HL7AI Studio · Product

FHIR BI

An advanced clinical BI that taps your HL7 messages — V2.x, V3, FHIR and DICOM — in real time. Beyond rejection analysis, FHIR BI lets clinicians model real data and produce reports on patients' health status.

Get started Pricing
V2.x
HL7 messaging
V3
HL7 version 3
FHIR
R4 · R5
DICOM
imaging
Understanding your data

Behind every chart, an HL7 message

No IT background needed. An HL7 message is simply the "slip" one system sends another at each event: an admission, a lab result, a prescription. Here is, in plain terms, what the few codes used in the reports below actually mean.

A real HL7 slip, dissected — a lab result
MSH|^~\&|LABO|CHUM|BI|CHUM|20240607083000||ORU^R01 Header — the CHUM lab sends a result on June 7 at 8:30 am.
PID|1||1958031712^^^RAMQ||TREMBLAY^JEAN||19551103|M Identity — TREMBLAY Jean, health-card number in the RAMQ registry, born Nov 3 1955.
PV1|1|I|SI^12^01|||… Location — the patient is admitted to the Intensive Care Unit (ICU).
OBR|1||A1234567|CREAT^Renal panel Exam — a renal panel was ordered (request number A1234567).
OBX|1|NM|2160-0^Creatinine||298|umol/L|59-104|H Result — Creatinine at 298 µmol/L, above normal (59–104), flagged "H" for "high".

The vocabulary, at a glance

ADT
Patient movements — the patient's ins and outs: admission (A01), unit transfer (A02), discharge (A03). The trail of their movements through the hospital.
PID
Patient identity — name, date of birth and health-insurance number. What ties everything back to the right patient.
PV1-3
Care unit — the ward and bed where the patient is: ER, ICU, Medicine…
OBR-4
Requested exam — the ordered test or imaging: renal panel, troponin, chest CT…
OBX
The result — the test and its measured value — e.g. "Potassium: 6.4 mmol/L". The heart of clinical BI: every point on the charts comes from an OBX.
DG1
Diagnosis — the coded diagnosis (ICD-10). Used to build a cohort: "all diabetic patients".
RXE · RXA
Medications — the drug prescribed (RXE) then actually administered (RXA) to the patient.
SPM
Specimen — the specimen itself (tube, blood culture) and the time it was collected.
Real data, in real time

A living clinical dashboard

Every HL7 message feeds the dashboard as it arrives. Clinical indicators update continuously — no extraction, no data warehouse.

FHIR BI — Live clinical view LIVE
HL7 messages / day
48 219
▲ 6,2 %
Active patients
3 142
▲ 1,8 %
Critical results (OBX)
137
to follow
Rejection rate
2,3 %
▼ 0,4 %

Critical results (OBX) — last 7 days

OBX-8 flag = 'H' or 'HH'
LunMarMerJeuVenSamDim

Admissions by unit (PV1-3)

Today
318
ER
214
Medicine
176
Surgery
58
ICU
96
CLSC
HL7 source — where these figures come from OBX-8 · Critical results — every lab result carries an abnormal flag (OBX-8: "H" = high, "L" = low, "HH/LL" = critical). Criticals received per day are counted. ADT^A01 · PV1-3 · Admissions by unit — each admission message names the patient's receiving ward in field PV1-3. ACK · Rejection rate — share of messages refused by the receiving system, flagged by a negative acknowledgement (ACK AR/AE).
Clinical modeling

Cross one clinical variable against another, freely

Pick a measure and a dimension: FHIR BI does the rest. A clinician gets the answer in two clicks — without writing a single query.

"Where do critical lab values concentrate?"
Measure Critical results OBX-8
×
Dimension Care unit PV1-3
=
Answer The wards most at risk in 2 clicks
Horizontal axis : the care unit · height : number of critical results
Filter by facility, period or source system
From the chart to the original HL7 message in a single click

Critical OBX results by care unit

Last 30 days
41
ICU
34
ER
22
Medicine
15
Surgery
8
CLSC

Exam distribution (OBR-4)

100 exams
  • Biochimie 34%
  • Hematology 27%
  • Microbiologie 18%
  • Imaging 14%
  • Pathologie 7%

Critical results to follow

Live PID · PV1 · OBX correlation
Patient (PID-5) Unit (PV1-3) Test (OBX-3) Value Status
GAGNON, M.ICUPotassium6,4 mmol/LCritical
TREMBLAY, J.ERTroponine0,92 µg/LCritical
CÔTÉ, L.MedicineHémoglobine71 g/LLow
ROY, P.ERGlucose22,1 mmol/LHigh
BÉLANGER, S.NephrologyCréatinine298 µmol/LCritical
HL7 source — how this screen is built OBX-5 · OBX-8 · PV1-3 — the bars count, per care unit (PV1-3), the results whose value (OBX-5) is flagged critical (OBX-8). The chart shows where critical values concentrate. OBR-4 — the donut breaks activity down by requested exam type, read from the OBR-4 exam code. PID-5 · OBX-3 · OBX-5 — the table joins, live, the patient name (PID-5), the performed test (OBX-3) and its measured value (OBX-5).
Clinical reporting

Model the health status of a patient population

Tracking a diabetic patient cohort from lab results received over HL7. Every HbA1c and glucose value arrives in a result message (the OBX segment): FHIR BI automatically groups them by patient and computes the averages. The clinician builds the report just by picking variables — no technical skill required.

