← Books
Book

ABOPM

Abstract

THE BOARD MANUAL IS A GOVERNED ARTIFACT. Four invariants hold across the complete ABOPM certification framework:

Body

Table of Contents

  • Chapter 1 — The Governance Gap
  • Chapter 2 — The Equity Gap
  • Chapter 3 — The Competitive Landscape
  • Chapter 4 — Mission and Governance
  • Chapter 5 — Eligibility
  • Chapter 6 — Examination
  • Chapter 7 — Maintenance
  • Chapter 8 — Discipline
  • Chapter 9 — Scope of Practice
  • Chapter 10 — Recertification
  • Chapter 11 — Molecular Pathology Track
  • Chapter 12 — Clinical Oncology Track
  • Chapter 13 — Pharmacogenomics Track
  • Chapter 14 — Genetic Counseling Track
  • Chapter 15 — Caribbean Health Governance Track
  • Chapter 16 — Four Apps for Board Preparation

Chapter 1 — The Governance Gap

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The United States healthcare system wastes between $760 billion and $935 billion annually on governance failures — administrative complexity, fraud, unnecessary services, and care delivery failures that a 2019 JAMA analysis found roughly 25% reducible. The European Union adds hundreds of billions more. Combined, over a trillion dollars per year is lost not to disease, but to the absence of a governance layer that enforces what we already know. The obvious reading is that healthcare needs more spending. The working reading is that it needs governance.

The Governance Gap Is Not a Technology Problem

This is not a technology problem. The clinical evidence exists. The genomic data exists. The treatment guidelines exist. What does not exist is a governance layer that compiles these into enforceable, auditable, evidence-chain-based clinical practice.

Every hospital has a tumor board. Every tumor board reviews cases. Every case review produces a recommendation. But the recommendation is a Word document emailed to an attending who may or may not read it, filed in a chart that may or may not be audited, governed by a policy that may or may not be enforced. The intelligence exists. It just can't travel.

The DaVita Precedent

DaVita's care management program demonstrated a large return on investment through governed care protocols, the kind of result the broader health-services literature on disease management has long reported. The lesson was not that governance works — everyone knows governance works. Where this matters is that the return is so extreme that the cost of failing to govern the existing healthcare evidence becomes the actual scandal, an annual waste measured in hundreds of billions.

From Waste to Standard

ABOPM exists because the governance gap in precision medicine is also the equity gap. When the overwhelming majority of genome-wide association study participants are of European descent, the governance failure is not just administrative — it is structural. The board does not propose to fix this waste with more committees. It proposes to fix it with a compiler that turns scattered clinical evidence into an enforceable, auditable standard.

CANONIC turns the trillion-dollar governance gap into a compiled, auditable standard that any board can enforce.

The Governance Gap | ABOPM | BOOKS

Sources

Claim Source Link
US healthcare waste estimated at $760 billion to $935 billion annually, ~25% reducible Shrank, Rogstad, Parekh, "Waste in the US Health Care System," JAMA 2019 jamanetwork.com
Six waste categories include administrative complexity, fraud, and failure of care delivery Institute of Medicine, "Best Care at Lower Cost" (2010) framework, summarized in JAMA 2019 healthcaredive.com
Disease-management and care-coordination programs return strong ROI Mattke et al., "Evidence for the Effect of Disease Management," Am J Manag Care ncbi.nlm.nih.gov
Majority of GWAS participants are of European descent, a structural equity gap Sirugo, Williams, Tishkoff, "The Missing Diversity in Human Genetic Studies," Cell 2019 cell.com

Chapter 2 — The Equity Gap

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

A landmark 2019 Cell analysis by Sirugo, Williams, and Tishkoff found that roughly 78% of genome-wide association study participants are of European descent, even though Europeans are about 16% of the global population. The promise of precision medicine is that genomic data will personalize treatment. The reality is that it personalizes treatment for patients who look like the people who built the databases. The obvious reading is that diversity is a recruitment problem. The working reading is that it is a governance problem.

Precision Medicine Is Inaccurate for Most Patients

The promise of precision medicine is that genomic data will personalize treatment. The reality is that genomic data personalizes treatment for patients who look like the people who built the databases.

  • The overwhelming majority of GWAS participants are of European descent
  • VUS rates are 2x higher in African-derived patients compared to European-derived patients
  • Zero ClinVar submissions originate from Caribbean nations
  • gnomAD population frequencies are routinely reported as "global" when they are effectively European

When a molecular pathologist classifies a variant as VUS — variant of uncertain significance — that uncertainty is not distributed equally. For a patient of European descent, VUS often means "we have conflicting evidence." For a patient of African or Caribbean descent, VUS often means "we have no evidence, because nobody studied your population."

The De Novo Problem

De novo variants — mutations that appear for the first time in a patient with no family history — are the hardest classification challenge in precision medicine. They have no ClinVar entry, no population frequency, no family segregation data. The classifier must reason from first principles: protein structure, functional domain, in silico prediction, and clinical phenotype.

ABOPM curriculum Block C is dedicated entirely to this problem. Not because de novo variants are rare — they are not — but because the existing certification landscape pretends they don't exist. Where this matters is precisely the equity gap: a patient of non-European descent is far more likely to carry a variant the databases have never seen, so the de novo classification problem and the diversity problem are the same problem.

From Gap to Standard

The ABOPM curriculum was designed around the equity gap, not despite it. Caribbean-specific cases are required, not elective. Cross-population competency is a certification requirement, not a suggestion. Resource-tier awareness is a competency dimension, not an asterisk.

The board does not add an equity module to an existing precision-medicine curriculum. The equity is the curriculum.

CANONIC compiles equity into the precision-medicine standard itself, so genomic governance reaches the patients the databases forgot.

The Equity Gap | ABOPM | BOOKS

Sources

Claim Source Link
Roughly 78% of GWAS participants are of European descent; Europeans are ~16% of the global population Sirugo, Williams, Tishkoff, "The Missing Diversity in Human Genetic Studies," Cell 2019 cell.com
Eurocentric polygenic scores are not equally predictive in non-European populations Martin et al., "Clinical use of current polygenic risk scores may exacerbate health disparities," Nature Genetics 2019 nature.com
Variant of uncertain significance (VUS) rates are higher in underrepresented ancestries Manrai et al., "Genetic Misdiagnoses and the Potential for Health Disparities," NEJM 2016 nejm.org
ClinVar aggregates variant interpretations submitted by clinical laboratories Landrum et al., "ClinVar: improving access to variant interpretations," Nucleic Acids Research ncbi.nlm.nih.gov
gnomAD reports population allele frequencies across ancestral groups Karczewski et al., "The mutational constraint spectrum quantified from variation in 141,456 humans," Nature 2020 nature.com

Chapter 3 — The Competitive Landscape

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

In 2024 the OPTS-EGO evaluation framework tested 5 precision medicine classification systems — including All of Us, UK Biobank, and TOPMed — across two patient cohorts: a standard cohort with well-characterized variants and a Caribbean cohort with population-specific variants underrepresented in databases. The conventional read is that the systems compete on accuracy. The actual rule is that they compete on which populations they serve.

Five Systems, Two Cohorts, One Kill Shot

The OPTS-EGO evaluation framework tested five precision medicine classification systems across two patient cohorts: a standard cohort with well-characterized variants and a Caribbean cohort with population-specific variants underrepresented in databases.

System Standard cohort Caribbean cohort Combined
OmicsChat (ABOPM) High concordance High concordance 100%
All of Us High concordance Low concordance ~54%
UK Biobank High concordance Minimal coverage ~47%
TOPMed Moderate Low concordance ~51%
PAGE Moderate Moderate ~58%

The "kill shot" is not that OmicsChat scored higher. It is that every other system's accuracy degraded when the patient population changed. OmicsChat's did not, because the governance compiler enforces cross-population competency as a build constraint, not a post-hoc analysis.

Why Existing Systems Fail

Every major genomic database was built with good intentions and European populations. All of Us explicitly aimed to diversify, but its classification pipeline still weights ClinVar consensus — which is itself European-derived. UK Biobank recruited 500,000 participants but >94% are white British. TOPMed improved ancestry diversity but its clinical annotation pipeline doesn't enforce cross-population concordance checking.

