Measurement integrity for clinical data
Know which patients move when your instruments change.
NakedSignal measures what an analyser, site or device change does to results at your clinical decision limits, and gives you a plan to manage the switch.
Runs on your own machine. Patient rows never leave.
10,579
paired patient results analysed
34
method comparisons
12
published studies
14
study sites
Why it matters
Every instrument change moves patients. The question is which ones.
Validation software reports the average bias. It does not tell you who now sits on the other side of a treatment threshold.
- In 23 of 29 published comparisons, the change moved more patients across a limit than repeat testing alone would.
- In the CD4 sample, 535 results changed category against about 238 expected.
- A re-test band removed 66% of that excess while repeating 15% of results.
Results moved across a decision limit
Public dataEach dot is one published method comparison. Above the dashed line, the change moved more patients than repeat testing alone would.
How it works
From the data you already have to a go-live plan.
Built on ISO 15189:2022 comparability of results and CLSI EP09 method comparison.
Export your paired results
The paired results from your method comparison or parallel testing. No new samples, no new study.
Measure at your decision limits
Bias where your clinicians act, and which patients cross each limit beyond what repeat testing explains.
Manage the switch
A re-test band for the transition and a one-page note for requesting clinicians, ready for go-live.
Runs inside your laboratory
The analysis runs on your own computer. Patient rows never leave; only aggregate results do, and you sign off what is released.
The change report
Everything you need to sign off a change.
Comparison line
How the new system relates to the old across your whole range, with its interval.
Bias at your limits
The difference at the concentrations where your clinicians act, not an average.
Patients across each limit
Who changed category, in each direction, set against repeat-testing noise.
Re-test band
A narrow range around each limit where results are repeated during the transition.
Clinician note
One page for requesting clinicians: what changed and what it means for their patients.
Network view
Across sites: which ones report interchangeably and which move patients.
Solutions
One method, wherever measurement changes.
Laboratories and networks
Replace an analyser without surprising your clinicians, and keep many sites reporting as one.
- Analyser and platform changes
- Reagent lot changes
- Multi-site comparability
Diagnostics makers and EQA
Show customers what a conversion does at their limits, and spot undeclared changes in scheme data.
- Conversion impact for customers
- Lot and recalibration checks
- Scheme-wide change detection
Clinical trials and medical AI
Separate site, lab and device effects from treatment effects, and test models on data they have not seen.
- Central versus local lab effects
- Device and site shifts
- Held-out model evaluation
Resources
Research and guides
Guide
Bias at clinical decision limits
Why the average bias hides the patients who matter, and how to measure at the limit.
ReadGuide
Jaffe versus enzymatic creatinine
What a creatinine method change does to eGFR staging, worked through on published data.
ReadResearch
MedEval-1
A versioned evaluation standard for medical AI, with a sealed, rotating held-out track.
ReadPlanning an analyser change?
Tell us what is changing and when. We will show you what the report would find.