Report — Diabetic cohort (HbA1c, glucose) Auto-updated
Patients tracked
412
▲ 9
Average HbA1c
7,8 %
▼ 0,3 %
At target (< 7%)
38 %
▲ 4 %
Retinal screening up to date
64 %
to improve

HbA1c distribution (%) — cohort

Patient count per band
84
< 6,5
72
6,5–7
121
7–8
78
8–9
57
> 9

Cohort average glucose — 6 months (mmol/L)

Downward trend
JanFévMarAvrMaiJun
HL7 source — how this report is built DG1 · The cohort — the selected patients are those whose diagnosis (DG1) is diabetes, coded E10–E11 in ICD-10. OBX-3 · OBX-5 · HbA1c — the distribution reads the HbA1c value (OBX-5) of results identified by their test code (OBX-3), then splits them into control bands. OBX-5 · Glucose — the curve plots the cohort's monthly average of measured glucose values (OBX-5).

A heart-failure cohort rebuilt from HL7 messages. BNP is read from lab results (OBX) and LVEF from echocardiography reports. A readmission is flagged automatically when an already-discharged patient (discharge message, ADT A03) is re-admitted (admission message, ADT A01) within 30 days — all with no technical skill.

Report — Heart failure (BNP, 30-day readmissions) Auto-updated
Patients tracked
286
▲ 5
30-day readmission
18,2 %
▼ 1,4 %
BNP median
540 pg/mL
▼ 6 %
On optimal therapy (GDMT)
71 %
▲ 6 %

Distribution by LVEF

100 patients
  • Reduced (< 40%) 48%
  • Mid-range (40–49%) 22%
  • Preserved (≥ 50%) 30%

30-day readmission — 12 months (%)

Below the 20% target
JFMAMJJASOND
HL7 source — how this report is built OBR-4 · OBX-5 · FEVG — the ejection fraction is read (OBX-5) from echocardiography reports, identified by their exam code (OBR-4), then grouped into categories (preserved / mid-range / reduced). OBX-5 · BNP — BNP (or NT-proBNP) comes straight from lab results (OBX-5). ADT^A03 → A01 · 30-day readmission — FHIR BI pairs, for the same patient (PID), a discharge (ADT A03) followed by a new admission (ADT A01) occurring within 30 days.

A chronic-kidney-disease cohort built from labs received over HL7. Creatinine and estimated GFR (eGFR) come from lab results (the renal-panel OBX), and the KDIGO stage — the severity of kidney impairment — is computed automatically from those values.

Report — Chronic kidney disease (eGFR, KDIGO stages) Auto-updated
Patients tracked
1 042
▲ 23
Stage G4–G5
14 %
at risk
Mean eGFR
52 mL/min
▼ 1,8
Albuminuria screened
68 %
to improve

Distribution by KDIGO stage

Patient count
180
G1
360
G2
230
G3a
130
G3b
95
G4
47
G5

Cohort mean eGFR — 12 months (mL/min)

Slowed decline
JFMAMJJASOND
HL7 source — how this report is built OBR-4 · OBX-5 · Creatinine & eGFR — creatinine and estimated GFR are read (OBX-5) from the renal panel (OBR-4). KDIGO stage — FHIR BI automatically classifies each patient by eGFR: G1 (≥ 90) to G5 (< 15). No manual entry, no calculation by the clinician. OBX-5 · Albuminuria — the urine albumin-to-creatinine ratio is also read from results (OBX-5) to flag un-screened patients.

An oncology pathway reconstructed from the dates carried by HL7 messages. Each step is dated automatically: suspicion (exam-request date), diagnosis (pathology result), the multidisciplinary tumor board and treatment start. FHIR BI then computes the delay between each step.

Report — Oncology: care pathway delays (suspicion → treatment) Auto-updated
New cases (month)
84
▲ 6
Median suspicion→treatment
31 j
▼ 3 j
Within target (≤ 28 d)
62 %
to improve
Reviewed at tumor board
96 %
▲ 2 %

Median delay per step vs target (d)

From referral to first treatment
Suspicion → consult
7 j / 7 j
Consult → diagnosis
12 j / 14 j
Diagnosis → tumor board
6 j / 7 j
Tumor board → treatment
9 j / 7 j
≤ target > target Target

Median suspicion→treatment by site (d)

Last 30 days
Prostate
38 j
Lung
34 j
Colon
30 j
Breast
26 j
Hematology
22 j
HL7 source — how this report is built OBR-7 · OBX-14 · Step dates — each step is dated by its HL7 message timestamp: the exam request (OBR-7), the pathology result (observation date OBX-14), the tumor board and treatment start. Pathway delays — FHIR BI subtracts these dates to get each step's delay, and automatically compares it to the target (≤ 28 days). DG1 · Tumor site — the breakdown by site comes from the coded diagnosis (DG1, ICD-O topography).

Antimicrobial-resistance surveillance from antibiograms received over HL7. Identified organisms and their antibiotic susceptibility are read from microbiology results (OBX), and antibiotic consumption is derived from drugs prescribed then administered (RXE, RXA).

Report — Antimicrobial stewardship & resistance (AMR) Auto-updated
SARM (% S. aureus)
12,4 %
▼ 0,8 %
Consumption (DDD/1000 pt-days)
642
▼ 4 %
C. difficile infections (month)
7
▼ 2
Compliant prophylaxis
88 %
to improve

Resistance (%): organism × antibiotic

Antibiograms from the last 90 days
AmoxCiproCeftriPip-TazoMero
E. coli58241261
K. pneumoniae72311893
P. aeruginosa1002235148
S. aureus841812102
Enterococcus124090305
susceptible resistant

Antibiotic consumption — 12 months (DDD/1000 d)

Downward trend
JFMAMJJASOND
HL7 source — how this report is built OBX-3 · OBX-5 · Antibiogram — for each organism, the tested antibiotic (OBX-3) and its result — Susceptible, Intermediate or Resistant (OBX-5) — feed the organism × antibiotic resistance map. SPM-4 · Organism & specimen — the identified organism and specimen type (blood culture, urine…) come from the specimen segment (SPM). RXA · Antibiotic consumption — the curve is computed from antibiotics actually administered (RXA), converted to defined daily doses (DDD).

