Evidence review
The evidence story is built around validation discipline, operational auditability, and honest status labeling while studies are still in progress.
Methodology02
Retrospective validation design, stated plainly.
Study type
Retrospective cohort validation for perioperative risk stratification.
Partner status
Teaching-hospital research collaboration structure in progress; named institution not yet published.
Sample target
10,000+ surgical encounters across varied case types and perioperative risk profiles.
Primary endpoint
CPRI discrimination against 30-day adverse event rate using AUC-ROC review.
Secondary endpoints
Sensitivity and specificity at high-risk thresholds plus recommendation-concordance review.
Clinical posture
Score outputs remain clinician-reviewable with override and visible driver transparency.
Preliminary Results03
Status shared honestly while validation is still underway.
AUC-ROC
Pending
Study in progress
Sensitivity
Pending
High-risk threshold under review
Specificity
Pending
Threshold calibration in progress
Public-facing performance claims stay intentionally conservative until the underlying cohort review, statistical analysis, and internal documentation are complete enough to stand up to external scrutiny.
Advisory Coverage04
Clinical oversight areas already defined, even where public names are not.
Chief of Anesthesiology
Reviews whether score outputs support perioperative decisions without introducing workflow friction.
Named advisor to be announced after public permission.Clinical Informaticist
Evaluates data provenance, chart mapping, and how score transparency appears inside clinical workflow.
Institution detail withheld until collaboration is public.Perioperative Medicine Lead
Assesses endpoint relevance, complications framing, and multidisciplinary handoff value.
Coverage area confirmed; public attribution pending.Pharmacogenomics Advisor
Reviews drug-gene integration logic and how genomic enrichment should be presented to clinicians.
Partner identity reserved until joint review is complete.Technical Documents05
Whitepapers first, publications when the evidence is ready.
Peer-reviewed publication is a goal, not a marketing shortcut. Until named studies and publication status can be stated responsibly, the current technical record lives in whitepapers and review materials.
Browse clinical whitepapersValidation Standards06
How Checksalus keeps the evidence bar credible.
- Retrospective validation before broad claims about predictive performance.
- Clear separation between pending metrics, modeled assumptions, and established outputs.
- Evidence basis and score drivers exposed so clinicians can interrogate why risk is elevated.
- Clinician override preserved so the system remains decision support, not opaque automation.
- Technical documents updated as validation milestones mature instead of pre-claiming outcomes.
Principle
AI in medicine has to be auditable before it is impressive.
Checksalus treats validation, explainability, and clinician override as product requirements rather than launch-stage polish.
Research Partnership
Interested in a clinical validation or study-design conversation?
Bring your endpoint priorities, review requirements, and collaboration goals. We’ll keep the conversation scoped to what is real today and what is still being developed.
Request Research ConversationGovernance
Evidence and auditability move together.
Validation review is paired with auditable deployment practices so informatics, compliance, and clinical leaders can examine not only what the model predicts, but how the surrounding workflow is governed.
Score-driver visibilityClinician overrideAudit logsControlled rolloutPending claims labeled clearly