Post-treatment response can add information about a later patient state.
PUBLIC SCIENTIFIC RECORD
Claims shown with their evidence boundary.
HISTAMOS publishes research questions, completed stress tests and proposed validation designs while keeping clinical authority, uncertainty and proprietary methods separate.
The available evidence does not establish a superior patient-level treatment policy.
Independent prospective validation and appropriate oversight remain required.
PUBLICATION 01
Treatment-response–informed precision medicine
A public proof of work showing both the value of response information and the point at which the evidence stops supporting treatment selection.
Human evidence, validation gap and economic relevance
The analysis distinguishes prognosis from treatment choice. It tests a plausible response-guided strategy against randomized human evidence and keeps the claim bounded when it does not outperform a simpler policy.
- Patient state can change after treatment begins.
- An informative response is not automatically a treatment selector.
- Prospective independent validation is required before patient-facing use.
Evidence statusSupported · bounded
Treatment allocationNot validated
Clinical useNone
Next requirementProspective validation
PUBLICATION 02
Lorundrostat renal–electrolyte rescue framework
A decision-time state and nested randomized rescue framework for preserving antihypertensive benefit while testing renal and electrolyte management strategies.
What the evidence supports now
Fixed lorundrostat 50 mg is the strongest replicated population-level exposure anchor in the supplied manuscript. The framework does not treat early creatinine change as proof of injury or as proof of benign adaptation. It first separates measurement effects from orthogonally supported filtration change.
- Lower exposure at 25 mg is an exact-product rescue candidate, not a validated rescue.
- CKD currently supports risk stratification, not treatment selection.
- The decisive experiment is nested randomization within a qualified decision-time state.
Article typeHypothesis and theory
Population anchorFixed 50 mg
Rescue candidate25 mg · unvalidated
Primary designNested randomization
CONTROLLED ANALYTICAL SOURCES
Validation data kept outside the public site
The public record exposes conclusions and boundaries. It does not publish row-level health data or internal model construction.
Treatment-response model stress test
7 / 10 configurations qualified
The strongest supplied clean pre-only configuration reached held-out RMSE 1.4779 and R² 0.2957, an 8.3% RMSE reduction versus the pre-event baseline. This is methodological evidence only and carries no biological or clinical authority.
RRT discontinuation dataset
124 × 84 records × variables
The workbook contains one KeGFR sheet with baseline, ICU, dialysis, kinetic renal-function and discontinuation-outcome variables. The row-level dataset is not served by the public website.
Disclosure boundary. The public site does not expose the HISTAMOS engine, rules, thresholds, feature engineering, ranking logic, source code, private hypotheses or patient-level source records.
Independent validation is the next useful step.
HISTAMOS is seeking qualified academic, clinical, payer and research partners who can evaluate these questions using governed patient-level data and prospective designs.