Development and Health System Deployment of an Electronic Health Record–Integrated Chatbot Intervention for Connecting Fall Risk Screening to Community Resources After Emergency Department Visits: Implementation Study
Bibliographic record
Abstract
BACKGROUND: Emergency departments (EDs) routinely screen for fall risk, but patients are rarely notified of their results or referred to preventive resources. There is a critical need for an intervention that notifies patients when they are at risk for falls and automates referrals to fall prevention programs without increasing clinician workload. Chatbots can be used to provide patient education and community resources in a conversational, friendly manner. We developed and implemented an automated intervention using our health system's electronic health record (EHR) and an artificial intelligence chatbot, Livi, to address this gap in fall prevention across 17 EDs. OBJECTIVE: This study aimed to share how we developed our fall risk notification and referral intervention and iteratively improved it based on end-user feedback. METHODS: We collaborated with the EHR and ED operations teams to automate patient notification of fall risk and referral. First, we leveraged existing fall risk screening questions in nursing documentation to identify patients at risk for falls. We then developed an EHR workflow that delivers a QR code in the after-visit summary for all high-risk patients at ED discharge. Scanning the QR code launches a conversation with Livi, guiding users to physician-vetted, evidence-based, free or low-cost fall prevention resources in their area. In this workflow, only ED patients who are screened as high risk receive linkage to Livi, and clinicians do not need to manually place referrals or enter specific fall prevention resources at discharge. We conducted rapid, iterative usability testing of the Livi falls chatbot with 93 community members during the development process at 3 community fairs in distinct settings. RESULTS: Rapid iterative testing led to enhancements in the intervention, such as increased font size, an option for Spanish language, additional geographic locations for fall prevention resources, home modification resources, the ability to self-assess for fall risk, fall prevention tips, and the ability for patients to leave feedback on the Livi chatbot. Because all EDs in the health system use the same instance of Epic, the EHR workflow was instantaneously deployed system-wide. The use of a QR code linked to the Livi chatbot also allows for the rapid updating of prevention resources. CONCLUSIONS: This study describes the formative development and system-wide implementation of the intervention. This scalable, EHR-integrated intervention demonstrates a novel and pragmatic approach to improving population health by capitalizing on existing clinical workflows and automating both risk notification and personalized resource referral for older adults without increasing clinician burden. The next steps include conducting a randomized controlled trial to assess the impact of the screening and referral tool on recurrent fall-related health care use compared with routine care in the ED. Formal evaluation of the implementation outcomes will be conducted in the planned trial.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.016 | 0.028 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.003 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".