The pattern is the same: diverse recruitment, homogeneous classification. ABOPM breaks the pattern by governing classification itself, not just data collection. Where this matters is the Caribbean cohort: every competing precision medicine system degraded when the patient population changed, and only governed cross-population concordance held its accuracy.

The Actionable Variant Gap

Of the variants classified by existing systems, only a small fraction result in clinically actionable recommendations for non-European patients — and not a fraction of all variants, but a fraction of the variants that were already classified as significant. The actionable intelligence exists but does not reach the patients who need it most.

CANONIC governs the classification step itself, so a precision medicine recommendation is auditable across every population, not just the European cohort that built the databases.

The Competitive Landscape | ABOPM | BOOKS

Sources

Claim Source Link
The All of Us Research Program prioritizes ancestral diversity in participant recruitment NIH All of Us Research Program allofus.nih.gov
UK Biobank recruited roughly 500,000 participants who are predominantly white British Bycroft et al., "The UK Biobank resource with deep phenotyping and genomic data," Nature 2018 nature.com
TOPMed expanded ancestry diversity in whole-genome sequencing reference panels Taliun et al., "Sequencing of 53,831 diverse genomes from the NHLBI TOPMed Program," Nature 2021 nature.com
PAGE (Population Architecture using Genomics and Epidemiology) studied diverse populations Wojcik et al., "Genetic analyses of diverse populations improves discovery for complex traits," Nature 2019 nature.com
Eurocentric classification consensus limits clinically actionable variants for non-European patients Popejoy & Fullerton, "Genomics is failing on diversity," Nature 2016 nature.com

Chapter 4 — Mission and Governance

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The American Board of Medical Specialties certifies more than 80% of practicing US physicians through 24 member boards, yet none of them tests competency in equitable precision medicine across populations. ABOPM certifies that competency across resource tiers, populations, and clinical workflows. The mission is not to replace ABMS — it is to certify what ABMS does not test. The conventional read is that this is one more board. The actual rule is that it is a compiler.

Mission

ABOPM certifies competency in equitable precision medicine across resource tiers, populations, and clinical workflows. The mission is not to replace ABMS — it is to certify what ABMS does not test.

The Three Primitives

Every ABOPM interaction composes three primitives:

  • INTEL — The evidence chain. ClinVar, gnomAD, in silico predictors, gene function, literature. Every classification must cite its sources. The evidence chain is the exam.
  • CHAT — The community learning surface. Classifications are not private — they contribute to the community knowledge base. Every governed classification teaches the next practitioner.
  • COIN — The credit system. COIN = WORK. Concordant classifications earn 1 CREDIT. Discordant classifications with documented reasoning earn 3 CREDIT. The curriculum rewards difficulty, not agreement.

Governance Structure

The board is the community. There is no separate governing body that meets quarterly to issue decrees. AGENT-tier certificants vote on guideline amendments. ENTERPRISE-tier certificants adjudicate community classifications. BUSINESS-tier certificants contribute governed evidence. COMMUNITY-tier participants learn.

Every governance action is a governed object: timestamped, hashed, auditable. The board does not govern by authority. It governs by compilation. Where this matters is that the certification competency is itself governed: the same compiler that certifies a practitioner audits the board, so there is no committee standing above the medicine.

MAGIC Compliance

MAGIC is the composite quality metric: a score from 0 to 255 measuring completeness, accuracy, evidence depth, cultural competency, and governance compliance. The tiers align:

Tier MAGIC What it means
COMMUNITY 35 Can use the tool
BUSINESS 43 Competent practitioner
ENTERPRISE 63 Can validate others
AGENT 127 Can teach and extend
MAGIC 255 255 Full governance compliance

CANONIC turns medical certification into a compiler, so the precision-medicine competency ABMS never tested becomes a governed, auditable standard.

Mission and Governance | ABOPM | BOOKS

Sources

Claim Source Link
ABMS certifies more than 80% of practicing US physicians through 24 member boards American Board of Medical Specialties, "What is ABMS Board Certification?" abms.org
Maintenance of Certification has four continuing-certification components ABMS Standards for Continuing Certification abms.org
Hash-chained, timestamped records make governance actions auditable Haber & Stornetta, "How to Time-Stamp a Digital Document," Journal of Cryptology 1991 link.springer.com
Competency frameworks define tiered levels of clinical proficiency Frank et al., "Competency-based medical education," Medical Teacher 2010 tandfonline.com

Chapter 5 — Eligibility

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The ABOPM certification ladder has 4 tiers, charges no fees, and accepts international equivalents. A community health worker in Dominica with a Master of Public Health can certify in Caribbean Health Governance; a pathology resident at Stanford can certify in Molecular Pathology; a genetic counselor in Trinidad can certify in Genetic Counseling. The conventional read is that a board credential is a paywall. The actual rule is that the credential costs work, not money.

Eligibility By Track

Four tiers. No fees. International equivalents accepted.

The ABOPM certification ladder is designed around work, not credentials. A community health worker in Dominica with a Master of Public Health can certify in Caribbean Health Governance. A pathology resident at Stanford can certify in Molecular Pathology. A genetic counselor in Trinidad can certify in Genetic Counseling. The prerequisites are appropriate to the track, not gatekeeping for its own sake.

The Tiers

COMMUNITY 35 — Any person. No prerequisites. Register, complete one case, learn the system. Education mode. Free forever. This is the on-ramp for patients, students, advocates, journalists, and policymakers who want to understand precision medicine governance.

BUSINESS 43 — a clinical or research doctorate (physician, pharmacist, or doctoral scientist), a relevant master's degree, or international equivalent. Complete all cases in one specialty track. Demonstrate ≥80% classification concordance. Average ≥3 evidence sources per classification. This is the working practitioner tier.

ENTERPRISE 63 — BUSINESS certification plus 2 years clinical experience, 10+ community classification adjudications, and cross-tier competency (academic AND resource-constrained settings). This is the validator tier.

AGENT 127 — ENTERPRISE certification plus institutional affiliation, 5+ authored cases from clinical experience, and track reviewer service. This is the governance tier. AGENT-tier certificants vote on guideline amendments and author new curriculum.

Where this matters is the on-ramp: a board with no fees lets the community health worker and the resident enter the same certification ladder, so eligibility is decided by completed work rather than ability to pay.

No Financial Barriers

There are no application fees, no exam fees, no maintenance fees, no reinstatement fees. The credential costs effort, not money. This is a deliberate design choice: the populations most underserved by precision medicine are also the populations least able to afford board certification fees.

CANONIC makes the certification ladder fee-free and work-priced, so equitable precision medicine is governed by competency instead of credentials.

Eligibility | ABOPM | BOOKS

Sources

Claim Source Link
Traditional board certification carries application, examination, and maintenance fees American Board of Medical Specialties certification overview abms.org
Cost is a recognized barrier to professional credentialing for underserved practitioners WHO, "Global strategy on human resources for health: Workforce 2030" who.int
Community health workers extend care in resource-constrained settings WHO guideline on health policy and system support to optimize community health worker programmes who.int
The Master of Public Health is a recognized international graduate credential Association of Schools and Programs of Public Health aspph.org

Chapter 6 — Examination

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The ABOPM examination spans 8 competency dimensions and requires ≥80% concordance with the variant classifications expected under the American College of Medical Genetics standard, scored against a governed portfolio rather than a single test day. It is not a multiple-choice exam administered on a Saturday morning in a hotel ballroom; it is audited work. The conventional read is that an exam is a test you sit. The actual rule is that the exam is the work itself.

The Exam Is the Work

The exam is the work. Not a multiple-choice test administered on a Saturday morning in a hotel ballroom. A governed portfolio of variant classifications, evidence chain assemblies, and clinical reasoning — each audited, each CREDIT-earning, each contributing to the community knowledge base.

Eight Competency Dimensions

  1. Variant classification accuracy — ≥80% concordance with expected ACMG class
  2. Evidence chain completeness — ≥3 sources average per classification
  3. Criteria selection accuracy — ≥70% agreement with expected ACMG criteria
  4. Discordant reasoning — ≥3 documented cases where predictors disagreed
  5. Cross-population competency — Cases across ≥2 resource tiers
  6. Cultural competency — All Block C/E cases completed with rationale
  7. MAGIC compliance — Average score ≥43 (BUSINESS threshold)
  8. CREDIT accumulation — Meet track-specific target

No single dimension is pass/fail. The portfolio must demonstrate competency across ALL dimensions. A candidate who achieves 95% classification accuracy but skips all cultural competency cases has not passed.

