Clinical Laboratory Diagnostics

AssayCore Med

AssayCore Med is built for clinical diagnostic laboratories that process thousands of samples daily. The system covers every step from order creation through result delivery, with built-in QC (Westgard rules, Levey-Jennings), bidirectional instrument connectivity via HL7v2 and ASTM E1394, and FHIR R4 export for HIS/EMR integration. Delta check, critical value alerts with escalation, and TAT monitoring are all natively supported.

Open Live Demo Request Demo

Compliance & Standards

ISO 15189:2022ISO/IEC 17025:2017FDA 21 CFR Part 11IVDR EU 2017/746CLIA (42 CFR 493)CAP Laboratory AccreditationHL7v2HL7 FHIR R4ASTM E1394Westgard QCIFCC reference intervalsGDPRHIPAA-ready

Platform Modules

Complete feature set covering every stage of your laboratory workflow.

Patient & Order Management

Patient registration with full demographic and insurance data
Order templates and test complexes (panels)
CITO / STAT priority flags with worklist highlighting
Pre-billing preview per order
Repeat order from previous visit

Sample Lifecycle

Code128 and DataMatrix barcode generation per tube
Aliquoting and split/merge operations
Routing to workstations by test type
Defect catalog with photo evidence and cancel/restore
Temperature and transport log

Results & Verification

Manual and batch result entry
Auto-verification with rule engine
Result history trend charts
Delta check (absolute and percentage) per analyte
Critical value confirmation with read-back dialog

Quality Control

All 13 Westgard rules (including R4s, 10x, 2of3-2s)
Levey-Jennings charts with exclude-point annotation
QC Convergence (new lot validation)
QC Establishment (mean/SD setup period)
Control material reference (Bio-Rad, Randox)

Instrument Integration

HL7v2 bidirectional connectivity
ASTM E1394 analyzer protocol
Worklists export to instruments
FHIR R4 Observation and DiagnosticReport
Real-time result import with mapping rules

Reporting & Compliance

PDF reports via Thymeleaf with QR verification code
Excel batch export with Apache POI
Full audit trail (21 CFR Part 11 e-signature ready)
TAT monitoring dashboard with SLA alerts
ISO 15189 accreditation support

Pain Points We Solve

Concrete laboratory pain — quantified — and how AssayCore Med addresses each.

Manual QC interpretation delays each batch release

8-15 min/batch lost per run

Auto-Westgard engine evaluates all 13 rules in <50 ms and surfaces the failing rule in plain language

Critical-value callbacks are not consistently read back

23% missed read-back per CAP audit average

Read-back dialog with HCP signature, escalation timer, and Kafka event for audit trail

Delta-check flags are noisy and ignored by techs

False-positive rate up to 40% on legacy LIS

Z-score + isolation-forest hybrid trained on patient history reduces false alerts ~3x

HL7 mappings rot silently when analyzers are swapped

1-2 production incidents/month industry average

Version-controlled HL7v2 + FHIR R4 mapping bundles with diff view and rollback

TAT SLA monitoring is reactive, not proactive

Outliers found only at end-of-day report

Real-time TAT clock per analyte with predictive SLA-breach alerts 30 min before deadline

Reference-interval updates require IT tickets

2-6 weeks lag for IFCC/CLSI guideline changes

Lab-configurable reference ranges per analyte/age/sex, audit-logged, no deploy needed

ML Models in Production

Bundled CPU-only inference. Open datasets, transparent metrics, lab-tunable thresholds.

Model	Task	Dataset	Metric	Latency	Status
med.anomaly_detector	Detect abnormal lab results before verification	MIMIC-IV demo (HF: itu/mimic-iv, ~100 patients)	F1 0.91 (synthetic holdout)	15 ms (CPU)	Production
med.delta_check	Alert on sudden patient-result changes (mix-ups, clots)	MIMIC-IV demo + synthetic patient histories	AUC 0.93	12 ms (CPU)	Production
med.terminology_mapper	Map free-text test names to LOINC / ICD-10	LOINC 2.78 + ICD-10 WHO release	Top-1 accuracy 94.2%	8 ms (CPU)	Production
med.ddx_suggester	Suggest top-5 differential diagnoses for clinician review	DDxPlus (HF: aai530/ddxplus, 1.3M patients, CC-BY)	Top-5 accuracy 87.6%	45 ms (CPU)	Production
med.clinical_qa	Retrieval QA over MedMCQA corpus	MedMCQA (HF: openlifescienceai/medmcqa, MIT, 194K)	Recall@5 0.81	60 ms (CPU)	Production
med.drug_lab_interaction	Flag drug-lab interactions (e.g. biotin → troponin)	DDI Corpus (1,025 docs) + curated KB	Precision 0.88 / Recall 0.79	20 ms (CPU)	Planned
med.qc_anomaly	Detect QC drift before Westgard rules trigger	Internal Levey-Jennings runs (50K points, synthetic)	F1 0.93 on shift/trend detection	25 ms (CPU)	Production

Formulas & Calculations

Calculations that actually run in AssayCore Med — every formula links to its source standard.

Levey-Jennings z-score

z = (x − μ) / σ

Westgard 1981 / CLSI C24-Ed4

Westgard 1-3s / 1-2s warning

1-3s: |z| > 3 → reject; 1-2s: |z| > 2 → warn

Westgard QC multirule (1981, 13 rules)

Westgard R-4s (range across two levels)

R = z_high − z_low; reject if |R| > 4

Westgard multirule

Delta check (percentage)

Δ% = |x_t − x_(t-1)| / x_(t-1) × 100; flag if Δ% > threshold

CLSI EP33; configurable per analyte

Coefficient of variation (CV%)

CV = σ / μ × 100

ISO 15189 §5.5; IFCC

Sigma metric

σ_metric = (TEa − bias) / CV

Westgard sigma-VP (Six Sigma in clinical labs)

Total Error (TEa)

TE = |bias| + 1.65 × CV

CLIA 2024 Acceptable Performance Specifications

Why AssayCore Med vs Competitors

Honest side-by-side. Where competitors lead we say so; where we lead, we show how.

Competitor	Their offer	AssayCore Med
LabWare LIMS	$150-400K implementation, 9-month deploy, custom QC scripting	$19.9-29.9K starter / $99K pro, 14-21 days deploy, all 13 Westgard rules out-of-box
STARLIMS (Abbott Informatics)	Strong workflow but proprietary HL7 mappings tied to vendor consultants	Open HL7v2 + FHIR R4 mappings versioned in Git, lab-editable, vendor-neutral
LabVantage Medical Suite	Heavy Oracle/Java EE stack, complex upgrades, no native ML	Kotlin + PostgreSQL + Kafka + 7 production CPU-only ML models bundled
Epic Beaker	Tight EHR coupling, requires full Epic suite ($5-50M); no standalone option	Standalone or FHIR-integrated; deploy on-prem or SaaS multi-tenant
Sunquest LIS	Mature but US-centric, IVDR coverage limited, weak Spanish/Russian i18n	EN/ES/RU UI, IVDR EU 2017/746 compliant, multi-region accreditation maps

Who Uses AssayCore Med?

Hospital central labs processing 1,000–10,000 samples per day
Outpatient diagnostic networks (chain labs) with multi-site reporting
Reference laboratories with high-complexity testing
Clinics requiring HIS/EMR integration via FHIR

Integrations

HL7v2ASTM E1394FHIR R4Roche cobasSysmex XNSiemens AtellicaBeckman DxHKeycloak SSOKafkaMinIO

Ready to see AssayCore Med in action?

Get a personalized demo from our lab informatics specialists.

Request a Demo Open Live Demo