Five-Stage Diagnostic Report Lifecycle & Pathologist Autoverification Engine

1. Clinical Laboratory Workflow & Accreditation Governance

In modern hospital systems, commercial diagnostic chains, and specialized pathology reference centers, diagnostic laboratory results influence upwards of seventy percent of all clinical decisions. A failure in report governance—such as premature release of unverified results, unflagged panic values, or untracked manual edits—presents immediate risk to patient safety.

Accreditation authorities, including the College of American Pathologists (CAP), ISO 15189, and CLIA (Clinical Laboratory Improvement Amendments), mandate rigorous digital controls:

1. Deterministic Specimen Custody: Every specimen must transition through verifiable status gates from phlebotomy venipuncture to final archive.
2. Age- and Gender-Stratified Reference Intervals: Analyte reference ranges cannot rely on static adult male baselines; intervals must dynamically bind to the patient exact chronological age and biological sex.
3. Reflex Testing Automation: Abnormal findings (such as an elevated Thyroid Stimulating Hormone or a reactive viral screen) must programmatically schedule secondary confirmatory assays without requiring manual order re-entry.
4. Non-Repudiation & Cryptographic Integrity: Released laboratory findings must be cryptographically hashed to guarantee that printed vouchers or patient portal downloads represent an untampered record.

Orbit Health enforces these invariants through a deterministic five-stage report state machine combined with an automated post-save clinical evaluation engine. The underlying mechanics are grounded in the production modules under apps/api/src/modules/health/services/lab-reports.service.ts (orchestrating atomic updates, state transitions, and digital approvals), apps/api/src/modules/health/reflex.ts (evaluating autoverification and conditional reflex cascade criteria), the relational lab_reports table schema, and the staff review console at apps/health/src/app/lab-reports/page.tsx.

2. The Five-Stage Diagnostic Report Lifecycle State Machine

Every laboratory requisition traverses five discrete, auditable operational states:

2.1 State Definitions & Transition Criteria

2.2 Reversion Rules & Terminal Completion

Once a report reaches the COMPLETED state, it enters an immutable terminal condition. The persistence layer prohibits in-place updates, deletions, or status regressions back to earlier phases. If an analytical discrepancy is discovered post-release, the laboratory must initiate an Amended Report Workflow.

3. Atomic Analyte Result Entry & Dynamic Range Binding

A single diagnostic test panel (such as a Comprehensive Metabolic Panel or Complete Blood Count) consists of multiple atomic analyte components. The entry of result values—whether transmitted via the ASTM and HL7 analyzer edge gateway or recorded via manual bench DataGrid entry—triggers dynamic clinical evaluation.

3.1 Patient-Specific Reference Interval Selection (pickRangeForPatient)

Analyte reference intervals vary significantly across pediatric, adolescent, adult, and geriatric demographics, as well as between biological sexes. When saving analyte results, the system queries configured interval bands using a deterministic prioritization algorithm:

1. Exact Gender & Age Match: Selects the band where the patient biological sex matches exactly and their chronological age falls within minAgeYears and maxAgeYears.
2. Wildcard Gender Fallback: If no sex-specific band exists, selects an age-matched row marked with gender wildcard all, any, or both.
3. Any In-Band Row: Selects any interval matching the chronological age criteria.
4. Global Default Fallback: If no age-stratified criteria match, applies the first configured default range row.

3.2 Reference State Evaluator & Flag Categorization

The evaluator parses the numerical or qualitative value and compares it against normal reference bounds (min, max) and critical panic thresholds (critLow, critHigh):

Flag =
  LL  if v < critLow              (Critical Low Panic)
  HH  if v > critHigh             (Critical High Panic)
  L   if v < min                  (Abnormal Low)
  H   if v > max                  (Abnormal High)
  N   if min <= v and v <= max    (Normal Physiologic Range)

Every analyte result record snapshots its evaluated range minimum, range maximum, flag status, and panic indicator at the moment of entry. This snapshot preserves historical clinical validity even if laboratory reference tables are recalibrated months later.

4. Visual Diagnostic Lifecycle & Autoverification Architecture

The SVG architecture below details the dual-track verification pipeline, illustrating how routine normal specimens follow the automated autoverification track, while abnormal, panic, or delta-breached results are diverted to the pathologist review desk:

5. Diagnostic Report Verification Console Interface

The diagnostic report console enables consulting pathologists and supervising laboratory technicians to review analyte values against demographic intervals, inspect longitudinal trends, evaluate reflex additions, and sign off on verified findings:

6. Pathologist Autoverification Rules Engine

In high-throughput reference facilities handling thousands of blood tubes daily, manual pathologist review of every routine Complete Blood Count (CBC) or normal lipid profile creates unacceptable bottlenecks. The Orbit Health Autoverification Engine safely automates report approval when specimens meet rigorous mathematical constraints.

6.1 Strict Autoverification Qualification Rules

A report is automatically transitioned from IN_PROGRESS to APPROVED if and only if all of the following rules evaluate true:

autoverified = tenant.autoverifyEnabled
  AND report.results.length > 0
  AND missingComponents(report).length == 0
  AND report.results.every(r => r.flag == 'N' AND NOT r.critical AND NOT r.deltaBreach)
  AND (report.hIndex == null OR report.hIndex <= 100)
  AND (report.iIndex == null OR report.iIndex <= 60)
  AND (report.lIndex == null OR report.lIndex <= 120)

1. Full Completion Invariant: Every single analyte component belonging to all ordered test profiles must contain a valid, parsed result value. Zero unfilled cells are permitted.
2. Strict Normalcy Invariant: Every analyte flag must be strictly N. Any value evaluating to L, H, LL, or HH immediately aborts autoverification and routes the file to the manual pathologist worklist.
3. Delta Invariant: The longitudinal delta check between the current result and the patient historical baseline must not exceed configured percentage tolerances.
4. Interference Index Invariant: Spectrophotometric serum interference indices for hemolysis, icterus, and lipemia must fall below assay-specific interference thresholds.