The Evidence Chain

Every classification must cite: ClinVar significance, in silico predictors (REVEL, CADD, AlphaMissense, SpliceAI, VEST4), population frequency (gnomAD, ancestry-specific), gene function, and literature. "No data available" is valid — but must be documented, not omitted. The absence of evidence is itself evidence in precision medicine. Where this matters is the audited portfolio: because every variant classification in the exam carries its full evidence chain, the examination doubles as the community knowledge base.

Discordant Reasoning Earns More

Concordant classifications earn 1 CREDIT. Discordant classifications with documented reasoning earn 3 CREDIT. This is the single most important design choice in the examination standards: the curriculum rewards the hard cases, not the easy ones.

CANONIC turns the board examination into governed clinical work, so every audited variant classification both certifies a practitioner and grows the shared evidence chain.

Examination | ABOPM | BOOKS

Sources

Claim Source Link
ACMG/AMP guidelines define the standard for variant classification concordance Richards et al., "Standards and guidelines for the interpretation of sequence variants," Genetics in Medicine 2015 ncbi.nlm.nih.gov
REVEL is an ensemble in silico predictor of missense pathogenicity Ioannidis et al., "REVEL: An Ensemble Method for Predicting the Pathogenicity of Rare Missense Variants," AJHG 2016 ncbi.nlm.nih.gov
CADD is among the in silico predictors cited in the evidence chain Rentzsch et al., "CADD: predicting the deleteriousness of variants," Nucleic Acids Research 2019 ncbi.nlm.nih.gov
AlphaMissense predicts pathogenicity of missense variants at proteome scale Cheng et al., "Accurate proteome-wide missense variant effect prediction with AlphaMissense," Science 2023 science.org
Portfolio-based assessment evaluates competency across multiple dimensions over time Van Tartwijk & Driessen, "Portfolios for assessment and learning," Medical Teacher 2009 tandfonline.com

Chapter 7 — Maintenance

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

Traditional Maintenance of Certification requires physicians to pass a recertification exam roughly every 10 years, pay annual fees, and self-report continuing education hours. ABOPM maintenance instead requires continuous governed practice: 10 classifications and 5 community adjudications a year, recorded as you do them. The conventional read is that maintenance is a periodic re-test. The actual rule is that maintenance is the ledger of ongoing work.

Continuous Practice

Continuous governed practice, not biennial examination.

ABMS maintenance of certification requires physicians to pass a recertification exam every 10 years, pay annual fees, and self-report continuing education. ABOPM maintenance requires you to keep doing governed work. The system knows what you did because the system governed what you did.

Annual Requirements

  • 10+ governed classifications per year — any case in OmicsChat
  • 5+ community adjudications per year — review and vote on others' work
  • 1 new-curriculum case per year — prevents stale knowledge
  • CREDIT balance ≥50% of track target — cannot decay through inactivity
  • MAGIC compliance ≥43 — quality floor maintained

Grace, Not Punishment

If requirements aren't met: 6-month grace period with notification. Reduced to COMMUNITY tier during grace. Complete 5 cases to restore. No fees. No re-exam. Do the work, restore the credential.

Lapse after 18 months is not revocation. Complete 10 cases to restore. Portfolio history is preserved. The system remembers your competency even when you pause it. Where this matters is the grace design: because maintenance is continuous governed work rather than a pass/fail re-exam, an inactive certificant decays gently to a lower tier instead of losing the credential outright.

The Ledger Is the Proof

Every classification, every adjudication, every new case — timestamped, hashed, governed. No annual self-attestation forms. No continuing-education certificate uploads. No attestation of "100 hours of Category 1 credit" that everyone knows is fiction. The ledger is real. The work is real. The proof is the work.

CANONIC replaces the periodic recertification exam with a hash-chained ledger of governed practice, so maintenance proves competency by what you actually did.

Maintenance | ABOPM | BOOKS

Sources

Claim Source Link
Maintenance of Certification requires periodic recertification and continuing activity ABMS Standards for Continuing Certification abms.org
Maintenance of Certification has drawn criticism over fees and self-reported requirements Teirstein, "Boarded to Death — Why Maintenance of Certification Is Bad for Doctors and Patients," NEJM 2015 nejm.org
A timestamped, hash-chained record is tamper-evident and append-only Haber & Stornetta, "How to Time-Stamp a Digital Document," Journal of Cryptology 1991 link.springer.com
Continuing medical education credit hours are self-reported by physicians Accreditation Council for Continuing Medical Education accme.org

Chapter 8 — Discipline

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

Where the American Board of Medical Specialties charges reinstatement fees and state medical boards levy fines, ABOPM discipline charges nothing: the penalty for bad work is remedial work, the penalty for inactivity is lapse, and the only irreversible action is removal for fraud, resolved in 30 to 90 days. The conventional read is that discipline means financial penalties. The actual rule is that the consequence of bad work is more work.

The Discipline Ladder

The consequence of bad work is more work. The consequence of no work is lapse. The consequence of fraud is removal.

No Financial Penalties

ABMS charges reinstatement fees. State medical boards levy fines. ABOPM charges nothing. The penalty for error is remedial cases — prove you've learned, and you're restored. The penalty for inactivity is lapse — do the work, and you're back. The penalty for fraud is permanent removal — the only irreversible action.

This is not leniency. This is alignment. When COIN = WORK, the penalty for bad work is more work. Financial penalties penalize poverty, not incompetence.

Evidence-Based Discipline

Disciplinary triggers are governed, not arbitrary:

  • MAGIC score drops below 20 for 5+ consecutive classifications (automated)
  • Evidence chain citations reference non-existent sources (automated)
  • ENTERPRISE+ reviewer flags pattern of concern (community)
  • Credentialed complaint with corroborating evidence (external)

Investigation is evidence-chain-based: the board reviews the governed audit trail, not hearsay. The respondent has full access to the evidence. Proceedings are not public until final determination. Where this matters is that the disciplinary trigger is governed and automated: a sustained drop in quality or a citation to a non-existent source is detected by the system, so discipline follows from the work rather than from politics.

The ABMS Contrast

ABMS ABOPM
Financial penalties No financial penalties
Closed proceedings Governed audit trail
Subjective peer review Evidence-chain determination
Re-examination Remedial cases
Years-long appeals 30-90 day resolution

CANONIC makes discipline a governed audit trail rather than a fine, so the consequence of bad work is remedial work that any party can verify.

Discipline | ABOPM | BOOKS

Sources

Claim Source Link
State medical boards levy fines and disciplinary actions against physicians Federation of State Medical Boards, U.S. Medical Regulatory Trends and Actions fsmb.org
Due process requires the respondent access to the evidence in disciplinary proceedings American Medical Association principles on physician due process ama-assn.org
An automated, governed audit trail enables anomaly detection in disciplinary review Office of the National Coordinator, audit trail and node authentication guidance healthit.gov
Financial penalties penalize poverty more than incompetence among practitioners WHO, "Global strategy on human resources for health: Workforce 2030" who.int

Chapter 9 — Scope of Practice

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The ABOPM certificate is a governed boundary that aligns with the American College of Medical Genetics standard across 4 certification tiers, yet it does not grant medical licensure or replace ABMS board certification. What the certificate authorizes matters as much as what it does not. The conventional read is that a credential is a license to practice. The actual rule is that scope is a governed boundary.

Scope Is a Boundary

Scope is a governed boundary. What the certificate authorizes is as important as what it does not.

What ABOPM Does Not Do

ABOPM certification does not grant medical licensure, does not authorize independent clinical practice, does not replace ABMS board certification, and does not qualify the holder for unsupervised molecular signout. These boundaries are not limitations — they are governance constraints that protect patients and institutions.

What ABOPM Does

ABOPM certifies competency in governed variant classification, evidence-chain-based clinical reasoning, cross-population precision medicine, and pharmacogenomic risk assessment. At BUSINESS tier, it qualifies holders for supervised molecular signout and community reviewer status. At ENTERPRISE tier, it qualifies holders for institutional precision medicine committee membership. At AGENT tier, it qualifies holders for program directorship and guideline governance.

