Skip to main content
NSAG / Continuing Education / Medicine · CME
ENROLLMENT NOT YET OPEN Accreditation applications in process across all 7 CE types. Book a discovery call to discuss institutional access or co-accreditation partnerships.
CE Specialty Track PREREQUISITE: Core CE Program Required

AI Governance and Integrative Medicine Evidence Standards for Physicians

An algorithm used for 200 million patients was systematically underestimating Black patients’ healthcare needs. Most hospitals still don’t have the governance infrastructure to detect this in their own clinical AI tools. Session M-CE1 gives you that framework. Session M-CE2 addresses integrative medicine governance — including lāʻau lapaʻau in Hawaiʻi — with the same evidence-based rigor.

2.0
CME Credit Hours
2
Add-On Sessions
On-Demand
Self-Paced Format
Download Track Overview PDF Book a Discovery Call
CME Specialty Track

Medicine CE

An algorithm used for approximately 200 million patients was producing systematic racial bias because it used healthcare cost as a proxy for health need. Remedying the disparity would have raised Black patients receiving extra care from 17.7% to 46.5%. The algorithm was commercially validated, FDA-cleared, and EHR-integrated — and it was producing that outcome for years before anyone measured it. Session M-CE1 gives physicians the governance framework for independent demographic impact assessment, before and after adoption.

Session M-CE2 addresses the governance gap on the other side of the evidence spectrum: institutions adopting traditional and complementary medicine — including lāʻau lapaʻau in Hawaiʻi — without evidence integration frameworks, cultural sovereignty agreements, or practitioner credentialing standards. The WHO identified governance infrastructure, not evidence absence, as the primary barrier. This session gives physicians the framework for both the evidence evaluation and the cultural sovereignty obligations.

Who This Is For
  • MDs and DOs (all specialties)
  • Residents and fellows
  • Physician executives (CMOs, medical directors)
  • Hospitalists and primary care physicians
  • Any physician whose practice involves clinical AI tools
  • Any physician whose patients use traditional or complementary medicine
Accreditation Status
ACCME (Accreditation Council for Continuing Medical Education)
2.0 AMA PRA Category 1 Credits™ — pending ACCME accreditation or joint provider arrangement
In Process

Pursuing joint providership with an ACCME-accredited provider partner as the primary pathway; the partner’s ACCME accreditation status must be confirmed. AMA PRA Category 1 Credit™ trademark language requires ACCME authorization.

What’s Included

Each Session Includes

  • ✓Pre-test (5 questions per session)
  • ✓Timed prerecorded video session (60 minutes)
  • ✓Post-test (5 questions per session; minimum 80% required for credit)
  • ✓Practice application exercise (active written or structured response required)
  • ✓Certificate of completion (issued on post-test completion; includes CE type and credit hours)
  • ✓Written materials package (substantive outline, authority map, and hypothetical with model analysis)
Prerequisite

Core CE Program Required

This specialty track is an add-on to the 15-session NSAG Core CE Program. The core curriculum must be completed before enrolling in this track. The core program introduces the four-tier governance framework, six-dimension scoring model, and module-specific evidence bases that this track builds on.

View Core CE Program →
Session Structure

60-Minute Timed Agenda · Every Session

Both sessions in this track follow the same 60-minute structure. Pre-test and post-test are required for credit. The practice application exercise requires an active written or structured response.