6.2 Reflex Rule Engine Mechanics (reflexMatches)

When analytical findings point toward secondary pathology, clinical guidelines dictate ordering confirmatory tests. Rather than requiring the attending physician to submit a new requisition, the reflex engine evaluates incoming results against configured database rules:

// Reflex evaluation logic executed during post-save processing
export function reflexMatches(
  rule: ReflexRuleRow,
  componentName: string,
  value: string | null | undefined,
): boolean {
  if (!rule.isActive) return false;
  if (componentName.trim().toLowerCase() !== rule.triggerComponent.trim().toLowerCase()) return false;
  const raw = (value ?? '').trim();
  if (raw === '') return false;

  switch (rule.operator) {
    case 'contains':
      return !!rule.thresholdText && raw.toLowerCase().includes(rule.thresholdText.toLowerCase());
    case 'eq': {
      if (rule.thresholdText) return raw.toLowerCase() === rule.thresholdText.toLowerCase();
      const v = Number(raw);
      const t = Number(rule.thresholdValue);
      return Number.isFinite(v) && Number.isFinite(t) && v === t;
    }
    case 'lt':
    case 'gt': {
      const v = Number(raw);
      const t = Number(rule.thresholdValue);
      if (!Number.isFinite(v) || !Number.isFinite(t)) return false;
      return rule.operator === 'lt' ? v < t : v > t;
    }
    default:
      return false;
  }
}

When a reflex match occurs:

• The system calls addTestToReport, appending the reflex test panel to the active accession document.
• The accession barcode is re-queued in the instrument edge gateway worklist for automatic aspiration and testing.
• An audit entry REFLEX_FIRED is permanently recorded with the triggering rule identifier.

7. CLIA & ISO 15189 Compliant Report Amendment Governance

Medical reports released to clinicians are legally binding diagnostic documents. In the event of an erroneous entry, specimen misidentification, or retrospective clinician dispute, the original findings cannot be expunged.

7.1 Immutable Cryptographic Hashing at Completion

Upon transition to the COMPLETED state, the system constructs a canonical JSON representation of all sorted analyte components, values, flags, and units:

Canonical String = JSON.stringify(sortedResults.map(r => ({ cid, val, flg, crt })))

The persistence engine computes a SHA-256 cryptographic digest over this string, storing the result in column resultHash. Any subsequent alteration of database rows results in an immediate hash validation mismatch during public QR code verification or clinical audit.

7.2 The Amendment Chain Architecture

When an approved pathologist issues an amendment:

1. Original Record Stays Immutable: The original report remains in COMPLETED state, retaining its original resultHash and signature stamps.
2. Creation of Versioned Child: A new report record is created with incremented version number:

reportNumber_amended = reportNumber_original + "-A" + version

3. Lineage Binding: The child report populates column supersedesId with the parent report ID.
4. Mandatory Audit Reason: The pathologist must supply a clinical justification string (e.g. "Specimen rerun following instrument recalibration revealed adjusted potassium measurement"), which is permanently committed to the audit service log.

8. Diagnostic Lifecycle Failure Modes & Recovery Matrix

The following operational matrix details frontline recovery procedures for lifecycle blocking states, autoverify aborts, and reflex collisions:

9. Cross-Functional Platform Architecture Dependencies

The diagnostic report verification engine interfaces with complementary clinical, quality, and enterprise ledgers:

• Hardware Peripheral Integration: Analytical measurements are ingested directly from astm and hl7 analyzer edge gateway integration.
• Longitudinal Patient Safety: Panic escalations and delta check tolerances are governed by delta-check validation and critical value escalation.
• Microbiology Diagnostics: Complex culture and antibiotic susceptibility workflows integrate via microbiology culture identification and ast susceptibility.
• Referring Doctor Commissions: Report completion automatically accrues financial commissions in the general ledger documented in double-entry general ledger and financial periods.

10. Regulatory Compliance, Pathology Audit & Accreditation FAQs

How does the system ensure compliance with CAP requirement GEN.43800 regarding autoverification?

The platform enforces annual autoverification validation audit tracking. Every autoverified report records the exact software release commit, rule criteria version, and timestamps. If any single criterion fails, the autoverify algorithm terminates instantly, requiring signed human intervention.

Can a laboratory director disable autoverification for specific high-risk panels?

Yes. Autoverification is configurable on a per-panel and per-test basis. Laboratories routinely configure high-risk panels (such as tumor markers, cardiac troponin, or toxicology screens) to bypass autoverification entirely, mandating pathologist manual sign-off for 100% of samples regardless of value normalcy.

What happens to attending physician notifications when an amended report is released?

When an amended report reaches completion, the distribution outbox identifies all recipients who received previous versions. It dispatches a high-priority push notification and SMS alert explicitly labeled with AMENDED REPORT NOTICE, highlighting the altered analyte lines alongside the clinical reason for correction.

How are qualitative test results evaluated for autoverification?

Qualitative results (such as "Negative", "Non-Reactive", or "Normal") match against exact permitted textual tokens defined in the test catalog. If the observed text exactly matches the designated normal token, the flag evaluates to N and qualifies for autoverification, provided all other systemic invariants hold.