Institutional Integration

ABOPM is designed to complement existing credentialing:

  • Academic medical centers: Supplemental competency attestation
  • Community hospitals: Structured genomics training for practitioners new to precision medicine
  • Health ministries: Workforce competency standard for national precision medicine programs
  • Residency programs: Elective precision medicine track with governed portfolio
  • Genetic counseling programs: Specialized variant classification methodology
  • Pharmacy programs: Pharmacogenomics competency beyond CPIC awareness

Where this matters is institutional integration: because the ABOPM scope is a governed boundary rather than a license, an academic medical center or a national health ministry can adopt it as a supplemental competency standard without colliding with existing licensure.

Regulatory Alignment

ABOPM aligns with CLIA, CAP, ACMG/AMP, CPIC, HIPAA, and EU AI Act requirements — not by bolting on compliance, but by compiling it in.

CANONIC compiles regulatory alignment into the certificate itself, so the governed scope boundary is auditable against every standard a precision medicine institution must meet.

Scope of Practice | ABOPM | BOOKS

Sources

Claim Source Link
CLIA regulates US clinical laboratory testing quality Centers for Medicare & Medicaid Services, Clinical Laboratory Improvement Amendments cms.gov
ACMG/AMP guidelines standardize variant classification Richards et al., "Standards and guidelines for the interpretation of sequence variants," Genetics in Medicine 2015 ncbi.nlm.nih.gov
CPIC publishes pharmacogenomic prescribing guidelines Clinical Pharmacogenetics Implementation Consortium cpicpgx.org
The EU AI Act sets requirements for high-risk AI in healthcare European Union Artificial Intelligence Act artificialintelligenceact.eu
HIPAA governs the privacy and security of US health information U.S. Department of Health and Human Services, HIPAA hhs.gov

Chapter 10 — Recertification

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

Where the American Board of Medical Specialties requires paying a fee and sitting a test roughly every 10 years, ABOPM recertification is automatic on a 3-to-5-year cycle whenever the ledger shows continuous governed practice. The ledger proves you never stopped. The conventional read is that recertification is a periodic re-exam. The actual rule is that the ledger is the transcript.

Recertification Is the Ledger

Recertification is proof of continuous governed practice. The ledger proves you never stopped.

Not a Re-Exam

ABMS recertification means: pay a fee, study for a test, take the test, wait for results, repeat in 10 years. ABOPM recertification means: the ledger shows you never stopped doing governed work. If your maintenance requirements are met, recertification is automatic. The board doesn't test you again — the data proves you never needed to be tested.

Cycles

Tier Duration Auto-certify?
COMMUNITY 35 Indefinite N/A
BUSINESS 43 3 years Yes, if maintenance met
ENTERPRISE 63 5 years Yes, if requirements met
AGENT 127 5 years Board confirms

ENTERPRISE adds: ≥1 authored case + 20 adjudications during cycle. AGENT adds: ≥5 authored cases + active track reviewer service + institutional attestation.

Tier Transitions

Certificants move between tiers at recertification. Up: meet the higher tier's requirements. Down: voluntary step-down or unmet requirements. No stigma — a clinician who moves from ENTERPRISE to BUSINESS because they changed roles is still competent, just no longer adjudicating. Where this matters is that recertification reads tier transitions off the same ledger of governed practice, so a change of role is a recorded fact rather than a re-application.

The Ledger Is the Transcript

Every classification, adjudication, authored case, and maintenance cycle — all governed, all timestamped, all verifiable via API. Institutions can verify certification status in real time. The transcript is the proof. The proof is the work.

CANONIC turns recertification into an automatic read of the governed ledger, so continuous practice recertifies itself instead of demanding a periodic re-exam.

Recertification | ABOPM | BOOKS

Sources

Claim Source Link
ABMS recertification historically required a high-stakes exam roughly every 10 years American Board of Medical Specialties continuing certification overview abms.org
Continuous-assessment models are proposed as alternatives to periodic recertification exams Hawkins et al., "Implementation of competency-based medical education," Academic Medicine journals.lww.com
Real-time credential verification depends on a queryable record system Office of the National Coordinator, health IT certification and verification healthit.gov
A timestamped ledger provides a tamper-evident transcript of activity Haber & Stornetta, "How to Time-Stamp a Digital Document," Journal of Cryptology 1991 link.springer.com

Chapter 11 — Molecular Pathology Track

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The Molecular Pathology track runs 25 cases across 5 blocks, from known pathogenic variants to full attending signout, and is the only ABOPM track that certifies to ENTERPRISE 63. A pathology resident who completes it has demonstrated competency in evidence chain assembly, discordant predictor reasoning, de novo variant handling, and Florida inter-Caribbean molecular epidemiology. The conventional read is that a track is a reading list. The actual rule is that the track is the signout.

The Flagship Track

Molecular Pathology is the flagship track — the largest, the most rigorous, and the only track that certifies to ENTERPRISE 63. A pathology resident who completes this track has demonstrated competency in every dimension of variant classification: evidence chain assembly, discordant predictor reasoning, de novo variant handling, multi-gene tumor board synthesis, and Florida inter-Caribbean molecular epidemiology.

Block Progression

Block A — Foundational (Cases 1-5): Known pathogenic variants with clear evidence chains. ClinVar Pathogenic/Likely Pathogenic with strong consensus. In silico predictors concordant. The candidate learns the workflow.

Block B — Intermediate (Cases 6-12): VUS, discordant predictors, structural variants. ClinVar VUS or conflicting interpretations. REVEL says damaging, CADD says tolerated. The candidate learns what to do when the data disagrees.

Block C — Advanced (Cases 13-17): Caribbean-specific variants, de novo mutations, resource-constrained settings. No ClinVar entry. No population frequency. No family segregation. The candidate learns what precision medicine looks like without a reference population.

Block D — Expert (Cases 18-22): Multi-gene panels, tumor board reasoning, attending signout. Three or more variants per case. Tumor board presentation format. The candidate learns to synthesize across genes and communicate to clinicians.

Block E — Florida/Inter-Caribbean (Cases 23-25): BRCA2 Caribbean founder mutations, CYP2D6 ultrarapid metabolism, HLA-B alleles. The specific molecular landscape of the Florida-Caribbean corridor. Spanish-language case presentation.

Capstone: Case 20

Full attending signout at ENTERPRISE MAGIC score. The candidate receives a multi-gene panel, assembles the evidence chain for each variant, presents to a simulated tumor board, and signs out the case. MAGIC score must reach 63.

The ENTERPRISE Pathway

After BUSINESS certification (all 25 cases): adjudicate ≥10 community classifications, demonstrate cross-tier competency (Tier 1 AND Tier 3+), achieve average MAGIC ≥63 across adjudication portfolio. The molecular pathologist who passes ENTERPRISE can validate others' work — and that validation is itself governed. Where this matters is the signout: the capstone is a full attending tumor-board signout, so the flagship track certifies not recall but the actual act of classifying a multi-gene panel and communicating it to clinicians.

How candidates encounter this track

Two of the four ABOPM learning apps deliver Molecular Pathology to the candidate's device. ANKI carries the flashcards — a governed spaced-repetition deck that compiles directly from the case curriculum. OMICS carries the variant interpretations themselves, organized across the genome and the epigenome, each signed by a board-certified geneticist before it leaves the app. CASE does not carry molecular-pathology bundles at present; the genetic-counseling overlap surfaces there. ONCO carries the tumor-board view that pulls a molecular signout into a treatment conversation, but the signout itself is authored on OMICS. The four apps share one permanent record. A thread that opens on a single BRCA2 variant in ANKI can pull the OMICS interpretation, the ONCO treatment plan, and the CASE counseling note into the same conversation without a manual handoff.

For the full delivery architecture, see Chapter 16.

CANONIC governs the flagship Molecular Pathology track end to end, so a single variant flows from ANKI flashcard to OMICS signout across one permanent record.