Time Duration Segment Content Credit-Eligible
0:00–0:03 3 min Welcome & Disclaimers No legal/medical advice disclaimer; speaker conflict of interest disclosure; CE accreditation status statement; content validation date stated No
0:03–0:08 5 min Pre-Test 5-question pre-test assessing baseline knowledge of the session’s governance domain content. Required for credit. Yes
0:08–0:12 4 min Overview & Learning Objectives Three learning objectives stated; governance problem for this session introduced in one sentence with the primary evidence source named Yes
0:12–0:22 10 min The Governance Problem Evidence base for the governance gap: primary source cited, verbatim key finding presented, governance implication stated Yes
0:22–0:35 13 min Evidence Base & Research Foundation 2–3 primary peer-reviewed sources: study summary, verbatim key finding, DOI displayed; governance implication per source Yes
0:35–0:45 10 min Six Governance Dimensions Each dimension defined and illustrated with a profession-specific institutional example Yes
0:45–0:50 5 min PIONEERING Governance Standards Full PIONEERING tier criteria for this module — observable, documentable governance standards that distinguish PIONEERING from EMERGING Yes
0:50–0:57 7 min Practice Application Exercise Structured exercise requiring active written or multiple-choice response. Completion required for credit. Yes
0:57–1:00 3 min Post-Test & Certificate 5-question post-test; minimum 80% score required for credit; certificate of completion issued on successful completion Yes
Learning Outcomes

What You’ll Be Able to Do

Session M-CE1 · 1.0 AMA PRA Category 1 Credit™

AI and Clinical Decision Support Governance

1

Assess a described clinical AI tool for demographic bias using the Obermeyer et al. (2019) proxy problem framework, identifying the proxy measure and the magnitude of disparity in a described scenario

Competency: Medical Knowledge; Systems-Based Practice Evidence: Post-test: given AI tool summary, identify proxy measure, affected population, and disparity
2

Evaluate a hospital’s AI governance against ONC/ASTP standards using Chang et al. (2025) Data Brief No. 80 findings, identifying gaps in demographic impact assessment and post-adoption monitoring

Competency: Systems-Based Practice; Medical Knowledge Evidence: Application exercise: score a described hospital’s AI governance across the six NSAG M11 dimensions
3

Identify at least two physician professional responsibility obligations arising from clinical AI tool use, citing applicable AMA Code of Medical Ethics provisions

Competency: Professionalism; Communication Evidence: Post-test: given a scenario involving physician reliance on biased AI output, identify AMA Code obligations
Session M-CE2 · 1.0 AMA PRA Category 1 Credit™

Evidence Standards for Integrative Medicine

1

Apply the Steel et al. (2025) WHO TCIM evidence integration framework to assess whether a described CAM modality has adequate evidence for institutional adoption

Competency: Medical Knowledge; Patient Care Evidence: Post-test: given a proposed integrative medicine program, evaluate evidence adequacy for institutional adoption
2

Assess a described integrative medicine program’s cultural sovereignty governance for Indigenous healing practices including lāʻau lapaʻau, using Kawakami et al. (2022) as the primary evidence source

Competency: Cultural Competency; Interpersonal and Communication Skills Evidence: Application exercise: identify cultural sovereignty governance gaps in a described hospital integrative medicine program
3

Identify physician disclosure and counseling obligations when a patient uses traditional or complementary medicine in combination with prescribed medications

Competency: Patient Care; Communication Evidence: Post-test: patient counseling scenario involving CAM and conventional care drug interaction
Sample Course Content

Practice Hypothetical & Model Analysis

The following is a representative hypothetical from this specialty track. Each session includes a practice application exercise based on a realistic institutional scenario. The model analysis below represents the level of analysis expected from practitioners completing this course.

Session M-CE1 · AI and Clinical Decision Support Governance

Hypothetical M-CE1: The Sepsis Algorithm

A 650-bed academic medical center’s CMO approves a commercially available AI sepsis prediction tool. The tool was FDA-cleared as a Class II medical device. The vendor provided validation data showing 88% sensitivity. The tool is integrated into the EHR and flags patients as ‘sepsis risk: high/moderate/low.’

Eighteen months after deployment, a quality improvement analyst finds that Black patients flagged as ‘moderate risk’ are being admitted to the ICU at 60% of the rate of white patients flagged as ‘moderate risk,’ controlling for vital signs, lab values, and documented symptoms. The CMO is informed. No action is taken pending ‘further analysis.’

