develop interventions to mitigate cognitive decline that individuals with low heath literacy can use and to modify the healthcare environment to better accommodate this population.

Improving Access to Online Health Information With Conversational Agents: A Randomized Controlled Experiment.

BACKGROUND: Conventional Web-based search engines may be unusable by individuals with low health literacy for finding health-related information, thus precluding their use by this population. OBJECTIVE: We describe a conversational search engine interface designed to allow individuals with low health and computer literacy identify and learn about clinical trials on the Internet. METHODS: A randomized trial involving 89 participants compared the conversational search engine interface (n=43) to the existing conventional keyword- and facet-based search engine interface (n=46) for the National Cancer Institute Clinical Trials database. Each participant performed 2 tasks: finding a clinical trial for themselves and finding a trial that met prespecified criteria. RESULTS: Results indicated that all participants were more satisfied with the conversational interface based on 7-point self-reported satisfaction ratings (task 1: mean 4.9, SD 1.8 vs mean 3.2, SD 1.8, P<.001; task 2: mean 4.8, SD 1.9 vs mean 3.2, SD 1.7, P<.001) compared to the conventional Web form-based interface. All participants also rated the trials they found as better meeting their search criteria, based on 7-point self-reported scales (task 1: mean 3.7, SD 1.6 vs mean 2.7, SD 1.8, P=.01; task 2: mean 4.8, SD 1.7 vs mean 3.4, SD 1.9, P<.01). Participants with low health literacy failed to find any trials that satisfied the prespecified criteria for task 2 using the conventional search engine interface, whereas 36% (5/14) were successful at this task using the conversational interface (P=.05). CONCLUSIONS: Conversational agents can be used to improve accessibility to Web-based searches in general and clinical trials in particular, and can help decrease recruitment bias against disadvantaged populations.

Racial disparities in patient activation: Evaluating the mediating role of health literacy with path analyses.

OBJECTIVE: The aim of this study was to use path analysis methods to determine if health literacy mediates the relationship between race and patient activation. METHODS: We performed a secondary analysis of data from a randomized controlled trial in elderly, urban, minority patients using path analysis. Path analysis was used to evaluate the mediation of race on patient activation through different variables. Several models were tested for best fit for their effects on patient activation. RESULTS: Across all models, significant mediation paths were identified from race to lower patient activation through health literacy. This relationship remained significant throughout alternative model testing for covariate combinations. The best-fit model included an indirect effect of sex on patient activation through health literacy indicating that the mediation effect of health literacy on patient activation was most profound for African American males. Health literacy had a bigger influence on patient activation for participants with a greater comorbidity than for those with fewer conditions. No significant direct effect was shown between race and patient activation in any of the models. CONCLUSION: Racial disparities in patient activation were fully mediated by health literacy skills. PRACTICE IMPLICATIONS: Future interventions to improve racial disparities in patient activation need to be targeted at improving health literacy.

Automated indexing of Internet stories for health behavior change: weight loss attitude pilot study.

BACKGROUND: Automated health behavior change interventions show promise, but suffer from high attrition and disuse. The Internet abounds with thousands of personal narrative accounts of health behavior change that could not only provide useful information and motivation for others who are also trying to change, but an endless source of novel, entertaining stories that may keep participants more engaged than messages authored by interventionists. OBJECTIVE: Given a collection of relevant personal health behavior change stories gathered from the Internet, the aim of this study was to develop and evaluate an automated indexing algorithm that could select the best possible story to provide to a user to have the greatest possible impact on their attitudes toward changing a targeted health behavior, in this case weight loss. METHODS: An indexing algorithm was developed using features informed by theories from behavioral medicine together with text classification and machine learning techniques. The algorithm was trained using a crowdsourced dataset, then evaluated in a 2×2 between-subjects randomized pilot study. One factor compared the effects of participants reading 2 indexed stories vs 2 randomly selected stories, whereas the second factor compared the medium used to tell the stories: text or animated conversational agent. Outcome measures included changes in self-efficacy and decisional balance for weight loss before and after the stories were read. RESULTS: Participants were recruited from a crowdsourcing website (N=103; 53.4%, 55/103 female; mean age 35, SD 10.8 years; 65.0%, 67/103 precontemplation; 19.4%, 20/103 contemplation for weight loss). Participants who read indexed stories exhibited a significantly greater increase in self-efficacy for weight loss compared to the control group (F1,107=5.5, P=.02). There were no significant effects of indexing on change in decisional balance (F1,97=0.05, P=.83) and no significant effects of medium on change in self-efficacy (F1,107=0.04, P=.84) or decisional balance (F1,97=0.78, P=.38). CONCLUSIONS: Personal stories of health behavior change can be harvested from the Internet and used directly and automatically in interventions to affect participant attitudes, such as self-efficacy for changing behavior. Such approaches have the potential to provide highly tailored interventions that maximize engagement and retention with minimal intervention development effort.