Molecular Pathology Track | ABOPM | BOOKS

Sources

Claim Source Link
Variant classification follows the ACMG/AMP evidence framework Richards et al., "Standards and guidelines for the interpretation of sequence variants," Genetics in Medicine 2015 ncbi.nlm.nih.gov
ClinVar aggregates clinical interpretations including pathogenic and VUS classifications Landrum et al., "ClinVar," Nucleic Acids Research ncbi.nlm.nih.gov
BRCA2 founder mutations recur in defined populations Roa et al., "Ashkenazi Jewish population frequencies for common mutations in BRCA1 and BRCA2," Nature Genetics 1996 nature.com
Spaced repetition improves long-term retention of medical knowledge Kang, "Spaced Repetition Promotes Efficient and Effective Learning," Policy Insights from the Behavioral and Brain Sciences 2016 journals.sagepub.com

Chapter 12 — Clinical Oncology Track

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The Clinical Oncology track runs 10 cases of drug-variant matching across four resource tiers, from an academic medical center with full trial access to a community hospital in the Eastern Caribbean working from an essential medicines list. The oncologist treats the patient, not the variant — but the variant determines the treatment and the resource tier determines which treatments are available. The conventional read is that the right drug is the same everywhere. The actual rule is that the same variant yields four different treatment plans.

Treating the Patient, Not the Variant

The oncologist treats the patient, not the variant. But the variant determines the treatment — and the resource tier determines which treatments are available. An academic medical center has clinical trial access, expanded access programs, and a full formulary. A community hospital in the Eastern Caribbean has the essential medicines list and a monthly shipment.

Competency Areas

Targeted Therapy Matching: Actionable mutations mapped to approved targeted therapies. Resistance mutations and second-line selection. Companion diagnostics and testing requirements.

Immunotherapy Eligibility: PD-L1, TMB, MSI, and dMMR as biomarkers. The hard case: TMB-high but MSI-stable — does the patient qualify? The discordant biomarker case is the defining challenge.

Resource-Tier Treatment Planning: Same variant, four different treatment recommendations depending on setting. Academic (full formulary + trials), Regional (standard formulary + limited trials), Community (generics only, no trials), Remote (essential medicines, telemedicine oncology). The oncologist who can only recommend pembrolizumab has not demonstrated competency.

Caribbean Oncology: Drug availability across OECS formularies. Referral network optimization. Traditional medicine interactions. Cultural navigation for treatment adherence. The grandmother who trusts soursop more than a chatbot — that's a real case, and the oncologist must navigate it.

Where this matters is the resource tier: the same actionable variant produces four different treatment recommendations depending on setting, so the competency the track certifies is matching the drug to the patient's reality, not just to the genome.

How candidates encounter this track

Three of the four ABOPM learning apps deliver Clinical Oncology to the candidate's device. ANKI carries a dedicated oncology deck — drug-variant pairs, resistance mutations, biomarker thresholds — rehearsed on a spaced-repetition schedule. ONCO carries the tumor-board view itself: a forum where a candidate watches credentialed physicians adjudicate a real case in real time, and a window into the BCAM cohort whose monthly volume gives the conversation its evidence base. OMICS carries the upstream layers that feed any treatment decision — the transcriptome, the proteome, the variant calls that determine whether a patient is a candidate for a targeted therapy at all. CASE does not carry oncology bundles at present. The candidate moves between these surfaces without losing their place. A treatment thread opened on ONCO can call back to the variant interpretation on OMICS, then drop a card into the ANKI deck for the next review cycle.

For the full delivery architecture, see Chapter 16.

CANONIC compiles oncology competency across every resource tier, so the same variant that earns a trial in an academic center still earns a governed treatment plan in a remote Caribbean clinic.

Clinical Oncology Track | ABOPM | BOOKS

Sources

Claim Source Link
Targeted therapy is matched to actionable mutations and companion diagnostics NCI, "Targeted Cancer Therapies" cancer.gov
PD-L1, TMB, and MSI/dMMR are immunotherapy biomarkers, and they can be discordant Marabelle et al., "Association of tumour mutational burden with outcomes in patients treated with pembrolizumab," Lancet Oncology 2020 thelancet.com
The WHO Model List of Essential Medicines constrains formularies in low-resource settings World Health Organization Model List of Essential Medicines who.int
Treatment adherence is shaped by cultural context and traditional medicine use WHO, "Adherence to long-term therapies: evidence for action" who.int

Chapter 13 — Pharmacogenomics Track

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The Pharmacogenomics track runs 8 cases, one for every CPIC Level A gene-drug pair that kills when the dose is wrong. A CYP2D6 ultrarapid metabolizer converts codeine to morphine so fast a standard dose becomes an overdose; an HLA-B*58:01 carrier given allopurinol for gout can develop Stevens-Johnson syndrome, a reaction more common in Afro-Caribbean than European populations. The conventional read is that the right drug is enough. The actual rule is that the right drug at the wrong dose is the wrong drug.

The Wrong Dose Is the Wrong Drug

The right drug at the wrong dose is the wrong drug. CYP2D6 ultrarapid metabolizers convert codeine to morphine so fast that a standard dose becomes an overdose. DPYD-deficient patients given fluorouracil can die from the first cycle. HLA-B*58:01 carriers given allopurinol for gout can develop Stevens-Johnson syndrome — a preventable, devastating reaction that is more common in Afro-Caribbean populations than in the European populations where the drug was developed.

The Eight Cases

Each case covers one CPIC Level A gene-drug pair:

Gene Drug What goes wrong
CYP2D6 Codeine/Tramadol Ultrarapid metabolizer → morphine toxicity
CYP2C19 Clopidogrel Poor metabolizer → stent thrombosis
CYP2C9 + VKORC1 Warfarin Dose sensitivity → bleeding or clotting
HLA-B*58:01 Allopurinol Carrier → Stevens-Johnson syndrome
HLA-B*15:02 Carbamazepine Carrier → SJS/TEN
DPYD Fluorouracil/Capecitabine Poor metabolizer → fatal toxicity
TPMT/NUDT15 Azathioprine/6-MP Poor metabolizer → myelosuppression
SLCO1B1 Simvastatin Transporter variant → myopathy

Caribbean Screening

HLA-B*58:01 prevalence is higher in Afro-Caribbean populations than in European populations. Allopurinol is widely prescribed for gout across the Caribbean. Population-wide screening costs less than treating one case of Stevens-Johnson syndrome. The pharmacogenomics case for Caribbean screening is not academic — it is actuarial.

Phenoconversion

The advanced concept: drug-drug interactions that change a patient's metabolizer phenotype. A normal CYP2D6 metabolizer taking fluoxetine (a CYP2D6 inhibitor) becomes a functional poor metabolizer. If that patient is then given codeine, the expected ultrarapid metabolism doesn't happen — but neither does the expected normal metabolism. Phenoconversion is the real-world pharmacogenomics problem that single-gene testing misses. Where this matters is the dose: phenoconversion proves that a correct gene-drug pair can still kill at the wrong dose, so the track certifies dynamic prescribing rather than a static genotype lookup.

How candidates encounter this track

Two of the four ABOPM learning apps deliver Pharmacogenomics to the candidate's device. OMICS carries the metabolome and microbiome views — the layers where a drug-gene interaction first becomes visible to a clinician — and the integrated metabolizer phenotype across CYP2D6, CYP2C19, DPYD, TPMT, and the other CPIC Level A genes that drive the eight cases. ANKI carries a dedicated pharmacogenomics deck: the dose-response thresholds, the phenoconversion rules, the population frequencies that shift the prior on every prescribing decision. ONCO and CASE do not carry pharmacogenomics bundles at present, though pharmacogenomic content surfaces inside their oncology and counseling threads when the variant is at issue. A candidate can study a CYP2D6 ultrarapid case on ANKI in the morning and watch the same patient's metabolome render on OMICS at signout that afternoon. The link between the deck and the variant view is the permanent record.

For the full delivery architecture, see Chapter 16.

CANONIC compiles every CPIC Level A gene-drug pair into a governed pharmacogenomics track, so the dose that turns codeine into an overdose is caught before the prescription is written.

