From Diagnosis
to Treatment

The complete analytics lifecycle for hospital material flow

Every hospital is a patient.
Its data is the specimen.

We built a diagnostic laboratory — not a consulting practice.
The lab processes specimens. It produces findings. It renders verdicts.
What happens next is a clinical decision.

Five layers. Four built. One ahead.

Treatment
what do we do about it?
Diagnoses
root cause identification
Hypotheses
business questions with verdicts
Assessments
entity health scores
Probes
detect symptoms in data

From detection to decision

1
What?
Probes detect anomalies. A billing gap. A missing implant. A duplicate charge.
2
How bad?
Assessments score entity health. CHF at risk. Severity. Affected scope.
3
Is it real?
Hypotheses weigh evidence. Confirmed, plausible, or not observed.
4
Why?
Diagnoses trace to root cause. Process, system, or behaviour.
5
What now?
Treatment recommends action. Investigation. Resolution. Verification.

A confirmed hypothesis
is not enough.

The CFO doesn't want to know what is happening.
They want to know why.

“Revenue leakage confirmed” is a finding.
“The SAP-OPALE interface drops 3% of billing events
on weekend transfers” is a diagnosis.

From signal to root cause

Confirmed
Hypothesis
→
Conditions
Check
→
Confidence
Compute
→
Root
Cause

The system evaluates structural conditions around each confirmed hypothesis —
temporal patterns, entity correlations, severity distributions —
to narrow the search space from “something is wrong” to “here is why.”

Six categories. One language.

⚙
Process Failure
A step in the workflow was skipped, delayed, or executed incorrectly.
⌨
System Failure
An interface dropped data, a mapping was stale, or a sync failed silently.
⚠
Data Quality
Missing fields, orphan references, type mismatches in the source data.
✎
Behavioural
Staff bypass, selective scanning, undocumented workarounds on the floor.
◆
Structural
Cost centre misalignment, organisational mapping gaps, taxonomy drift.
☁
External
Regulatory changes, supplier catalogue updates, seasonal demand shifts.

From finding to root cause

Finding
Revenue leakage: CHF 142K unbilled usage in Q3
↓
Hypothesis
Confirmed: systematic unbilled consumption in orthopaedics
↓
Diagnosis
Root cause: billing workflow gap — weekend transfers not triggering invoice events. Confidence: 85%.
↓
Recommendation
Check ERP interface logs for Saturday/Sunday batch failures. Review billing trigger configuration.
The diagnostic lab identifies the disease.
It doesn't perform the surgery.

This is not a limitation. It is a design decision.
The lab must stay pure to stay trusted.

What the system cannot know

?
Who is responsible?
The data shows where the gap is. It does not show who owns the process or who can fix it.
↺
What has been tried?
Past remediation attempts, workarounds, and institutional memory live outside the data.
✓
Is it intentional?
Some anomalies are accepted trade-offs. The system flags them. Humans decide if they matter.
⚖
What does the fix cost?
The cost of a diagnosis is quantified. The cost of treating it requires operational context.

Floor knowledge meets
data intelligence

The System
Detects patterns. Quantifies risk. Identifies root causes. Suggests where to look.
↔
The Human
Knows the context. Weighs trade-offs. Assigns ownership. Makes the decision.

The system suggests. Humans decide.
This is where analytics becomes action.

Three capabilities to close the loop

☷
Investigation Templates
Structured checklists generated from the diagnosis. What to verify. Where to look. What data to collect from the floor.
☑
Resolution Tracking
Record what was decided, what was changed, and when. Link resolutions back to the original diagnosis for audit trail.
★
Cross-Tenant Knowledge
When hospital A resolves a diagnosis, the resolution pattern becomes available to hospitals B, C, and D facing the same root cause.

The cycle that makes the system smarter

Detect
→
Assess
→
Hypothesise
→
Diagnose
→
Treat
Human Decision
→
Verify

After treatment, the next data load re-runs the probes.
Did the finding disappear? Did the money at risk decrease?
The system verifies its own prescriptions.

When hospital A resolves a diagnosis,
hospital B learns from it.
The fleet gets smarter.

Same root cause. Same probe. Different hospital.
The resolution that worked for one becomes a recommendation for all.
Every tenant makes the platform stronger.

The diagnostic lab stays pure.
The treatment ward is where humans work.

Treatment
investigation, resolution, knowledge
Diagnoses
root cause identification
Hypotheses
business questions with verdicts
Assessments
entity health scores
Probes
detect symptoms in data

From specimen to treatment. The complete lifecycle.

nuMetrix