Clinical cohorts

Diabetes, renal failure, oncology… define a population by its diagnosis (DG1 segment) or by a lab result (OBX segment).

Health trends

Track indicator trends (HbA1c, glucose, blood pressure) over time.

Shareable reports

Export and share the report with the clinical team or management.

Clinicians & non-technical users

A true clinical HL7 BI, with no IT training

Free cross-analysis

Cross any clinical variable against another — e.g. critical results by the care unit where the patient is.

Meaningful comparisons

Compare, for instance, CLSC vs hospital admissions by diagnosis or exam.

Instant charts

Trends, distributions, comparisons: the chart is built live, with no wait.

Full autonomy

Explore clinical data directly at the source, with no detour through the IT team.

PACS · DICOM imaging

PACS analytics, organized by domain

Every C-STORE association feeds these dashboards, grouped by theme: volume, performance, rejects, conformance, dose, metadata and capacity.

Studies today
1 284
▲ 4,1 %
Stored images
4,7 M
▲ 22 k
C-STORE intake
99,4 %
▲ 0,2 %
Storage used
38,2 To
76 %
1 · PACS

Activity & volume

Sources — DICOM : Modality (0008,0060), StudyDate (0008,0020), StudyDescription (0008,1030) · HL7 : ORM^O01, ORU^R01 · FHIR : ImagingStudy, ServiceRequest

PACS load: day × hour

Studies received (C-STORE) — typical week
00–0404–0808–1212–1616–2020–24
Lun12402101809030
Mar10382251909528
Mer114224020010032
Jeu9392301959230
Ven134525020511035
Sam82095804018
Dim71670603014
low high

Top exams

StudyDescription (0008,1030) — 30 days
Brain CT
318
Chest X-ray
295
Abdo-pelvic CT
210
Lumbar MRI
165
Abdominal US
142
Mammography
96

Distribution by modality

100 studies
  • CT 31%
  • CR/DX 28%
  • US 16%
  • MR 14%
  • MG 6%
  • NM/XA 5%

Study volume — 12 months (thousands)

Upward trend
JFMAMJJASOND
2 · PACS

Performance & image access time

Sources — DICOM : C-STORE / C-MOVE, StudyTime (0008,0030), WADO-RS · HL7 : ORU^R01 (report timestamp) · FHIR : DiagnosticReport.issued

Median TAT
42 min
▼ 6 min
Studies within SLA
87 %
▲ 3 %
Median access time
0,6 s
▼ 0,1 s
Served from cache
92 %
▲ 2 %

SLA compliance

Report < 60 min
87 %
target 90% · 87% met

Retrieval by storage tier

Median retrieval time
Online (SSD)
0,4 s
Near-line (disk)
1,8 s
Archive (tape)
22 s

Access time by percentile

Time to first image (TTFI)
1 s
P50
2 s
P90
4 s
P95
18 s
P99

Modalities: volume × delay × size

One bubble per modality
0275582109Volume (studies / day)Median TAT (min)XAXACTCTMRMRNMNMUSUSMGMGCR/DXCR/DX
Bubble size ∝ images per study
3 · PACS

Capacity & storage

Sources — DICOM : PixelData (7FE0,0010), NumberOfFrames (0028,0008), Rows/Columns (0028,0010/0011), BitsAllocated (0028,0100)

Storage capacity

Primary PACS array
76 %
38.2 TB of 50 TB used

Cumulative storage — 12 months (TB)

Growth ~ 0.7 TB / month
JFMAMJJASOND
4 · PACS

Exam rejects by reason

Sources — DICOM : MPPS (status DISCONTINUED), Key Object Selection (quality reject), ImageComments (0020,4000) · HL7 : OBX (reason)

Reject rate
3,8 %
▼ 0,3 %
Rejects today
47
▼ 5
Top reason
Motion / blur
34 %
Repeats
39
added dose

Rejects by modality

Last 30 days
CR/DX
142
CT
38
MR
21
US
12
MG
9

Reject reasons

100 rejects
  • Motion / blur 34%
  • Positioning 26%
  • Artifact 17%
  • Exposure 13%
  • Wrong protocol 6%
  • Other 4%

Reject rate — 12 months (%)

Continuous improvement
JFMAMJJASOND
5 · PACS

Protocol conformance

Sources — DICOM : ProtocolName (0018,1030), KVP (0018,0060), Exposure (0018,1152), ConvolutionKernel (0018,1210) · HL7 : OBR-4 (exam code)

Overall conformance
96,3 %
▲ 1,4 %
Protocols audited
42
this month
Open deviations
7
to address
Approved parameters
98,1 %
▲ 0,6 %

Conformance by protocol (target 95%)

% of exams following the approved protocol
Chest X-ray
99,1 % / 95 %
Brain CT
98,2 % / 95 %
Mammography
97,8 % / 95 %
Chest CT
96,5 % / 95 %
Abdo-pelvic CT
94,1 % / 95 %
Lumbar MRI
91,4 % / 95 %
≥ target < target 95% target