Discussion Questions
  1. Apply Obermeyer et al. (2019): what proxy measure is the tool likely using if Black patients with equivalent clinical presentations are flagged lower? Explain the mechanism.
  2. The tool was FDA-cleared and commercially validated. Why is this insufficient to rule out demographic bias in this hospital’s patient population?
  3. Under AMA Code 1.1.6 and 1.2.11, what obligations does the CMO have upon receiving the QI finding? What does ‘further analysis pending’ constitute as a governance response?
  4. Draft a suspension threshold criterion (NSAG M11 Dimension 6) that would have triggered a defined response earlier in this scenario.
Model Analysis
  1. Proxy mechanism: The most common proxy in clinical AI tools is healthcare utilization (prior visits, procedures, medications). Obermeyer et al. (2019) demonstrated this proxy encodes racial disparities because Black patients — facing structural barriers including under-insurance and geographic access limitations — access healthcare at lower rates than white patients with equivalent illness severity. If the sepsis tool uses prior utilization as a feature, Black patients will generate lower scores not because they are less sick, but because their prior healthcare contact is lower.
  2. FDA clearance is insufficient: FDA 510(k) clearance evaluates substantial equivalence to a predicate device, not demographic performance in a specific hospital’s patient population. Chang et al. (2025) found 71% of hospitals use EHR-integrated predictive AI while only 57% evaluate all or most models for bias. FDA clearance is a regulatory baseline, not a governance substitute. Physician governance requires independent demographic impact assessment for the institution’s specific patient population.
  3. AMA Code obligations: AMA Code 1.1.6 (patient safety reporting) requires physicians to report conditions that harm or risk harming patients. The QI finding is precisely such a condition. AMA Code 1.2.11 (artificial intelligence) requires physicians to advocate for AI systems that meet patient care standards for safety, efficacy, and equity. ‘Further analysis pending’ as the governance response constitutes a failure to take required corrective action. Required steps include: notifying the clinical team, implementing human override review for affected patients, commissioning independent demographic assessment, and disclosing the finding to the vendor.
  4. Suspension threshold (M11 Dimension 6): ‘If post-adoption monitoring reveals that any demographic group’s ICU admission rate for a given risk tier differs by more than 15% from the aggregate rate after controlling for clinical presentation, clinical leadership is authorized and required to suspend AI-assisted risk stratification for affected patient subgroups and revert to clinical judgment pending independent investigation, with findings reported to the CMO within 30 days.’

This model analysis represents the depth of reasoning expected in the practice application exercise. The post-test (minimum 80%) assesses conceptual mastery of the session’s evidence base and governance framework.

View the full Evidence Library → for the primary sources cited in this and all other specialty track sessions.

Related Tracks

Practitioners Who Take This Track Also Benefit From

CNE
Nursing Specialty Track

The Nursing track (Session N2) addresses clinical AI governance for EHR-integrated tools from the nursing perspective, with specific attention to documentation burden and EHR workflow design as governance domains. The Medicine and Nursing tracks together provide a cross-disciplinary view of clinical AI governance.

View track →
CLE
Law Specialty Track

The Law track (Session L2) covers algorithmic accountability from the attorney’s perspective, using the same COMPAS evidence base as a case study. Physicians who consult to or serve as expert witnesses in cases involving clinical AI benefit from understanding the legal governance framework.

View track →
Institutional Access

License This Track for Your Organization

Institutional licensing includes cohort enrollment across both the Core CE Program and this specialty track, facilitator support, and options to integrate with existing CE tracking systems. Pricing includes all 15 core sessions plus this specialty add-on.

$5,000
1–50 employees
Core CE + This Specialty Track
$10,000
51–200 employees
Core CE + This Specialty Track
$15,000
201+ employees
Core CE + This Specialty Track
Inquire About Institutional Access
Download Track Overview

Get the condensed 2–3 page program overview

Program description · All learning outcomes · Full 60-minute timed agenda · Accreditation status · No enrollment required.

The overview is not gated. Download it directly, without an email address. If you do enter one it is sent to NSAG so we can tell you when accreditation completes, and nothing else — see the privacy notice.