Health literacy and usability of clinical trial search engines.

Several web-based search engines have been developed to assist individuals to find clinical trials for which they may be interested in volunteering. However, these search engines may be difficult for individuals with low health and computer literacy to navigate. The authors present findings from a usability evaluation of clinical trial search tools with 41 participants across the health and computer literacy spectrum. The study consisted of 3 parts: (a) a usability study of an existing web-based clinical trial search tool; (b) a usability study of a keyword-based clinical trial search tool; and (c) an exploratory study investigating users' information needs when deciding among 2 or more candidate clinical trials. From the first 2 studies, the authors found that users with low health literacy have difficulty forming queries using keywords and have significantly more difficulty using a standard web-based clinical trial search tool compared with users with adequate health literacy. From the third study, the authors identified the search factors most important to individuals searching for clinical trials and how these varied by health literacy level.

Limited health literacy and decline in executive function in older adults.

Limited health literacy is associated with worse executive function, but the association between limited health literacy and decline in executive function has not been established because of a lack of longitudinal studies. The authors aimed to examine this association by studying a prospective cohort in the setting of a randomized controlled trial to promote walking in older adults. Participants were community-dwelling older adults (65 years of age or older) who scored 2 or more on the Mini-Cog, without depression (score of less than 15 on the 9-item Patient Health Questionnaire), and who completed baseline and 12-month evaluations (n = 226). Health literacy was measured using the Short Test of Functional Health Literacy in Adults. Executive function measured at baseline and 12 months using the Trail Making Test (TMT), Controlled Oral Word Association Test, and Category Fluency. The associations between health literacy and 12-month decline in each test of executive function were modeled using multivariate linear regression. Health literacy was found to be limited in 37% of participants. Limited health literacy was associated with reduced performance on all 3 executive function tests. In fully adjusted models, limited health literacy was associated with greater 12-month decline in performance on the TMT than higher health literacy (p = .01). In conclusion, older adults with limited health literacy are at risk for more rapid decline in scores on the TMT, a measure of executive function.

A randomized controlled trial of an automated exercise coach for older adults.

OBJECTIVES: To compare the efficacy of a computer-based physical activity program (Embodied Conversational Agent-ECA) with that of a pedometer control condition in sedentary older adults. DESIGN: Single-blind block-randomized controlled trial stratified according to clinic site and health literacy status. SETTING: Three urban ambulatory care practices at Boston Medical Center between April 2009 and September 2011. PARTICIPANTS: Older adults (N = 263; mean age 71.3; 61% female; 63% African American; 51% high school diploma or less). INTERVENTION: ECA participants were provided with portable tablet computers with touch screens to use for 2 months and were directed to connect their pedometers to the computer using a data cable and interact with a computer-animated virtual exercise coach daily to discuss walking and to set walking goals. Intervention participants were then given the opportunity to interact with the ECA in a kiosk in their clinic waiting room for the following 10 months. Control participants were given a control pedometer intervention that only tracked step counts for an equivalent period of time. Intervention participants were also provided with pedometers. MEASUREMENTS: The primary outcome was average daily step count for the 30 days before the 12-month interview. Secondary outcomes were average daily step count for the 30 days before the 2-month interview. Outcomes were also stratified according to health literacy level. RESULTS: ECA participants walked significantly more steps than control participants at 2 months (adjusted mean 4,041 vs 3,499 steps/day, P = .01), but this effect waned by 12 months (3,861 vs 3,383, P = .09). For participants with adequate health literacy, those in the ECA group walked significantly more than controls at both 2 months (P = .03) and 12 months (P = .02), while those with inadequate health literacy failed to show significant differences between treatment groups at either time point. Intervention participants were highly satisfied with the program. CONCLUSION: An automated exercise promotion system deployed from outpatient clinics increased walking among older adults over the short-term. Effective methods for long-term maintenance of behavior change are needed.

Employing virtual advisors in preventive care for underserved communities: results from the COMPASS study.

Electronically delivered health promotion programs that are aimed primarily at educated, health-literate individuals have proliferated, raising concerns that such trends could exacerbate health disparities in the United States and elsewhere. The efficacy of a culturally and linguistically adapted virtual advisor that provides tailored physical activity advice and support was tested in low-income older adults. Forty inactive adults (92.5% Latino) 55 years of age and older were randomized to a 4-month virtual advisor walking intervention or a waitlist control. Four-month increases in reported minutes of walking/week were greater in the virtual advisor arm (mean increase = 253.5 ± 248.7 minutes/week) relative to the control (mean increase = 26.8 ± 67.0 minutes/week; p = .0008). Walking increases in the virtual advisor arm were substantiated via objectively measured daily steps (slope analysis p = .002). All but one intervention participant continued some interaction with the virtual advisor in the 20-week poststudy period (mean number of poststudy sessions = 14.0 ± 20.5). The results indicate that a virtual advisor delivering culturally and linguistically adapted physical activity advice led to meaningful 4-month increases in walking relative to control among underserved older adults. This interactive technology, which requires minimal language and computer literacy, may help reduce health disparities by ensuring that all groups benefit from e-health opportunities.