Overall conformance

All protocols
96,3 %
target 95% · 42 protocols audited
6 · PACS

CT dose tracking (CTDIvol / DLP)

Sources — DICOM : Radiation Dose SR (RDSR), CTDIvol (0018,9345), DLP, MPPS · IHE REM

Mean CTDIvol
9,2 mGy
▼ 0,5
DLP median
520 mGy·cm
▼ 4 %
Exams > DRL
8,4 %
watch
Dose alerts (7 d)
23
reviews due

Median DLP vs DRL, by protocol

Diagnostic reference level (DRL) shown as the tick
Brain CT
850 mGy·cm / 1 000 mGy·cm
Chest CT
380 mGy·cm / 400 mGy·cm
Abdo-pelvic CT
745 mGy·cm / 800 mGy·cm
Cardiac CT
920 mGy·cm / 700 mGy·cm
Aortic CTA
1 150 mGy·cm / 1 200 mGy·cm
Lumbar spine CT
680 mGy·cm / 600 mGy·cm
≤ DRL > DRL NRD

Median DLP — 12 months (mGy·cm)

Continuous protocol optimization
JFMAMJJASOND
7 · PACS

DICOM metadata quality

Sources — DICOM : PatientID (0010,0020), AccessionNumber (0008,0050), ReferringPhysicianName (0008,0090) · HL7 : OBR-18 (Accession), PID (ADT) · FHIR : Patient.identifier

Overall completeness
96,8 %
▲ 1,1 %
Incomplete studies (today)
18
▼ 7
AccessionNumber missing
2,6 %
to fix
Most-missing tag
ReferringPhysician
15,7 %

Metadata completeness by tag

% of studies with a valid value · 99% threshold
PatientID (0010,0020)
99,8 % / 99 %
PatientName (0010,0010)
99,9 % / 99 %
Modality (0008,0060)
100 % / 99 %
StudyDate (0008,0020)
99,7 % / 99 %
AccessionNumber (0008,0050)
97,4 % / 99 %
StudyDescription (0008,1030)
92,1 % / 99 %
ReferringPhysician (0008,0090)
84,3 % / 99 %
≥ threshold < threshold 99% threshold

Conformance score

Key-tag metadata
96,8 %
weighted average · target 99%
Cross-standard analytics

Cross V2.x, V3, FHIR and DICOM on a single key

FHIR BI unifies the standards. Define a "join key" — a patient identifier (NAM, NIM, IPM, etc.) — and the tool automatically reconciles V2.x, V3 messages, FHIR resources and DICOM studies for the same patient.

The join key in action

Each standard exposes the patient identifier in a different field
HL7 V2PID-3 · PID-18 → NAM · NIM
HL7 V3id/@extension → NIM
FHIRPatient.identifier → NAM · IPM
DICOM(0010,0020) PatientID → IPM
Join key
NAM · NIM · IPM
Master patient index
Unified patient record
Admissions (V2)3
Lab results (V2/V3)27
FHIR observations12
DICOM studies4

Volume by unit, by standard

Messages joined on the patient key — 24 h
318
ER
215
Medicine
190
Surgery
92
ICU
114
CLSC
  • HL7 V2
  • HL7 V3
  • FHIR
  • DICOM

Activity by standard — 12 months

Reconciled monthly volume
JFMAMJJASOND
  • HL7 V2
  • HL7 V3
  • FHIR
  • DICOM

Critical results: unit × test type

OBX (V2/V3) crossed with PV1-3 — 30 days
K+TroponineGlucoseCreat.Hb
ER614935
ICU117496
Medicine42758
Surgery21324
CLSC10512
low high

Clinical example — patient journey

TREMBLAY, Jean · NAM TRJE 8203 1517 · joined on the key
08:12
V2ADT^A01 — Admission to ER
PV1-3 : ER
08:47
V2ORU^R01 — Lab result
OBX : Troponine 0,92 µg/L · critical
09:05
FHIRObservation — Abnormal ECG
Patient.identifier = NAM · IPM
09:40
DICOMStudy — Coronary angiography
1 series · 48 images
11:20
V3Document — Cardiology report
id/@extension = NIM
14:00
V2ADT^A02 — Transfer to ICU
PV1-3 : ICU
PACS example

One imaging request, from order to report

The accession number links the HL7 order, the DICOM images and the report — a single request record, across all standards.

Join key — Accession number

The carrier field in each standard
HL7 V2OBR-18 / OBR-3 → ACC
DICOM(0008,0050) AccessionNumber → ACC
FHIRServiceRequest.identifier → ACC
HL7 V3id/@extension → ACC
Key
ACC — A1234567

Request journey

Events joined on accession A1234567
10:02
V2ORM^O01 — Chest CT order
OBR-4 · OBR-18 = A1234567
10:15
V2Modality worklist (MWL)
Scheduled procedure
10:48
DICOMStudy received — Chest CT
(0008,0050) = A1234567 · 1 series · 220 images
11:30
FHIRImagingStudy
identifier = A1234567
14:10
V2ORU^R01 — Radiology report
OBR-18 = A1234567 · finalized
14:12
FHIRDiagnosticReport.issued
identifier = A1234567
Logistics example

Specimen traceability, from collection to result

The container identifier links the order, transport, lab receipt and analysis — a full chain of custody.

