Radically different
SAN FRANCISCO — OpenEvidence rode the wave of early enthusiasm for large language models by building a free chatbot for doctors. Physicians, especially trainees, have flocked to the platform to help make patient care decisions; the company claims that about 650,000 U.S. physicians use it actively.
In just four years, OpenEvidence has leveraged that popularity into a $12 billion valuation, going directly to clinicians and avoiding health tech’s traditional hospital procurement process. But the company is facing competitive pressures and questions about whether its ad-based business model can continue to propel the company forward.
At the STAT Breakthrough Summit West, OpenEvidence chief technology officer Zachary Ziegler said that while the company has grown by building for individual clinicians, it’s time to aim bigger.
Background: Osteoporosis poses a significant global health burden and is responsible for over 8.9 million fragility fractures annually. Despite evidence-based guidelines and treatment, a substantial care gap persists, with only a low percentage of fracture patients receiving guideline-concordant care. Primary care physicians (PCPs) are pivotal in community-based fracture prevention but face challenges in translating knowledge into practice. While hospital-based fracture liaison services are effective, their reach is limited, necessitating scalable alternatives. Virtual communities of practice and web-based learning tools offer promising avenues for PCP professional education; however, their application in osteoporosis management remains underexplored. The Community Fracture Capture (CFC) Learning Hub was developed as an interactive, case-based platform to address these gaps by enhancing PCPs’ knowledge, confidence, and engagement in osteoporosis care. Objective: The study aimed to conduct a qualitative evaluation of PCPs’ interactions with the CFC Learning Hub, focusing on barriers and facilitators of the online learning tool and exploring PCP perceptions of the program. Methods: A qualitative analysis was performed using data from 55 PCPs across four 6-week cycles of the CFC Learning Hub (May 2022-October 2023). Data sources included discussion forum comments and responses to open-ended questions in end-of-cycle evaluations. Relational content analysis was used, with 2 researchers independently coding data using semantic and latent approaches. Themes were identified through iterative discussions and validated against existing literature. Results: Four themes emerged from PCP interactions: (1) user experience–guided platform design, where participants emphasized intuitive navigation, minimized fragmented sections, and clarity of interface as critical for engagement; (2) learning-supportive course structure, highlighting the importance of explicit links between case studies and foundational knowledge, weekly summaries, and quizzes aligned with content; (3) learners’ different styles and preferences, with diverging needs for synchronous vs asynchronous learning, didactic sessions, and peer-to-peer interactions; and (4) program content, where participants requested expanded topics and postprogram refreshers. Conclusions: The CFC Learning Hub demonstrated efficacy as a specialist- and peer-to-peer–supported online learning model for PCPs, addressing osteoporosis care gaps through user-centered design, adaptable content delivery, and collaborative moderation. Key successes included resolving usability issues iteratively and accommodating diverse learning preferences. These findings underscore the potential of the Hub to enhance primary care professional education and fracture prevention. The study advocates for broader adoption of the platform to bridge osteoporosis care disparities.
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/d3f5093e9bc12109eae0449b55cb24ec" />
Background: In recent years, digital patient portals have become an increasingly common feature of care in various medical fields. Despite growing scientific evidence of their effectiveness and the benefits they offer to patients and caregivers, their implementation, especially in hospital mental health settings, lags behind expectations. Objective: The study aimed to identify the barriers and facilitators to implementing a patient portal in a public mental health hospital setting in Germany. Moreover, it aimed to develop recommendations for implementing a patient portal. Methods: Three psychiatric clinics in the early stages of implementing an online portal for patients participated in this implementation study. We assessed objective usage data (log data from the patient portal) and performed qualitative interviews with professionals and questionnaire surveys with both patients and professionals. We combined the results to develop generic recommendations for the implementation of patient portals in a mental hospital setting using a 2-stage Delphi method with a group of professionals and patients. Results: Portal log data from 71 patients indicated variation in the use of the portal functions. On average, users logged in 9.5 (SD 14.9) times (median 4, IQR 2-7 times). The variability in the number of logins per patient, ranging from 1 to 72, indicated a high variance in the frequency of use. On average, the portal was used for 47 (SD 59) days (median 27, IQR 2-62 days). Questionnaire data from 27 patients showed satisfaction with the portal and elucidated perceived barriers to usage. Qualitative interview data from 15 professionals revealed patient-related, professional-related, organizational, structural, and technical facilitators and barriers to the implementation process. We developed 10 actionable recommendations for the implementation of digital patient portals in psychiatric hospitals, which were rated by an expert group on different dimensions. Conclusions: To our knowledge, this is the first implementation study in a German mental health hospital setting that provides experience-based recommendations for advancing the implementation of digital patient portals in hospital mental health care. The next steps will include the analysis of a larger number of users and functions, which will help to specify recommendations for different target groups and settings. Trial Registration: German Clinical Trials Register DRKS00036894; https://drks.de/search/en/trial/DRKS00036894
<img src="https://jmir-production.s3.us-east-2.amazonaws.com/thumbs/ae6f58f3685b46339a09abf49a82cd13" />