An artificially intelligent, natural language processing chatbot designed to promote COVID-19 vaccination: A proof-of-concept pilot study

Objective Our goal is to establish the feasibility of using an artificially intelligent chatbot in diverse healthcare settings to promote COVID-19 vaccination. Methods We designed an artificially intelligent chatbot deployed via short message services and web-based platforms. Guided by communication theories, we developed persuasive messages to respond to users' COVID-19-related questions and encourage vaccination. We implemented the system in healthcare settings in the U.S. between April 2021 and March 2022 and logged the number of users, topics discussed, and information on system accuracy in matching responses to user intents. We regularly reviewed queries and reclassified responses to better match responses to query intents as COVID-19 events evolved. Results A total of 2479 users engaged with the system, exchanging 3994 COVID-19 relevant messages. The most popular queries to the system were about boosters and where to get a vaccine. The system's accuracy rate in matching responses to user queries ranged from 54% to 91.1%. Accuracy lagged when new information related to COVID emerged, such as that related to the Delta variant. Accuracy increased when we added new content to the system. Conclusions It is feasible and potentially valuable to create chatbot systems using AI to facilitate access to current, accurate, complete, and persuasive information on infectious diseases. Such a system can be adapted to use with patients and populations needing detailed information and motivation to act in support of their health.

An artificially intelligent, natural language processing chatbot designed to promote COVID-19 vaccination: A proof-of-concept pilot study.

Objective: Our goal is to establish the feasibility of using an artificially intelligent chatbot in diverse healthcare settings to promote COVID-19 vaccination. Methods: We designed an artificially intelligent chatbot deployed via short message services and web-based platforms. Guided by communication theories, we developed persuasive messages to respond to users' COVID-19-related questions and encourage vaccination. We implemented the system in healthcare settings in the U.S. between April 2021 and March 2022 and logged the number of users, topics discussed, and information on system accuracy in matching responses to user intents. We regularly reviewed queries and reclassified responses to better match responses to query intents as COVID-19 events evolved. Results: A total of 2479 users engaged with the system, exchanging 3994 COVID-19 relevant messages. The most popular queries to the system were about boosters and where to get a vaccine. The system's accuracy rate in matching responses to user queries ranged from 54% to 91.1%. Accuracy lagged when new information related to COVID emerged, such as that related to the Delta variant. Accuracy increased when we added new content to the system. Conclusions: It is feasible and potentially valuable to create chatbot systems using AI to facilitate access to current, accurate, complete, and persuasive information on infectious diseases. Such a system can be adapted to use with patients and populations needing detailed information and motivation to act in support of their health.

Digital Facilitation to Support Patient Access to Web-Based Primary Care Services: Scoping Literature Review.

BACKGROUND: The use of web-based services within primary care (PC) in the National Health Service in England is increasing, with medically underserved populations being less likely to engage with web-based services than other patient groups. Digital facilitation-referring to a range of processes, procedures, and personnel that seek to support patients in the uptake and use of web-based services-may be a way of addressing these challenges. However, the models and impact of digital facilitation currently in use are unclear. OBJECTIVE: This study aimed to identify, characterize, and differentiate between different approaches to digital facilitation in PC; establish what is known about the effectiveness of different approaches; and understand the enablers of digital facilitation. METHODS: Adopting scoping review methodology, we searched academic databases (PubMed, EMBASE, CINAHL, Web of Science, and Cochrane Library) and gray literature published between 2015 and 2020. We conducted snowball searches of reference lists of included articles and articles identified during screening as relevant to digital facilitation, but which did not meet the inclusion criteria because of article type restrictions. Titles and abstracts were independently screened by 2 reviewers. Data from eligible studies were analyzed using a narrative synthesis approach. RESULTS: A total of 85 publications were included. Most (71/85, 84%) were concerned with digital facilitation approaches targeted at patients (promotion of services, training patients to improve their technical skills, or other guidance and support). Further identified approaches targeted PC staff to help patients (eg, improving staff knowledge of web-based services and enhancing their technical or communication skills). Qualitative evidence suggests that some digital facilitation may be effective in promoting the uptake and use of web-based services by patients (eg, recommendation of web-based services by practice staff and coaching). We found little evidence that providing patients with initial assistance in registering for or accessing web-based services leads to increased long-term use. Few studies have addressed the effects of digital facilitation on health care inequalities. Those that addressed this suggested that providing technical training for patients could be effective, at least in part, in reducing inequalities, although not entirely. Factors affecting the success of digital facilitation include perceptions of the usefulness of the web-based service, trust in the service, patients' trust in providers, the capacity of PC staff, guidelines or regulations supporting facilitation efforts, and staff buy-in and motivation. CONCLUSIONS: Digital facilitation has the potential to increase the uptake and use of web-based services by PC patients. Understanding the approaches that are most effective and cost-effective, for whom, and under what circumstances requires further research, including rigorous evaluations of longer-term impacts. As efforts continue to increase the use of web-based services in PC in England and elsewhere, we offer an early typology to inform conceptual development and evaluations. TRIAL REGISTRATION: PROSPERO International Prospective Register of Systematic Reviews CRD42020189019; https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=189019.

Randomized Clinical Trial of Inreach With or Without Mailed Outreach to Promote Hepatitis C Screening in a Difficult-to-Reach Patient Population.

INTRODUCTION: Hepatitis C virus (HCV) treatment can significantly reduce the risk of liver-related mortality; however, many patients remain unaware of their infection in clinical practice. The aim of this study is to compare the effectiveness of inreach, with and without mailed outreach, to increase HCV screening and follow-up in a large, difficult-to-reach patient population. METHODS: We conducted a pragmatic randomized clinical trial from August 2018 to May 2019 in a large safety-net health system. Patients born between 1945 and 1965 were randomly assigned (1:1) to inreach with an electronic health record reminder to providers (n = 6,195) or inreach plus mailed HCV screening outreach (n = 6,191) to complete HCV antibody screening. Outreach also included processes to promote HCV RNA testing among those with a positive HCV antibody and linkage to care among those with positive HCV RNA. The primary outcome was completion of HCV antibody testing within 3 months of randomization (ClinicalTrials.gov NCT03706742). RESULTS: We included 12,386 eligible patients (median age 60 years; 46.5% Hispanic, 33.0% Black, and 16.0% White). In intent-to-treat analyses, HCV screening completion was significantly higher among inreach-plus-outreach patients than inreach-alone patients at 3 months (14.6% vs 7.4%, P < 0.001) and 6 months (17.4% vs 9.8%, P < 0.001) after randomization. Among those who completed HCV screening within 6 months, a higher proportion of inreach-plus-outreach patients with positive antibody results completed RNA testing within 3 months than inreach-alone patients (81.1% vs 57.1%, respectively, P = 0.02); however, linkage to care within 3 months of HCV infection confirmation did not significantly differ between the 2 groups (48.1% vs 75.0%, respectively, P = 0.24). DISCUSSION: Among difficult-to-reach patients, a combination of inreach and mailed outreach significantly increased HCV screening compared with inreach alone. However, HCV screening completion in both arms remained low, highlighting a need for more intensive interventions.