Join key — Container identifier

The carrier field in each standard
HL7 V2SPM-2 / SAC-3 → SID
FHIRSpecimen.accessionIdentifier → SID
DICOM(0040,0512) ContainerIdentifier → SID
HL7 V3specimen/id → SID
Key
SID — C-88231

Chain of custody

Events joined on container C-88231
07:20
V2ORM^O01 — Order — Blood culture
SPM-2 = C-88231
07:35
V2Specimen collected
SPM-17 · timestamp
08:05
FHIRSpecimen — In transport
accessionIdentifier = C-88231
08:40
V2Received at lab
SAC · container status
09:10
DICOMDigitized slide (pathology)
(0040,0512) = C-88231
15:00
V2ORU^R01 — Microbiology result
SID = C-88231 · validated
Real-time alerts

Inform clinicians within seconds

Connected to the HL7 stream, HL7AI detects the critical event as the message arrives and notifies the right clinician immediately — instead of waiting for them to open the chart. Every second saved counts.

LaboratoryCRITICAL

Potassium 6.8 mmol/L

TriggerORU^R01 · OBX-5 = 6,8 · OBX-8 = HH
DetectionPanic value above threshold, in real time
Attending (PV1-7) notified8 s
Benefitvs ~2 h by manual callback · arrhythmia risk
MicrobiologyURGENT

Positive blood culture — S. aureus

TriggerORU^R01 · OBX positive culture
DetectionBacteremia → sepsis protocol, in real time
Care team + stewardship15 s
Benefittargeted therapy sooner · lower mortality
PharmacyURGENT

Interaction — Warfarin + NSAID

TriggerRDE^O11 · new order
DetectionCross-check with active meds, in real time
Prescriber — before dispensingreal time
Benefitprevents a bleeding risk
PharmacyCRITICAL

Allergy — Penicillin

TriggerRDE^O11 · Amoxicillin
DetectionCross-check with patient AL1 allergy, in real time
Prescriber + pharmacist5 s
Benefitprevents a severe allergic reaction
Diagnostic — ImagingCRITICAL

Pulmonary embolism — CT angio

TriggerORU^R01 · report · critical-finding code
DetectionCritical imaging result, in real time
Ordering physician + ER12 s
Benefitimmediate anticoagulation · better outcome
LaboratoryCRITICAL

Troponin 0.92 µg/L

TriggerORU^R01 · OBX-8 = H
Cardiology + ER10 s
BenefitACS managed without delay
LaboratoryCRITICAL

INR 8.2 — anticoagulant overdose

TriggerORU^R01 · OBX-5 = 8,2
Attending9 s
Benefitvitamin K · bleeding risk avoided
LaboratoryCRITICAL

Glucose 1.8 mmol/L

TriggerORU^R01 · OBX-8 = LL
Unit nurse (PV1-3)6 s
Benefitimmediate glucose correction
LaboratoryURGENT

Creatinine 320 µmol/L — AKI

TriggerORU^R01 · rapid rise
Nephrology14 s
Benefitdose adjustment · nephrotoxics stopped
LaboratoryURGENT

Hemoglobin 62 g/L

TriggerORU^R01 · OBX-8 = LL
Care team12 s
Benefittransfusion considered sooner
LaboratoryCRITICAL

Sodium 118 mmol/L

TriggerORU^R01 · OBX-8 = LL
Attending8 s
Benefitcontrolled correction · neuro risk
LaboratoryCRITICAL

Lactate 5.2 mmol/L

TriggerORU^R01 · hypoperfusion
ICU7 s
Benefitsepsis/shock · early resuscitation
PharmacyCRITICAL

Overdose — dose > maximum

TriggerRDE^O11 · RXE-dose
Prescriberreal time
Benefitdose error blocked
PharmacyURGENT

Therapeutic duplication

TriggerRDE^O11 · 2 active PPIs
Pharmacist6 s
Benefitsimplified prescription
PharmacyURGENT

High-alert medication — Insulin

TriggerRDE^O11
Double-check required5 s
Benefitadministration error prevented
ImagingCRITICAL

Tension pneumothorax

TriggerORU^R01 · critical finding
Ordering physician + ER11 s
Benefitimmediate drainage
ImagingCRITICAL

Intracranial hemorrhage

TriggerORU^R01 · TDM brain
Neurosurgery + ER9 s
Benefitimmediate neuro management
ImagingFOLLOW-UP

Incidental finding to monitor

TriggerORU^R01 · follow-up recommendation
Callback loop to requestertracked
Benefitno finding lost to follow-up
AdmissionINFO

High fall risk

TriggerADT^A01 · history + age
Unit teamat admission
Benefitprevention from arrival
MonitoringURGENT

Early-warning score NEWS2 ≥ 7

TriggerORU^R01 · vitals (OBX)
Rapid response team20 s
Benefitdeterioration caught before arrest

Median notification delay

9 s
from message arrival to clinician alert
HL7 technical support

Diagnose and resolve integration incidents faster

For integration teams: every rejected message is analyzed automatically — faulty segment, acknowledgement code (AR/AE), cause and suggested fix — instead of reading logs by hand.

