Abstract Introduction Our program triages 3,000+ sleep referrals annually, across 11 sub-specialty clinic types. To ensure patients are scheduled into the best clinic for their sleep concerns, patient families complete an intake questionnaire. Historically, this process involved multiple steps: • Questionnaire sent to patient family upon receiving the referral. • Completed questionnaire received and forwarded to provider for review. • Provider returned triage recommendation for specific clinic. • Clinic recommendation sent to scheduling call center. • Call center contacted patient family and scheduled. This workflow was resource-intensive for successful completion. As such, we hypothesized that implementing an automated scoring algorithm would yield more predictable triage decisions, fewer staff touch points, and improved time to scheduling. Methods Based on the historical 43-item screening questionnaire, an abbreviated 13-item screening questionnaire with automated algorithmic logic to assign clinics was created. The new questionnaire was validated against historical clinical triage decisions across three, blinded providers. Quality Improvement tools and LEAN methodology to improve the triage process. The team worked to complete a current state and future state value stream map (VSM). Tasks to complete each step in the process were documented on the VSM. Each task and the wait time between tasks were timed and averaged for randomly selected patients. The team then brainstormed possible kaizen opportunities documenting them on the VSM. Results The historical process had a lead time of 174 hours and a process time of 12 minutes per referral and a 60% provider agreement on triaged subspeciality clinic. The redesigned process reduced lead time to 94 hours and process time to 10 minutes. Additionally, the provider agreement improved to 95% for the triaged subspecialty clinic. Conclusion This approach reduces process delays, minimizes provider bias, and staff touch points through an integrated scoring algorithm. By automating triage, we have achieved more consistent triage decision-making, improved patient and provider experience, and operational efficiencies. The success of this initiative demonstrates the potential for scalable automation in high-volume sleep referrals and fine tuning of triage algorithms to maximize referral efficiency and potentially paving the way for broader application across other specialties. Support (if any)
Hjelm et al. (2026) studied this question.