Combating "Referral Madness Syndrome"in Classical Hematology

Background: The Medical University of South Carolina (MUSC) serves a large geographic area. In recent years, MUSC has expanded its network to include hospitals across the state. We recently described the "referral madness syndrome" (Blood Adv (2023) 7 (21):6702-6704) as a major strain on the benign hematology workforce, practitioners, communications, and hospital system operations. There is a significant negative impact of high-volume, fast-paced medical decision-making that can lead to mistakes and poor outcomes. While the root cause of increasing referrals is complex, there is a need to provide a solution to the cognitive strain placed on physicians and to safely meet the needs of Classical Hematology patients.

Approach: We have implemented several strategies to assist with collecting referral information, scheduling, connecting to telehealth, dictating notes, placing orders, and communicating with patients. We report the nature of the resources required to manage this new clinical practice model and focus on implementing Telehealth services and AI to provide safe and effective care.

Results: The MUSC Health network serves the 30,000 square miles of South Carolina. MUSC hospitals are in urban and rural areas, with 8 MUSC Health Medical Centers in Columbia, Kershaw, Florence, Orangeburg, Marion, Lancaster, and Greenville. The distance from these facilities to Charleston varies from 75 to 180 miles. Charleston's tri-county population is 800,000 compared to 5.2 million statewide, with a net influx of 62,000 new residents moving to SC yearly. The distance between the major population centers in the state varies from 100 to 200 miles.

The telehealth program for Classical Hematology started in 2020 in response to the growing number of referrals, limited number of hematologists, and was accelerated by COVID-19. The total number of new patients increased from 628 in 2021 to 1,925 in 2023, while established patients grew from 776 to 1,340. Telephone-only visits decreased from 586 to 236. Total visits increased from 1,990 in 2021 to 8,210 including 2021-23 visits.

Screening has been a constant challenge, with 18.9% of patients scheduled with inadequate data to understand the nature of the referral. We recently hired a nurse to assist with screening. The no-show rate for in-person visits averaged 17.6% for 5 years before implementation. Since implementing Telehealth services, the no-show rate was 4.2%, and the same-day cancellation rate was 5.6%, totaling approximately 9.8%.

We utilize the ANDOR for secure telehealth visits, allowing up to two video patients per hour. The Claude 3.0 Pro version using Sonnet 3.5 is now used solely as an AI program to assist in correcting voice dictation, reducing the average number of corrections in voice-dictated notes from 11.8 per note to less than 0.5. The AI program was used to verify decision-making processes and provide additional rationale for testing and treatment options. A demonstration will be provided.

Limitations: Challenges exist in obtaining proper information prior to the referral, and this needs to be corrected to reduce no-shows and cancellations. AI is being developed to automate data collection and enable same-day addition of patients to the schedule.

Conclusion: While MUSC Health Charleston has a System-based Hematologist to manage their hospital-based needs, Telehealth and AI have the potential to assist in inpatient care at other facilities and harmonize care throughout the network. We have identified a major strain on the healthcare system related to referrals for common and rare hematological problems. By utilizing Telehealth and AI, we aim to educate the community and improve practice. Future plans include training medical students and hematology fellows in telehealth and AI use.