Rejected inbound message
MSH|^~\&|SIU|CLINIQUE|DPI|CHUM|20260630103000||ADT^A01|MSG7781|P|2.5.1 EVN|A01|20260630103000 PID|1||‹ PID-3 / NAM absent ›||GAGNON^MARIE||19731105|F PV1|1|I|MED^4^B|||||8812^ROY^PAUL — — — — — — — — — — — — — — — — MSA|AR|MSG7781|Missing required field PID-3

Automatic diagnosis

Reject analysis on arrival
StatusREJECTED · AR
CauseMissing patient identifier (PID-3 / NAM)
FieldPID-3
ACKMSA-1 = AR · MSA-3 = « Missing required field PID-3 »
SourceSIU / CLINIQUE
FixEnrich via the patient index (MPI) or return to source system

Common support cases

Symptom, detected cause, HL7 field and action
Symptom Detected cause Segment / field Action
NAK (AR) on admissionsMissing patient identifierPID-3Enrich via MPI / return to source
Duplicate messagesIdentical control IDMSH-10Automatic de-duplication
Sequence gapNon-contiguous sequence numberMSH-13Missing-message detection + replay
Corrupted accentsISO vs UTF-8 encodingMSH-18UTF-8 normalization
Unrouted resultUnmapped exam codeOBR-4LOINC mapping table
Unrecognized unitNon-standard unitOBX-6UCUM normalization
Interface downMLLP connection lost—Reconnect + queue replay
End-to-end flows

Trace every message, from source to application

The Sankey diagram reveals the real data journey: which source systems feed which standards, and which clinical applications they are finally routed to. Each flow's thickness is proportional to its volume.

Source systems → standards → clinical applications

Daily volume · thickness ∝ message count
SIL → V2: 220SIL → FHIR: 60RIS → DICOM: 140RIS → V2: 40ADT → V2: 180ADT → V3: 50PHA → V2: 70PHA → FHIR: 40V2 → DSQ: 270V2 → DOS: 190V2 → BI: 50V3 → DSQ: 50FHIR → BI: 60FHIR → DOS: 40DICOM → PACS: 120DICOM → BI: 20SILRISADTPharmacyHL7 V2HL7 V3FHIRDICOMDSQClinical recordPACSFHIR BI
HL7 technical support

An accelerator for analyzing rejections and errors

For teams handling HL7 rejections, errors and inconsistencies.

  • Clear visualization of faulty or inconsistent variables
  • Correlation across multiple HL7 fields (e.g. PV1-2 vs PV1-3, OBR-4 vs OBX-3)
  • Statistical analysis of rejection causes by facility, source system or period
  • Quickly understand the "why" of a rejection, beyond the raw technical error message

Rejection causes by type

Last 7 days
64
Bad PID
41
No PV1
28
Bad OBR-4
19
Duplicate
Real time

Connected live to the National or Organizational Integration Agent's HL7 flow

Connected in real time to the integration agent's HL7 flow, the tool enables:

Immediate detection

Analysis of incoming messages as they arrive.

Near real-time tracking

Continuous tracking of rejections and critical alerts.

Proactive intervention

Act on at-risk flows before incidents occur.

Monitoring

Real-time integration monitoring

Monitor interface health: volumes, acknowledgements (ACK/NAK) and per-flow latency.

Messages / day
52 410
▲ 3,4 %
ACK rate
99,6 %
▲ 0,2 %
Median latency
36 ms
▼ 5 ms
Active listeners
14
all UP

ACK rate — 12 months (%)

Positive acknowledgements
JFMAMJJASOND

Median latency by interface

Receive → ACK
ADT
24 ms
Pharmacy
33 ms
SIL
38 ms
Lab
41 ms
RIS
52 ms
Incidents

Incidents & resolution time

Categorize interface incidents and track the mean time to resolution (MTTR).

Open incidents
6
▼ 3
MTTR
3,5 h
▼ 0,4 h
SLA met
94 %
▲ 2 %
Recurring
2
to fix

Incidents by category (30 d)

Technical causes
12
Connection
9
Format/parse
7
Mapping
5
Timeout
3
Other

Mean time to resolution (MTTR) — 12 months (h)

Continuous improvement
JFMAMJJASOND
Logistique

Flows, transport, stock and traceability

Reconstructed from HL7 (ADT, ORM, SPM) and DICOM messages — patient transport, specimens, reagents, sterilization, queues.

Transports / day
486
▲ 2,6 %
Specimens in transit
312
18 late
Bed occupancy
87 %
target 85%
Stock-outs (30 d)
7
▼ 3

Patient transports — 12 months

Portering (per day)
JFMAMJJASOND

Source HL7 — ADT^A02 (transferts) + messages de brancardage. Each patient transfer fires an A02; events are counted per day to measure portering load.

Avg transport delay by unit

Request → pickup
ER
22 min
Imaging
18 min
OR
15 min
Medicine
12 min
Rehab
9 min

Source HL7 — ORM^O01 (demande) → ADT (prise en charge). The delay is the gap between the transport-request timestamp and arrival in the destination unit.

Bed occupancy by unit (%)

Bed management (PV1-3)
110 %
ER
96 %
Geriatrics
92 %
Medicine
84 %
Surgery
72 %
Obstetrics

Source HL7 — PV1-3 (localisation), ADT^A01/A02/A03. Admissions/discharges/transfers keep bed occupancy per unit current: occupied ÷ available beds.

Collection → lab receipt (min)

By collection site
North clinic
52 min
Home
44 min
Outpatient
28 min
Inpatient
14 min

Source HL7 — SPM-17 (date/heure de prélèvement) → OBR-14 (réception). Collection → lab-receipt time, computed per requesting unit.

ADT flow — 12 months

Admissions / transfers / discharges
JFMAMJJASOND
  • Admissions
  • Transfers
  • Discharges

Source HL7 — ADT^A01 (admission), A02 (transfert), A03 (sortie). The three plotted volumes come straight from the event type carried in MSH-9 / EVN-1.

Specimens in transit — status

312 spec.
  • In transport 19%
  • Received 38%
  • In analysis 37%
  • Late 6%

Source HL7 — SPM-11 (rôle) + SAC (statut contenant). Each specimen's status (collected, in transit, received, rejected) is broken down into proportions.

