2025 CLINICAL CHALLENGES

Projects should align with the aim under the case title, but may not need to address all aspects of the clinical vignette.

{Neuromuscular Monitoring}

Aim: Leverage AI to track, monitor, and manage Parkinson’s Disease 

  • Michael Hayes, a 65-year-old man with Parkinson’s disease, has been managing his symptoms for several years. Despite medication, he’s noticed a gradual worsening of his tremors, muscle stiffness, and cognitive function. He visits his neurologist regularly, but struggles to communicate the subtleties of his symptoms, which seem to fluctuate day-to-day. Michael is interested in an AI-powered health platform that can track changes in his motor skills and cognitive function over time, including speech patterns, handwriting, and cognitive test results (e.g. Mini Mental Cognitive Test, Montreal Cognitive Assessment, Clock Drawing Test). 

    This system should provide a detailed, long-term view of Michael’s symptoms, allowing his doctor to detect subtle changes and tailor interventions more effectively. Ultimately, he hopes for a continuous, data-driven approach that streamlines communication between him and his healthcare team, ensuring both stay aligned in managing his Parkinson’s disease more effectively.

{Schedule Optimization}

Aim: Smart scheduling for improved efficiency and patient outcome

  • Dr. Emily Hoang, a primary care physician at an outpatient clinic in Washington, DC, faces ongoing scheduling inefficiencies that impact both patient care and clinic operations. Patients still hospitalized are mistakenly booked for clinic follow-ups, while those with flu symptoms are scheduled for virtual visits when they need in-person care. Insurance approvals frequently cause last-minute rescheduling, and incorrect time slots for sick visits and wellness exams lead to wasted time or rushed care. No-shows remain a persistent issue, particularly among patients with transportation challenges or those who forget their appointments, making it harder for others to secure timely visits. These inefficiencies contribute to longer patient wait times, delays in necessary care, and difficulty accessing their doctor when needed.

    She envisions a scheduling system that prevents premature hospital follow-ups, verifies insurance ahead of time, and triages patients appropriately for in-person or virtual visits. To improve patient access and experience, she wants it to proactively address mobility issues and social needs—reminding elderly patients to bring forms and medications while optimizing appointment times. Most importantly, she hopes for a system that reduces wait times, increases the likelihood of receiving timely care, and ensures patients can see their doctor when they need to, minimizing scheduling disruptions that delay treatment.

{Call Scripts}

Aim: Improve efficiency by automating post-discharge follow-up script 

  • Jenna Thompson, an ICU nurse at MedStar Georgetown, balances a demanding workload—monitoring vitals, administering medications, and coordinating patient care. On top of this, she must ensure that recently discharged patients adhere to post-hospitalization instructions, take medications correctly, and schedule follow-up appointments. These follow-ups are crucial for patient safety, but the process is inefficient—nurses spend valuable time making repetitive calls, waiting on hold, and tracking down patients, all while juggling in-hospital responsibilities.

    Jenna envisions a system that automates routine follow-ups, minimizing the need for manual calls. An intelligent platform could send personalized messages with key discharge details—medication schedules, follow-up appointments, and self-care instructions—directly to patients. If a patient fails to confirm adherence, reports worsening symptoms, or misses an appointment, the system would immediately notify the urgent care team to follow up. By eliminating time-consuming calls and unnecessary hold times, this system would allow nurses to focus on in-hospital care while ensuring discharged patients receive the support they need.

{Mental Health}

Aim: AI companion that provides support and resources for healthcare professionals  

  • Sarah, a 28-year-old Internal Medicine resident, has been struggling with persistent anxiety, low mood, and overwhelming stress over the past few months. She finds it difficult to concentrate, experiences sleep disturbances, and feels emotionally drained by the demands of her work. Despite recognizing her distress, Sarah hesitates to seek professional help due to time constraints and uncertainty about where to start. She would greatly benefit from an AI-powered mental health companion that offers a comprehensive self-assessment of her mood, sleep patterns, and stress levels while providing personalized, easily accessible resources, guidance on finding a therapist who fits her schedule and understands the challenges of working in healthcare, and real-time support tailored to her needs and goals.

{Pediatric Health Education}

Aim: Empower parents with AI-driven interactive education to better manage their children's health

  • Mateo’s parents recently learned that their 7-year-old son has type 1 diabetes, but they struggle to fully understand how to manage his condition. They find medical jargon confusing, feel unprepared to handle emergencies, and worry about teaching Mateo lifelong self-care habits. While doctors provide guidance, the information can be difficult to retain, and pamphlets feel insufficient.

    They wonder if an AI-powered tool—such as personalized coaching, interactive simulations, or intuitive visual guides—could simplify complex medical concepts, offer real-time support, and empower them to make informed decisions. The ideal solution will enhance caregiver confidence, improve health education for common pediatric conditions—such as type 1 diabetes, asthma, food allergies, or epilepsy—and be accessible in multiple languages to support diverse families.

{Patient Safety Challenge}

Aim: Prevent medical errors and improve patient safety with a consumer-driven approach

  • Medical errors harm millions of U.S. patients each year, claim approximately 250,000 lives, and cost billions of dollars. These preventable mistakes—ranging from medication errors and infections to surgical and diagnostic safety failures—underscore the urgent need for bold new thinking in patient safety. While many solutions focus on healthcare providers, a critical gap remains: empowering patients and their families to actively prevent harm.

    The Patient Safety Technology Challenge calls on H2AI teams to build the most consumer-focused, innovative, and impactful solution that clearly understands the problem identified. Innovations must be consumer-driven and address one of the five leading patient safety challenges across the continuum of care: medication errors, procedural and surgical errors, errors during routine patient care (e.g., pressure ulcers, blood clots, falls), infections, and diagnostic safety.

{AI-powered Compliance}

Aim: Enhancing documentation accuracy with real-time error detection, regulatory guidance, and training

  • Dr. Laura Bennett, a Vice Chair of Education and Compliance in the Department of OB/GYN at a major Tennessee healthcare system, faces persistent challenges with sterilization consent forms required for Medicaid reimbursement under TennCare. The federally mandated paper form must be completed accurately and submitted with billing requests for sterilization procedures, but outdated workflows and manual processes across hospitals often lead to errors. When forms are missing or incorrectly filled out, claims are denied, requiring resubmission within a strict 90-day window. In some cases, entire hospital stays—including labor and delivery—go unpaid if the form is not corrected in time, leading to significant financial losses and delayed payments for providers.

    Dr. Bennett envisions an AI-powered training agent designed to help OB/GYN staff in Tennessee navigate these documentation challenges. By analyzing completed forms in real time and integrating key regulatory knowledge, AI could flag potential errors before submission, ensuring compliance with TennCare policies. Additionally, interactive training modules could simulate real-world billing scenarios, reinforcing best practices and standardizing workflows across hospitals. This AI-driven approach would improve accuracy, streamline resubmissions, and reduce financial risk—allowing healthcare providers to focus more on patient care while ensuring timely reimbursement.

    To facilitate development, participating teams can review the TennCare Medicaid Sterilization Consent Form here. This resource provides a detailed example of instructions on form completion and compliance requirements, serving as a foundational tool for staff education and AI-assisted verification.