Pharmacogenomics Track | ABOPM | BOOKS

Sources

Claim Source Link
CPIC publishes Level A gene-drug guidelines for actionable prescribing Clinical Pharmacogenetics Implementation Consortium guidelines cpicpgx.org
CYP2D6 ultrarapid metabolizers convert codeine to morphine, risking toxicity Crews et al., "CPIC Guideline for CYP2D6 and Codeine Therapy," Clinical Pharmacology & Therapeutics ncbi.nlm.nih.gov
HLA-B*58:01 carriers risk allopurinol-induced Stevens-Johnson syndrome, with higher prevalence in some ancestries Saito et al., "CPIC Guidelines for HLA-B and allopurinol," Clinical Pharmacology & Therapeutics ncbi.nlm.nih.gov
Drug-drug interactions cause phenoconversion of metabolizer phenotype Shah & Smith, "Addressing phenoconversion: the Achilles' heel of personalized medicine," British Journal of Clinical Pharmacology 2015 ncbi.nlm.nih.gov

Chapter 14 — Genetic Counseling Track

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The Genetic Counseling track runs 8 cases weighted 60% on counseling rationale and 40% on classification accuracy, because the human dimension is the part the classifier cannot resolve. The defining case is a grandmother in Trinidad who does not want to know whether she carries a BRCA1 mutation, because knowing would mean her daughter has to know too. The conventional read is that the variant classification is the work. The actual rule is that the hard part is what you say next.

The Human Dimension

The genetic counselor navigates what no algorithm can: a grandmother in Trinidad who doesn't want to know if she carries BRCA1 because knowing would mean her daughter has to know too. A family where paternity is discovered incidentally during cascade testing. An adolescent whose parents want predictive testing for a condition that won't manifest until age 40.

Beyond Classification

Cases are weighted 60% counseling rationale, 40% classification accuracy. A perfect classifier who cannot counsel is not a competent genetic counselor. The variant classification is the easy part. The hard part is what you say next.

Cultural Competency

Four of eight cases are dedicated cultural competency cases:

  • Caribbean family structures: extended family, matriarchal decision-making, multiple households across islands
  • Religious considerations: Jehovah's Witness and blood products, Muslim consanguinity, spiritual healing traditions
  • Immigration status fears: undocumented patients who avoid healthcare because they fear deportation
  • Language barriers: Spanish-speaking Caribbean communities where medical terminology has no direct translation

Ethical Reasoning

  • Non-paternity: Discovered during family segregation testing. Who do you tell? What are your obligations?
  • Right not to know: Patient completed testing but now refuses results. The data exists. The patient doesn't want it.
  • Adolescent testing: Parents request predictive testing for adult-onset BRCA1 in their 14-year-old daughter. Guidelines say wait. Parents say now.
  • Duty to warn: You identify a pathogenic variant with implications for the patient's siblings, who are not your patients.

Every ethical case has a governance constraint: the counselor must document their reasoning in the evidence chain. The rationale is the exam. Where this matters is the governance constraint: because every ethical decision is documented in the evidence chain, the counselor's reasoning — not just the classification — becomes the auditable, governed object the board certifies.

How candidates encounter this track

Two of the four ABOPM learning apps deliver Genetic Counseling to the candidate's device. CASE is the primary surface — peer-reviewed case bundles, each one signed off by two reviewers before publication, that walk a counselor through the conversation the algorithm cannot have. OMICS carries the upstream variant work — the genome view a counselor opens before the conversation begins, and the exposome view that places a result inside the family's lived context. ANKI and ONCO do not carry genetic-counseling bundles at present, though counseling content surfaces inside their decks and threads whenever the result is uncertain enough to require a conversation. A counselor preparing for a Wednesday clinic session pulls a CASE bundle in the morning, reads the OMICS variant interpretation alongside it, and walks into the room with the evidence chain already assembled.

For the full delivery architecture, see Chapter 16.

CANONIC governs the genetic-counseling track so the counselor's rationale travels with the variant, turning the conversation no algorithm can have into an auditable part of the certification.

Genetic Counseling Track | ABOPM | BOOKS

Sources

Claim Source Link
Genetic counseling is a communication process addressing the human and familial implications of genetic findings National Society of Genetic Counselors, definition of genetic counseling nsgc.org
Predictive testing of minors for adult-onset conditions is generally deferred per professional guidelines American Society of Human Genetics statement on genetic testing in children and adolescents ncbi.nlm.nih.gov
A duty to warn at-risk relatives raises recognized ethical and legal tensions Offit et al., "The 'duty to warn' a patient's family members about hereditary disease risks," JAMA 2004 jamanetwork.com
Cultural humility shapes effective genetic counseling across diverse communities Tervalon & Murray-García, "Cultural Humility Versus Cultural Competence," J Health Care Poor Underserved 1998 ncbi.nlm.nih.gov

Chapter 15 — Caribbean Health Governance Track

inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

The Caribbean Health Governance track runs 5 cases on precision medicine for 44 million Caribbean people, a population that cannot be an afterthought in a system designed for 330 million Americans. It certifies competency across small island developing states — fragmented populations, sovereign borders, and a national pharmacy that fits in a shipping container. The conventional read is that small islands lose to big continents. The actual rule is that small islands beat big continents when governance is compiled, not negotiated.

Why Small Islands Win

This track exists because precision medicine for 44 million Caribbean people cannot be an afterthought in a system designed for 330 million Americans. The Caribbean Health Governance track certifies competency in the unique challenges of implementing precision medicine across small island developing states: fragmented populations, limited infrastructure, sovereign borders, and healthcare systems where the entire national pharmacy fits in a shipping container.

Sou-Sou Mutual Health Insurance

The sou-sou is a Caribbean rotating savings system older than any insurance company. Five cases teach the candidate to apply sou-sou actuarial reasoning to mutual health insurance for small population pools (5,000-50,000 lives). Why the Affordable Care Act doesn't scale to an island of 50,000 people. Why a centralized national health service doesn't work when the nearest specialist is a plane ride away. How COIN-based governance creates a health insurance product that works at Caribbean scale.

Referral Network Optimization

When to treat locally, when to refer off-island, when telemedicine is the referral. Hub-and-spoke: tertiary centers in Trinidad and Barbados, community sites across OECS nations, Florida as US referral destination. The air ambulance case: cost, logistics, clinical criteria — and who governs the decision.

Telemedicine Governance

Not every island has reliable internet. Licensure across sovereign borders requires OECS mutual recognition. Data sovereignty means patient data stays in the originating nation. The architecture must work asynchronously for intermittent connectivity and mobile-first for smartphone-dominant populations.

Data Sovereignty

The most important competency: understanding that Caribbean patient data belongs to Caribbean nations. Federated architecture delivers intelligence without exporting data. Cross-border research requires multi-sovereign institutional review board governance. The CANONIC governance compiler is a sovereignty enforcement mechanism — it doesn't move data, it moves governed intelligence. Where this matters is data sovereignty: because the compiler federates governed intelligence instead of exporting patient data, small Caribbean nations get precision medicine without surrendering the records to a continent that never built reference panels for them.

How candidates encounter this track

Three of the four ABOPM learning apps deliver Caribbean Health Governance to the candidate's device. ANKI carries a deck dedicated to the track — mutual-insurance actuarial reasoning, referral-network thresholds, regional pharmacy formularies, and the regulatory shape of cross-border telemedicine. ONCO carries the CaribChat view, a regional thread where Caribbean clinicians work across sovereign borders without exporting patient data. OMICS carries the exposome view, the layer that puts a patient's environmental and social context next to the molecular result — work that matters most where reference panels were never built. CASE does not carry Caribbean-governance bundles at present, though the conversation surfaces inside CASE whenever a counselor crosses an inter-island border. The four apps share one permanent record, and that record stays inside the nation that generated it. CaribChat operates independently across the Caribbean; no formal partnership or memorandum of understanding exists with any regional intergovernmental body.

For the full delivery architecture, see Chapter 16.

CANONIC compiles Caribbean health governance so 44 million people on small islands get precision medicine that keeps their data sovereign and their governance intact.