Reagent stock vs threshold (%)

% of reorder level
Hematology
140 % / 100 %
Chemistry
115 % / 100 %
Microbiology
92 % / 100 %
Rapid kits
64 % / 100 %
≥ target < target Target

Source HL7 — OMS^O05 (réquisition de stock) + RXA (consommation). Remaining reagent stock compared to the reorder threshold to anticipate stock-outs.

Stock-outs by category (30 d)

4
Reagents
2
Consumables
1
Pharmacy

Source HL7 — OMS^O05 (statut « en rupture »). Count of stock-outs detected per supply category over the period.

Sterilization cycles / day

Typical week
42
Lun
45
Mar
44
Mer
46
Jeu
48
Ven
20
Sam
14
Dim

Source HL7 — ORM^O01 (ordres de retraitement d'instruments). Sterilization cycles launched per day, from the CSSD reprocessing orders.

ER wait: day × hour (min)

Median time
00–0404–0808–1212–1616–2020–24
Lun456018021015080
Mar425817520514578
Mer405518521515582
Jeu446217820814880
Ven486519522516590
Sam30401201409555
Dim28361001208048
low high

Source HL7 — ADT^A04 (admission urgence) + horodatages PV1-44/PV1-45. Time from triage to care, crossed by weekday × time slot.

No-show rate — 12 months (%)

Missed appointments
JFMAMJJASOND

Source HL7 — SIU^S12 (rendez-vous) vs ADT^A04 (présence). Scheduled appointments with no matching admission = no-show rate.

Pneumatic transport — sends/day

By service
320
Lab
180
Pharmacy
140
ER
90
OR

Source HL7 — OML^O21 (échantillon lié à l'ordre). Pneumatic-tube send volumes, correlated to lab orders per service.

Equipment utilization (RTLS)

100 devices
  • In use 64%
  • Available 22%
  • Maintenance 9%
  • Missing 5%

Source — RTLS feed (real-time location), outside HL7. Mobile-equipment utilization status: in use, available, in maintenance, missing.

Room turnover time (min)

Discharge → ready
ICU
78 min
Surgery
62 min
Medicine
54 min
Obstetrics
40 min

Source HL7 — ADT^A03 (sortie) → ADT^A01 (admission suivante). Time from a patient's discharge to the room being ready again (cleaning included).

Logistics message throughput — 12 months (k/day)

Stable queue
JFMAMJJASOND

Source HL7 — MSH (tous messages logistiques). Integration-engine throughput: ADT/ORM/OML messages processed per day, a queue-load indicator.

Finance

Revenue, costs, billing and coding

Clinical activity (ADT, procedures, DG1, results) becomes financial indicators: revenue by service, cost per stay, billing rejections, coding completeness.

Billing / month
4,2 M$
▲ 3,1 %
Billing rejection
6,4 %
▼ 0,7 %
Days sales outstanding
42 j
▼ 3 j
Overall margin
11,8 %
▲ 0,5 %

Revenue by service (k$)

Current month
920
Surgery
740
Imaging
610
Medicine
480
ER
350
Lab

Source HL7 — DFT^P03 (transaction financière), FT1-16 (service), PV1-10. Each billed procedure produces a DFT; amounts are aggregated by requesting service.

Monthly billing ($M) — 12 months

Billed procedures
JFMAMJJASOND

Source HL7 — DFT^P03 / FT1-6 (montant). Monthly sum of FT1 amounts across all financial transactions.

Avg cost per stay (DRG) — $

Top 5 DRGs
CABG
28 500 $
Hip replacement
18 200 $
AVC
12 400 $
Pneumonia
7 600 $
Delivery
4 300 $

Source HL7 — DRG (groupe diagnostic), DG1, PV1 (durée). Average cost per stay grouped by APR-DRG, from diagnosis and length of stay.

Billing rejection rate — 12 months (%)

Continuous improvement
JFMAMJJASOND

Source HL7 — ACK/MSA-1 (AR/AE) sur les DFT. Share of financial transactions rejected by the payer, per month.

Budget vs actual by line (k$)

Current quarter
1 940
Staff
650
Drugs
430
Supplies
420
Imaging
315
Lab
  • Actual
  • Budget left

Source HL7 — DFT^P03 (réel) vs budget planifié (ERP). Actuals come from FT1 transactions; stacked against budget per expense category.

Payer mix

100 claims
  • RAMQ 68%
  • Private / insurer 18%
  • Other province 8%
  • Self-pay 6%

Source HL7 — IN1-4 (nom du payeur), IN1-2 (régime). Claim distribution by payer: RAMQ, private, CNESST, SAAQ, out-of-province.

Imaging cost by modality ($/study)

420 $
IRM
260 $
TDM
95 $
US
45 $
X-ray
70 $
Mammo

Source — HL7 DFT + DICOM Modality (0008,0060). Average cost per study by modality (CT, MRI, XR, US), joining billing with the DICOM header.

Drug cost by class (k$)

Current month
Oncology
820 k$
Anti-infectives
340 k$
Immunology
290 k$
Cardio
180 k$
Others
150 k$

Source HL7 — RXA (administration), RXE-2 (code produit). Cost of administered drugs grouped by therapeutic class.

Margin by service (%)

22 %
Imaging
18 %
Surgery
14 %
Lab
6 %
Medicine
3 %
ER

Source HL7 — DFT (revenus) − coûts affectés. Margin per service = billed revenue minus allocated direct costs, as a percentage.

