• AOC Header

    Analytics Oversight Committee (AOC)

The Analytics Oversight Committee (AOC) is the Health Data Oversight program subcommittee that oversees advanced analytic models, including artificial/augmented intelligence (AI) and machine learning (ML), intended for real time or near-real time clinical decision-making.  The AOC evaluates models using the S.A.F.E. criteria: Safety/Risk, Accuracy, Fairness/Bias, and Evidence.  The AOC also oversees the work done by the Health Analytics Core (HAC), which provides data science and biostatistics resources for model development and performance evaluation.

 

The AOC aims to work with requestors to enable innovative, effective, and meaningful solutions to the health system's biggest challenges in support of the clinical strategic plan and organizational vision.

 

Considerations Regarding the AOC Evaluation
An AOC approval does not equate to an overall project/initiative approval. ​Most aspects of an overall project/initiative approval are outside of AOC scope, and can include IT support, project management, regulatory and compliance, data sharing, procurement, clinical and operational implementation, and funding.

Coordinating AI Oversight Across UC Davis Health
Artificial intelligence tools are evaluated through the collaborative efforts of multiple oversight
bodies at UC Davis Health, each bringing a unique lens to the review process: 
  • Procurement: Ensures tools are acquired through appropriate channels and contracts
  • Legal: Ensures compliance with applicable laws and regulations
  • Privacy: Evaluates appropriate use/disclosure of patient personal and protected information
  • Compliance: Reviews adherence to UC and UC Davis Health policies
  • Risk Management: Assesses institutional liability and risk exposure
  • Information Technology (IT): Reviews technical feasibility, integration, and infrastructure needs
  • Intellectual Property (IP) and Tech Transfer: Analyzes ownership, licensing, and commercialization concerns
  • Health Ventures: Assesses alignment with health system innovation goals and potential scalability
  • Institutional Review Board (IRB): Determines requirements for research involving human subjects

Because each group uses different criteria and priorities, recommendations may vary. In cases of disagreement or complex evaluations, the AOC plays a key role in facilitating cross-functional discussion and resolution.

Ultimately, the AOC's mandate is to determine whether a tool is “safe enough” to be used in the UC Davis Health environment, particularly for direct or indirect patient care all while considering ethical, technical, and clinical impacts of deployment in real-world settings.


Frequently Asked Questions

Models (including devices, algorithms or other technology utilizing AI) that are used for real-time or near real-time clinical decisions or clinical research should be reviewed by the AOC. If you are unsure, email the HDOC Support Team (HDOC@health.ucdavis.edu).

Use the Scope Decision Trees (staff only) to determine if your request is in- or out-of-scope. Models that do not influence clinical decisions (like simple rule based Epic BPAs or dashboards) are usually out of scope but still need to be listed in the UCDH Model Registry.

Note: Sometimes leadership requests AOC review of a request that would typically be out-of-scope. If this is the case, email the HDOC Support Team (HDOC@health.ucdavis.edu) to determine next steps for your specific use case.

  1. Confirm your request is In-Scope (staff only)
  2. Complete the Pre-AOC Checklist (staff only) to ensure all criteria are met. Some important prerequisites are:
    1. Roles, Responsibilities and Ownership Confirmed (e.g. Clinical/Business Champion, Executive Sponsor, Downstream Impacted Groups)
    2. Strategic Alignment with Impacted Groups
    3. IT "Request New Technology Initiatives" Ticket Submitted via ServiceNow
    4. Implementation Workflows and Analysis Plan Clearly Defined and Documented
    5. If applicable, Vendor Assessment Completed (third party, non-Epic)
    6. If applicable, IRB Approved Protocol Required for Intended Clinical Research Implementation
  3. Submit an AOC Analytics Evaluation/Consult (staff only) ticket in ServiceNow.

You can request an AOC review by completing the AOC Analytics Evaluation/Consult (staff only).

These are the Frameworks the OAC uses to evaluate models requested to be deployed at UCDH.

The AOC uses the S.A.F.E. Criteria, as well as the impact to the current standard of care. 

The S.A.F.E. criteria are: 

  • Safety/Risk: Is the model reasonably safe to use on patients? 
  • Accuracy: Is the model reasonably accurate for what it is trying to predict? 
  • Fairness/Bias: Are there any fairness or bias concerns? 
  • Evidence: Is there high-quality evidence that supports the use of the model? 

After you submit, the AOC triage team will review and reach out for any additional information needed. The review process is outlined here, and the anticipated timeline depends on the complexity of the request:

Analytics Oversight Committee (AOC) Review Process

Anticipated timeframes for an AOC decision are as follows. Timelines may vary depending on whether the request meets the AOC pre-requisites and includes all necessary information:

  • Low complexity: <2-weeks
  • Medium: 2-4 weeks
  • High: 5-8 weeks
  • The AOC meets twice per month (2nd Wednesday & 4th Tuesday)

Once approved, your model moves into the appropriate AOC Phase (e.g., Pre-Implementation → Pilot → Go-Live → Ongoing Monitoring). To move into the next Phase, you will need to re-engage the AOC and pass the next Phase Gate. See the AOC Phases & Phase Gates (staff only) for additional information. If it has been over 1-year, the AOC may periodically check-in to confirm the model is still in-use and performing as expected.

Here are the most common situations that would require you to re-engage the AOC:

  • Ready to Progress to the Next Phase: If the model were approved for a phase and you are ready to move to the next phase (e.g., you got approval to pilot and now you would like to implement). See the AOC Phases & Phase Gates (staff only) for additional information on the AOC 5-Phases.
  • Material Change: Any material changes and/or updates (e.g., retraining, new features, major workflow impacts, new setting or impacted population, etc.)
  • Replace: There is a desire to implement a new model in replace the existing (previously approved) model
  • Retire or Decommission: No longer wish to use the existing (previously approved) model and no replacement is identified - no AOC review is needed but the UCDH Model Registry will need to be updated.

The decision will outline the areas of concern as well as recommended next steps. Requestors may be referred to as resources to help mitigate the concerns. If they disagree, Requestors can choose to escalate - details on how to initiate this process will be provided in the decision document.

Always feel free to email the AOC Program Manager, Rachelle Carson (racarson@health.ucdavis.edu) or the HDOC Support Team (HDOC@health.ucdavis.edu) with any questions you may have.