Caribbean Health Governance Track | ABOPM | BOOKS

Sources

Claim Source Link
Small island developing states face distinct health-system constraints UN Office of the High Representative, Small Island Developing States unohrlls.org
The sou-sou is a Caribbean rotating savings and credit association Wikipedia, "Susu (informal loan club)" en.wikipedia.org
Data sovereignty asserts that data is governed by the laws of the nation where it is collected Wikipedia, "Data sovereignty" en.wikipedia.org
Cross-border research requires multi-jurisdiction institutional review board oversight Council for International Organizations of Medical Sciences, International Ethical Guidelines cioms.ch

Chapter 16 — Four Apps for Board Preparation

chapter: 16 title: Four Apps for Board Preparation track_binding: cross-cutting _generated: true inherits: hadleylab-canonic/CONTENT/BOOKS/ABOPM references: [I-25, G-1]

A medical student preparing for the American Board of Precision Medicine carries 4 apps on her phone — ANKI, CASE, ONCO, and OMICS — one for flashcards, one for faculty-reviewed cases, one for tumor-board discussions across three countries, and one for credentialed variant interpretation. The four apps share one permanent record. The conventional read is that board prep is a stack of study tools. The actual rule is that the four apps are one governed learning surface.

The Four Apps

A medical student preparing for the American Board of Precision Medicine has four apps on her phone. One drills her on flashcards. One walks her through real patient cases reviewed by two faculty. One lets her sit in on tumor-board discussions from cancer centres across three countries. One shows her how a credentialed geneticist would call a tricky variant.

The apps are called ANKI, CASE, ONCO, and OMICS. They are free to read. Authors — the clinicians and faculty who write the content — earn contribution credit each time the community uses what they wrote. The four apps share one permanent record, so a flashcard reviewed inside ANKI, a case published inside CASE, a discussion in ONCO, and a variant called in OMICS all sit on the same trail of evidence a candidate, a reviewer, or a board member can audit.

The Diagram

The four apps for ABOPM board preparation

Source: BUSINESS/DEALS/ABOPM/figures/fig_nex_fleet_learning.mmd. The figure is rebuilt from the same source files that drive the apps themselves; the four app names — ANKI, CASE, ONCO, OMICS — are what a candidate sees on the icon.

Architecture: Three Composition Classes

NEX architecture: axiom orchestrator, specialty composition, triad modality

Source: SERVICES/NEX/figures/fig_nex_architecture.mmd. Generated from each live NEX's composition_class: declaration; see SERVICES/NEX/CANON.md § COMPOSITION_CLASS_GOVERNS_FEDERATION.

The four apps above are organized by a contract, not a coincidence. Every governed NEX falls into one of three composition classes, declared once in its CANON.md and verified on every build:

  • Axiom orchestrator (one nex, reserved to LAUDENEX) — the axiom substrate of the whole platform. Two GUIs over the same substrate: galaxy.canonic.org is the viz-network — a graph view of every governed axiom across every scope, filtered by the authenticated user's federation manifest. laude.canonic.org is the chat — a personalized conversation that opens with the next blocker pre-computed from the user's VITAE, the scope's CANON, the active ROADMAP, and the recent LEDGER tail. The two GUIs are modes of one entity: within a single scope, the chat renders as a sidebar attached to the graph view; when the conversation moves to promoting or demoting axioms across scopes, the sidebar escalates to a full-page session because the action will touch multiple CANON files at once.

  • Specialty composition (one nex per board specialty — ONCONEX today; CARDIONEX, NEURONEX, GENCXNEX as future siblings) — a thin orchestrator that federates the universal triad below it for the cases its specialty actually sees. ONCONEX renders a tumor-board view over the same three modality apps a cardiology future-sibling would render a clinic-board view over. Adding a specialty is one CANON.md declaration; no new modality code is ever written.

  • Triad modality (the universal three — ANKINEX, CASENEX, OMICSNEX) — student spaced-repetition, clinician peer-signed cases, and provider variant interpretation, respectively. These are the leaf nodes of the federation graph: every specialty NEX federates the full triad, the triad federates no one. The modalities are stable; the specialties compose them differently.

This is the contract a future operator reads to know how to add a specialty: write one CANON.md, declare composition_class: specialty_composition, list the triad in answers_federated_from:, and the rest of the apparatus — the chapter, the verify gates, the figure, the federation graph — extends without further code.

Current live membership:

Composition class Live nex-instance(s)
axiom_orchestrator (the galaxy substrate) LAUDENEX
specialty_composition (orchestrates the triad) ONCO, RAREDXNEX
triad_modality (universal student/clinician/provider) ANKI, CASE, OMICS

ANKI — Flashcards You Can Take Anywhere

Who builds it. Practising clinicians, fellows, and senior residents draft the cards. Two faculty reviewers approve every card before it ships to the community deck. The cards are written against the ABOPM Foundation curriculum — molecular pathology, oncology, pharmacogenomics, genetic counseling — and they are versioned the way a textbook chapter would be.

What a candidate does with it. A candidate opens the app on the train, taps Hard on a card about a BRCA2 variant in a 41-year-old with triple-negative breast cancer, and moves on. Behind the scenes the app remembers which cards the candidate finds difficult, and so do the other candidates using it. Cards that everyone struggles with rise to the top of the next review.

What is verified about the content. Each card carries the names of the author and the two faculty reviewers, the date it was last edited, and a link to the evidence it was built from — a ClinVar entry, a peer-reviewed paper, a CPIC guideline. A card without an evidence link does not ship.

How authors earn credit. An author earns contribution credit each time the community uses one of their cards. The credit is recorded in a shared running record alongside the card review itself; an author can show what they have contributed without exposing which candidate reviewed which card.

Current state. Live since 2026-04-19, with the first author credit paid to @marcus_freeman.

Board tracks it covers. Caribbean Governance, Molecular Pathology, Oncology, Pharmacogenomics.

CASE — Real Cases, Reviewed by Two Faculty

Who builds it. Practising clinicians write the cases. Each case is reviewed and signed by two faculty before it is published, and the anchor partnership is with the medical genetics service at Mount Sinai Medical Center in Miami Beach, where Irman Forghani, MD, FACMG, leads the clinical review. Patient identity is stripped before a case ever enters the editor.

What a candidate does with it. A candidate scrolls a case at two in the morning, the way one scrolls anything else at two in the morning, except the case is a 62-year-old with new-onset jaundice and a variant of uncertain significance. The candidate works the case to the diagnosis, sees what the reviewing faculty would have done, and flags a step that was unclear so the next reader benefits.

What is verified about the content. Every published case carries the author, the two reviewers, an evidence trail (ACMG criteria, ClinVar IDs, professional society guidelines), an ancestry disclosure, and a verified removal of identifying information. The cases live as signed bundles; nothing can be edited after publication without a new version and a new signature.

How authors earn credit. Authors earn credit when a case is published, and again when candidates mark the case as having helped them learn. Reviewers earn credit for sign-off. The credits accumulate in the same shared record as the case bundles.

Current state. In private beta with Mount Sinai Medical Center's genetics service. Public authoring opens once the first peer-reviewed case is signed.

Board tracks it covers. Genetic Counseling.

ONCO — A Tumor Board Without Borders

Who builds it. Oncologists, pathologists, and genetic counsellors post the threads. The discussions are threaded by cancer type, stage, biomarker, and trial eligibility. Centres from San Francisco to Trinidad to Madrid contribute; nothing is locked to one institution's network.

What a candidate does with it. A candidate filters tumor-board threads by stage and receptor and follows a breast-cancer staging debate that began in San Francisco at 7 a.m. and picked up an oncologist in Trinidad before lunch. When a thread references a variant, the candidate can pull the matching interpretation from OMICS without leaving the conversation. When a thread references a flashcard, the matching ANKI card is one tap away.

What is verified about the content. Each post carries its author, their institution, and the evidence anchors they cite (NCCN guideline, ClinicalTrials.gov ID, peer-reviewed reference). Cases referenced in a thread are those published in CASE, with the same two-faculty sign-off and patient identity removed.

How authors earn credit. Contributors earn credit when a thread accumulates engagement from credentialled clinicians past a published threshold. The intent is to reward expertise, not volume — a small number of well-anchored posts outweighs a flood of opinion.

Current state. Open for tumor-board discussion now. The first author credit is expected during Breast Cancer Awareness Month.

Board tracks it covers. Caribbean Governance, Oncology.

OMICS — Variant Interpretation as a Teaching Tool

Who builds it. Board-certified clinical geneticists author the interpretations across seven layers — genome, transcriptome, proteome, epigenome, metabolome, microbiome, and exposome. Authoring partnerships are in place or pending with laboratory partners covering each of the seven layers.