Revenue leakage (unbilled, k$)

By service — 30 d
58 k$
ER
42 k$
Imaging
31 k$
Consults
19 k$
Lab

Source HL7 — ORU^R01/OBR (acte réalisé) sans DFT associé. Revenue leakage: procedures documented by a result but with no matching billing transaction.

Coding completeness (%) vs target

DG1 / procedures
Diagnoses (DG1)
94 % / 95 %
Procedures
91 % / 95 %
Comorbidities
86 % / 95 %
Complications
82 % / 95 %
≥ target < target Target

Source HL7 — DG1 (diagnostics), PR1 (actes). Coding completeness: DG1/PR1 fields present vs expected per stay, against target.

Days sales outstanding — 12 months (d)

Downward trend
JFMAMJJASOND

Source HL7 — dates DFT (émission) → paiement. Average days-sales-outstanding: days between billing and payment.

Cost per bed-day — 12 months ($)

JFMAMJJASOND

Source HL7 — PV1-44/45 (durée de séjour) + coûts. Total cost divided by occupied bed-days, derived from admission/discharge dates.

Lab spend by test family (k$)

Chemistry
210 k$
Hematology
160 k$
Microbiology
120 k$
Immunology
95 k$
Genetics
70 k$

Source HL7 — OBR-4 (test demandé), OBX. Lab spend grouped by test family (chemistry, hematology, microbiology, etc.).

Revenue by facility (k$)

5 200
CHU
3 400
Regional
1 800
Clinic A
900
CLSC

Source HL7 — MSH-4 (établissement émetteur) + DFT. Revenue aggregated by source facility, identified by the sending field in the header.

Administration

Administrative & operational steering

Management indicators are computed directly from HL7 admit/transfer/discharge (ADT) flows, FHIR resources and DICOM volumes — with no manual entry.

Admissions / day
142
▲ 3 %
Occupancy rate
87 %
target 85%
Average length of stay
5,8 d
▼ 0,3
30-day readmission
11,4 %
▼ 0,6 %

Indicator scorecard

Value, target, trend and HL7/FHIR/DICOM source per indicator
Indicator Value Target Trend Source
Admissions / day142—▲ 3 %HL7 ADT^A01 PV1 · FHIR Encounter
Average length of stay5,8 d≤ 6 d▼ 0,3HL7 PV1 · FHIR Encounter.period
Bed occupancy rate87 %85 %▲ 2 %HL7 PV1-3 · FHIR Location
30-day readmissions11,4 %≤ 12 %▼ 0,6 %HL7 ADT^A01 ADT^A03
Median ER wait3 h 12≤ 4 h▼ 14 minHL7 ADT PV1 (timestamps)
ALC / alt-level patients38—▲ 4HL7 PV1-2 (patient class)
Cancelled surgeries4,1 %≤ 5 %▼ 0,3 %HL7 SIU^S12
Discharges before noon32 %≥ 35 %▲ 3 %HL7 ADT^A03
In-hospital mortality1,8 %—▼ 0,1 %HL7 ADT^A03 · FHIR Encounter.hospitalization
Imaging exams / day1 284—▲ 4,1 %DICOM C-STORE Modality (0008,0060)
Administration

Patient activity & flow

Sources — HL7 : ADT^A01, ADT^A02, ADT^A03, PV1 · FHIR : Encounter, Patient

Admissions vs discharges — 12 months

ADT^A01 / ADT^A03 events
JFMAMJJASOND
  • Admissions
  • Discharges

Average length of stay by service (d)

Computed from PV1 / Encounter.period
Geriatrics
9,8 j
Medicine
6,4 j
Surgery
5,1 j
ICU
4,5 j
Pediatrics
3,2 j
Obstetrics
2,6 j
Administration

Occupancy & capacity

Sources — HL7 : PV1-3 (unit/bed), ADT^A02 · FHIR : Location, Encounter.location

Occupancy rate by unit (%)

PV1-3 (location) · 85% target
110 %
ER
96 %
Geriatrics
92 %
Medicine
88 %
ICU
84 %
Surgery
72 %
Obstetrics

Overall occupancy

All beds
87 %
target 85% · 87% reached
Administration

Quality & performance

Sources — HL7 : ADT (timestamps), PV1 · FHIR : Encounter.hospitalization

Key indicators vs target (lower is better)

Readmission %, ER wait h, cancellations %, ALOS d
30-day readmission (%)
11,4 / 12
ER wait (h)
3,2 / 4
Cancelled surgeries (%)
4,1 / 5
Avg length of stay (d)
5,8 / 6
≤ target > target Target

Median ER wait — 12 months (h)

Triage → provider
JFMAMJJASOND
In summary

From opaque technical flow to clinical knowledge

FHIR BI turns HL7 into a source of clinical and operational knowledge.

Faster incident resolution

Reduces incident resolution times.

Better data quality

Improves the quality of exchanged data.

Support for clinical teams

Supports clinicians in modeling and analysis.

Real-time analysis

Real-time analysis when connected to the National/Organizational Integration Agent.

Health-status reports

Models patients' health status from HL7 data.

Stronger HL7 maturity

Strengthens the organization's HL7 autonomy and maturity.

See your HL7 messages differently

Create a free account and explore FHIR BI today.

Create account HL7AI Integration Agent
An unhandled error has occurred. Reload 🗙

Rejoining the server...

Rejoin failed... trying again in seconds.

Failed to rejoin.
Please retry or reload the page.

The session has been paused by the server.

Failed to resume the session.
Please retry or reload the page.