AI-Enabled Diagnostic Imaging Requisition, Protocoling and Scheduling Workflow

Problem Statement:

Demand for diagnostic imaging continues to grow while radiologist, technologist and clerical capacity remains constrained. Every scheduled CT and MRI exam depends on a chain of manual steps before the patient reaches the scanner. The requisition is received, often by fax or on paper. It is reviewed for missing or unclear information, routed to the right modality, protocolled by a radiologist, assigned a priority, booked into an appropriate slot, and followed up with the patient. Each step is a handoff where delay, variation or error can occur.

Protocoling is a recognized bottleneck. At many sites, radiologists still review paper requisitions in batches, and clerks must then re-enter signed protocols into the radiology information system before scheduling can happen. Inconsistent protocol and priority selection contributes to repeat exams, avoidable contrast use, and variable image quality. Upstream, incomplete requisitions trigger clarification loops with referring clinicians. Downstream, booking offices field high volumes of calls about referral status, preparation instructions and rescheduling.

The opportunity is not simply to hire more radiologists and booking staff, which is a slow and costly lever, but to increase the effective capacity of the existing imaging workforce by automating and standardizing the front end of the imaging pathway, without changing who is accountable for clinical decisions. The participating Edges are seeking a solution that pre-processes requisitions, recommends protocols and priorities for human review, and returns completed outputs to existing hospital systems. For Edges seeking broader scope, the solution should also extend automation into requisition intake, scheduling and patient communication.

Objectives:

• Digitize and standardize the CT and MRI protocoling workflow, with a structured, prioritized radiologist worklist and protocol recommendations for radiologist review and sign-off
• Reduce the time from requisition receipt to completed protocol and to booking-ready status
• Identify incomplete, conflicting or unclear requisitions early and route them for clarification, with tracked resolution
• Apply configurable local protocol rules and provincial wait-time priority frameworks consistently
• Integrate with existing EMR and RIS environments through standards-based interfaces, returning outputs into existing workflows rather than creating parallel ones
• Maintain human-in-the-loop control and a complete audit trail of recommendations, overrides and final decisions
• For Edges seeking end-to-end scope, automate intake of faxed, paper and electronic requisitions, rules-based scheduling into appropriate slots, and patient and booking office communication
• Provide reporting that measures turnaround, workload, concordance and rework, so Edges can build an evidence-based procurement business case

Essential (mandatory) outcomes:

• Reduced protocoling turnaround time, from requisition receipt to completed protocol
• Reduced radiologist and technologist time spent on routine protocoling
• Reduced manual handling, re-keying and staff touches per requisition
• More consistent protocol and priority selection, with high concordance against expert review
• Fewer requisitions returned for missing or unclear information, and fewer repeat or corrected exams
• No increase in inappropriate routing or protocol errors, with all clinical decisions remaining with accountable clinicians
• Reliable integration with existing hospital systems and full traceability of every requisition
• A validated, replicable workflow and integration pattern that other hospitals, health authorities and imaging providers can adopt

Required capabilities (all respondents):

• Receipt of imaging orders and requisitions from hospital systems, presented to radiologists in a structured, prioritized protocoling worklist
• Configurable protocol templates by modality, body region, contrast use and acquisition parameters, with AI-assisted or rules-based protocol recommendations
• Radiologist review, edit, override and electronic sign-off, with the signed protocol returned to downstream RIS, EMR, scheduling and PACS workflows without manual re-entry
• Priority recommendations aligned to provincial wait-time frameworks and local rules
• Standards-based integration (HL7v2, FHIR, DICOM, and vendor-supported APIs where available) with Epic, including Epic Radiant, and with other hospital EMR and RIS environments
• Human-in-the-loop design, with no autonomous final protocol or priority assignment
• Complete audit trail of source data, recommendations, reviewer actions, overrides and final outcomes
• Support for non-production testing and a shadow-mode validation phase before live clinical use
• Ontario privacy and security compliance (PHIPA) and other applicable provincial requirements, Canadian data residency (Canadian cloud or on-premises), role-based access controls, encryption and audit logging
• Operational reporting covering volume, queue age, protocoling turnaround, concordance and overrides

Additional capability modules:

Some participating Edges require one or more of the following modules in addition to the required capabilities above. Respondents are not required to offer every module but should clearly state which modules their solution supports today and which are on their roadmap.