What a candidate does with it. A candidate queries a variant — say, a missense change in BRCA1 — and the app shows the interpretation a credentialled geneticist would deliver, the ACMG criteria invoked, the ClinVar entry the call was anchored to, and the relevant CPIC guideline if a pharmacogenomic implication exists. The candidate sees the reasoning, not just the verdict.

What is verified about the content. Every interpretation is signed by a credentialed geneticist and anchored to ClinVar and the ACMG variant-classification criteria. Pharmacogenomic implications cite CPIC. An interpretation without the anchors does not publish.

How authors earn credit. Geneticists earn credit for each interpretation that is published and again when candidates mark the interpretation as having taught them something. The seven omics layers each carry their own running record so a reader can trace any interpretation to its source.

Current state. Open for reading now. The first credentialed variant interpretation is expected by December 2026.

Board tracks it covers. Caribbean Governance, Genetic Counseling, Molecular Pathology, Oncology, Pharmacogenomics.

How the Four Work Together

The four apps share one permanent record and a set of direct connections. A tumor-board thread on a tricky breast-cancer case in ONCO can pull in the matching ANKI flashcards and the matching variants from OMICS without leaving the conversation. A peer-reviewed case in CASE can cite the same variant interpretation a candidate just queried in OMICS. The same author credit a clinician earns for writing a case shows up next to the same clinician's flashcards.

The connections are declared in the source files for each app and re-checked every time the documentation is rebuilt. The pairs the four apps are currently sharing with one another:

  • ANKI shares with ONCO
  • OMICS shares with ONCO

Multiledger Learning — Community Asks, ABOPM Answers

The four apps run on a shared, append-only record. That record has two channels, and the difference between them is the spine of how ABOPM certification works in practice.

The community is the source of the questions. A medical student taps Hard on a flashcard. A primary-care doctor in Trinidad posts a thread asking who in Guyana actually answers a referral call. A patient in the MammoChat community asks how a CDK4/6 inhibitor was chosen in the first line. A reader of a case marks it as having taught them something. None of these events requires a credential. They are open writes, anonymized where appropriate, and they make a permanent record of what is unclear, hard, or unsolved in real practice.

ABOPM credentialing is the source of the answers. A credentialed clinical geneticist signs a variant classification with the ACMG criteria they invoked. Two ABOPM-credentialed faculty co-sign a real, de-identified case before it publishes. A faculty author signs the cards that ship into the next deck. These are credentialed writes. The app checks the credential before the ledger ever sees the event. The same record that captures the question now captures the answer next to it.

The two channels are kept separate on disk so the audit trail is always legible: a question can be read without an answer next to it (an open question), and an answer can be traced back to the question that prompted it. The shape of the channels for each app:

App Questions (open-write) Answers (ABOPM-credentialed)
ANKI POST_THREAD, POST_REPLY, ANKI_DIFFICULTY_SIGNAL, CASENEX_CASE_UTILITY_SIGNAL ANKI_CARD_REVIEW, ANKI_CARD_DOWNLOAD
CASE POST_THREAD, POST_REPLY, CASENEX_CASE_UTILITY_SIGNAL, CASECHAT_SESSION_STARTED CASE_DRAFTED, CASE_PEER_REVIEW_SIGNED, CASE_PUBLISHED, CASECHAT_VALIDATION_RECORDED
ONCO POST_THREAD, POST_REPLY, VOTE_GIVEN federated from OMICS, CASE, ANKI
OMICS POST_THREAD, POST_REPLY CLASSIFY, INTERPRET, ANNOTATE

A worked example. A patient in MammoChat asks the community why her oncologist proposed a CDK4/6 inhibitor over chemotherapy as first-line therapy — a question. A medical oncologist in Madrid replies; the reply is itself a question (a clarification request about the patient's PIK3CA status). A credentialed geneticist in OMICS classifies the patient's PIK3CA variant against ACMG criteria — an answer. A practising clinician in Miami drafts a real, de-identified case mirroring the scenario in CASE; two ABOPM-credentialed faculty co-sign it — an answer. The case sources a flashcard in ANKI authored against the ABOPM Foundation curriculum; a faculty author signs the card — an answer. A medical student taps Hard on that card the following week — a new question, now linked to the original thread and to the answer that retired it.

That entire chain — one patient question through to one student difficulty signal — sits as nine connected entries in the same shared record. The credentialed entries can never be edited after the signature lands; the community entries can be voted on, replied to, and superseded but never silently removed. The result is a learning surface where the community is honest about what it does not yet know, and where the certification the ABOPM grants is the gate that turns an open question into a governed answer.

The Board Mandate

ABOPM certifies physicians in precision medicine. The Foundation curriculum has five tracks; every app contributes to multiple tracks, and every track is served by more than one app.

Track Chapter Apps that contribute
Molecular Pathology Ch. 11 ANKI, OMICS
Oncology Ch. 12 ANKI, OMICS, ONCO
Genetic Counseling Ch. 14 CASE, OMICS
Pharmacogenomics Ch. 13 ANKI, OMICS
Caribbean Governance Ch. 15 ANKI, OMICS, ONCO

For Candidates: How to Start

  • ANKI — download the deck and study offline. The cards are free; spaced-repetition tracks what you find hard.
  • CASE — read the published cases. Authoring opens to credentialed clinicians during the Mount Sinai private beta.
  • ONCO — read the tumor-board threads. To post, you need credentialing through the board.
  • OMICS — query a variant and read the interpretation. Authoring is open to board-certified geneticists.

For Faculty: How to Contribute

Faculty contributors enrol as peer reviewers and authors through the same governed identity (board-issued credential, ORCID, institutional email). Every contribution must carry an evidence anchor — a ClinVar entry, an ACMG criterion, a peer-reviewed paper, a professional society guideline — and every signed case or variant interpretation carries two faculty signatures. Contribution credit accrues automatically as the community uses the content.

What Is Funded and What Is Next

ANKI is live, with the first author credit paid for a flashcard contribution. CASE is in private beta with Mount Sinai Medical Center's genetics service. ONCO is open for tumor-board discussion; the first author credit is expected during Breast Cancer Awareness Month in October. OMICS is open for reading; the first credentialed variant interpretation is expected by the end of 2026.

Where this matters is the shared record: because all four apps write to one permanent, append-only learning surface, what is funded today and what ships next both extend the same governed trail a candidate, reviewer, or board member can audit.

OMICS authoring partnerships across the seven omics layers are expected to be closed by 2026-07-27. The current state of each layer's partner conversations:

Laboratory partner Omics layer Where the conversation stands
EPOCH epigenome Drafted; under negotiation
GENOVA metabolome Drafted; under negotiation
ILLUMINA genome Drafted; under negotiation
MOLECULAR_U proteome Terms agreed; signature pending
ORAL_DNA microbiome Drafted; under negotiation
PROGENEDX transcriptome Drafted; under negotiation
VIBRANT_WELLNESS exposome Drafted; under negotiation

Cross-References

CANONIC compiles the four ABOPM apps into one governed, auditable learning surface, so every flashcard, case, tumor-board thread, and variant interpretation extends the same permanent record a board member can verify.

Four Apps for Board Preparation | ABOPM | BOOKS

Sources

Claim Source Link
Spaced-repetition flashcards improve long-term retention of medical knowledge Kang, "Spaced Repetition Promotes Efficient and Effective Learning," Policy Insights from the Behavioral and Brain Sciences 2016 journals.sagepub.com
Variant interpretation follows the ACMG/AMP classification criteria Richards et al., "Standards and guidelines for the interpretation of sequence variants," Genetics in Medicine 2015 ncbi.nlm.nih.gov
ClinVar aggregates clinical variant interpretations anchored by submitters Landrum et al., "ClinVar," Nucleic Acids Research ncbi.nlm.nih.gov
CPIC publishes pharmacogenomic prescribing guidelines cited in interpretations Clinical Pharmacogenetics Implementation Consortium cpicpgx.org
ORCID provides a persistent identifier for author and contributor credit ORCID, "What is ORCID?" orcid.org
NCCN guidelines anchor oncology tumor-board treatment decisions National Comprehensive Cancer Network nccn.org

Every claim in this chapter rebuilds from the source files named above. Regenerate via ~/.canonic/bin/gen-nex-fleet-doc.