Mosaic Health Solutions, Inc.
MosaicRx provides a patient-specific medication risk assessment tool for value-based care teams managing complex older adults. The platform conditions drug burden scores on the individual's current physiology—such as stage four kidney disease—and surfaces the top three risks at the prescribing moment with a transparent evidence trail. A deterministic rules engine generates all clinical claims, while a language model only renders explanations in readable form.
- Healthcare Technology
- Digital Health
- Data & Analytics
- Software Only
Funding
Founders
Founder details are not available yet.
Product
Problem
Standard medication burden scoring treats each drug as carrying a fixed weight, ignoring how a patient's current physiology—such as kidney function—changes the actual risk of a given medication list. This leads to prescribing cascades where side effects are misread as new conditions, new drugs are added to treat them, and the compounding burden results in falls, delirium, and emergency department visits that are difficult to trace back to any single error.
Solution
MosaicRx provides a patient-specific medication risk assessment tool that conditions drug burden scores—such as anticholinergic load—on the individual's current physiology, adjusting both the score and the threshold it is measured against. At the prescribing moment, the system highlights the top three risks for that specific patient, each with a transparent evidence trail showing contributing medications, timing, and the questions worth reviewing. The platform is designed as a prioritization layer around clinical judgment, not another alert: it surfaces risks, shows what it ruled out and why, and keeps the clinician in control—nothing auto-prescribes or auto-stops. Every clinician response is captured to improve the engine's accuracy over time.
Target Audience
Primary users are value-based care teams—including physicians, pharmacists, and nurse practitioners—managing complex older adults who are at elevated risk for medication-related adverse events and prescribing cascades.
Features
- Phenotype-specific risk scoring conditioned on the patient's longitudinal trajectory and current physiology, not static drug-drug tables
- Glass Box explainability exposing the evidence trail for every determination, including contributing medications, timing, and review questions
- Deterministic rules engine generates all clinical claims with a complete decision log; the language model only renders explanations and cannot generate, alter, or extend clinical claims
- Rule-out reporting that shows hypotheses considered, pharmacologic and temporal reasoning behind rejections, and agents examined and excluded with grounds
- Clinician feedback capture across four states—acted on, rationale documented, not relevant, or not actionable today—to identify where the engine is wrong
- Prioritized top-three risk presentation at the prescribing moment rather than a high-volume alert stream