• Module A, Requisition intake and validation: automated intake of faxed, paper and electronic requisitions; extraction and normalization of patient, exam and clinical indication data; detection of missing or conflicting information; routing to the correct modality or work queue; and tracked communication with referring clinicians until resolved
• Module B, Scheduling: rules-based scheduling of outpatient CT, MRI, ultrasound and mammography exams into appropriate appointment slots within the Edge’s existing scheduling system
• Module C, Patient communication and booking office call management: automated booking confirmations, preparation instructions, status updates and rescheduling, plus support for inbound booking office calls

Complementary / Nice to have features:

• Appropriateness decision support based on recognized imaging guidance (for example, CAR or ACR criteria)
• Ability to extend protocoling to emergency department and outpatient requisitions, and to additional modalities such as ultrasound and nuclear medicine/PET
• Configurable dashboards for imaging operations leaders
• Patient-facing multilingual communication

Out of Scope:

• Image interpretation, diagnostic reporting, or AI-generated radiology reports
• Autonomous final protocol, priority or appropriateness decisions without clinician review
• Replacement of EMR, RIS, PACS or enterprise scheduling systems
• Changes to imaging modality hardware
• General X-ray and other unscheduled, walk-in imaging services
• Automation of in-person patient registration
• Secondary use, external model training, or commercialization of Edge data without separate approvals and agreements

Multiple CAN Health Edges have identified this challenge and are participating in this Call for Innovation. Their environments differ. Some operate Epic, including Epic Radiant, while others use a separate RIS and EMR. Some still receive most requisitions by fax or on paper, while others receive electronic orders that still require manual review before booking. Their scope also differs. Some Edges are seeking a protocoling solution only, some want protocoling combined with requisition intake and validation, and others are seeking an end-to-end platform covering intake, protocoling, scheduling and patient communication. Each Edge will run its own commercialization project and may select different companies or modules based on fit.

Edge-reported estimates illustrate the scale of the problem:

• Radiologists at one Edge can spend over an hour each morning protocoling paper requisitions, which clerks then re-enter into the RIS
• One Edge receives over 100,000 scheduled CT, MRI, ultrasound and mammography requisitions per year, with booking offices handling 60 to 80 calls per area per day

This pattern reflects a broader national challenge. Protocoling and requisition management are required at every hospital and imaging provider that offers scheduled CT and MRI, across hundreds of sites in Canada. Provincial investment in central intake and wait-time strategies for CT and MRI further increases the need for consistent, auditable intake and prioritization.

This gap exists largely because radiologist and booking staff time, not scanner capacity alone, is the binding constraint, and most imaging workflows still assume a person must touch every requisition at every step. At the same time, available technology tends to sit at one of two extremes: core EMR and RIS platforms built for ordering and scheduling that do not pre-process requisitions or recommend protocols, or point solutions that address a single step as a bolt-on tool outside the core clinical system. Neither supports the integrated, human-in-the-loop workflow the participating Edges are seeking.

Without a connected workflow, imaging departments will continue to absorb growing demand through manual work, patients will continue to wait longer than necessary for scans, and referring clinicians will continue to chase status updates. A validated solution would be replicable well beyond the initial sites, including additional hospital sites, emergency and outpatient protocoling streams, additional modalities, regional central intake programs, and community imaging clinics. Participating Edges see potential to expand within 12 to 24 months of a successful commercialization project and are willing to act as reference sites and champion the solution to peer organizations.