Automated interventions for multiple health behaviors using conversational agents.

OBJECTIVE: An automated health counselor agent was designed to promote both physical activity and fruit and vegetable consumption through a series of simulated conversations with users on their home computers. METHODS: The agent was evaluated in a 4-arm randomized trial of a two-month daily contact intervention comparing: (a) physical activity; (b) fruit and vegetable consumption; (c) both interventions; and (d) a non-intervention control. Physical activity was assessed using daily pedometer steps. Daily servings of fruit and vegetables were assessed using the NIH/NCI self-report Fruit and Vegetable Scan. RESULTS: Participants in the physical activity intervention increased their walking on average compared to the control group, while those in the fruit and vegetable intervention and combined intervention decreased walking. Participants in the fruit and vegetable intervention group consumed significantly more servings per day compared to those in the control group, and those in the combined intervention reported consuming more compared to those in the control group. CONCLUSION: Automated health intervention software designed for efficient re-use is effective at changing health behavior. PRACTICE IMPLICATIONS: Automated health behavior change interventions can be designed to facilitate translation and adaptation across multiple behaviors.

Feasibility of a virtual exercise coach to promote walking in community-dwelling persons with Parkinson disease.

OBJECTIVE: The short-term benefits of exercise for persons with Parkinson disease (PD) are well established, but long-term adherence is limited. The aim of this study was to explore the feasibility, acceptability, and preliminary evidence of the effectiveness of a virtual exercise coach to promote daily walking in community-dwelling persons with Parkinson disease. DESIGN: Twenty subjects with Parkinson disease participated in this phase 1, single-group, nonrandomized clinical trial. The subjects were instructed to interact with the virtual exercise coach for 5 mins, wear a pedometer, and walk daily for 1 mo. Retention rate, satisfaction, and interaction history were assessed at 1 mo. Six-minute walk and gait speed were assessed at baseline and after the intervention. RESULTS: Fifty-five percent of the participants were women, and the mean age was 65.6 yrs. At the study completion, there was 100% retention rate. The subjects had a mean satisfaction score of 5.6/7 (with 7 indicating maximal satisfaction) with the virtual exercise coach. Interaction history revealed that the participants logged in for a mean (SD) of 25.4 (7) days of the recommended 30 days. The mean adherence to daily walking was 85%. Both gait speed and the 6-min walk test significantly improved (P < 0.05). No adverse events were reported. CONCLUSIONS: Sedentary persons with Parkinson disease successfully used a computer and interacted with a virtual exercise coach. Retention, satisfaction, and adherence to daily walking were high for 1 mo, and significant improvements were seen in mobility.

The role of information technology in health literacy research.

Without concerted effort, the current explosion in health information technology will further widen the digital health divide for individuals with inadequate health literacy. However, with focused investment of time and energy, technology has the potential for reducing disparities through intelligent, usable, and accessible systems that tailor information, advice, counseling, and behavioral support to an individual's need at a given time and place.

A reusable framework for health counseling dialogue systems based on a behavioral medicine ontology.

Automated approaches to promoting health behavior change, such as exercise, diet, and medication adherence promotion, have the potential for significant positive impact on society. We describe a theory-driven computational model of dialogue that simulates a human health counselor who is helping his or her clients to change via a series of conversations over time. Applications built using this model can be used to change the health behavior of patients and consumers at low cost over a wide range of media including the web and the phone. The model is implemented using an OWL ontology of health behavior change concepts and a public standard task modeling language (ANSI/CEA-2018). We demonstrate the power of modeling dialogue using an ontology and task model by showing how an exercise promotion system developed in the framework was re-purposed for diet promotion with 98% reuse of the abstract models. Evaluations of these two systems are presented, demonstrating high levels of fidelity to best practices in health behavior change counseling.

Usability of conversational agents by patients with inadequate health literacy: evidence from two clinical trials.

Embodied Conversational Agents (ECA) are computer-animated characters that simulate face-to-face conversation with patients. These agents can be programmed with best practices in human-human health communication and used for automated health education and behavior change counseling interventions. Evidence is presented from two ongoing clinical trials demonstrating that patients at different levels of health literacy find these agents acceptable and easy to use for automated health communication interventions. Innovative computer interface systems can be used to ensure that inadequate health literacy not serve as a barrier to interventions using health information technology.