Continuous Learning

Clinical confidence, with models that improve over time.

Our patented Continuous Learning framework aggregates real-world clinical data from across our network, using it to continuously optimize pharmacokinetic models, keeping predictions accurate as patient populations and clinical practice evolve.
  1. Home
  2. Platform
  3. InsightRX Continuous Learning

Static models stagnate, InsightRX models improve.

Most precision dosing platforms deliver a fixed model, trained once on historical data and deployed indefinitely. As prescribing patterns shift, patient populations change, and clinical practice evolves, those models lose accuracy without anyone noticing.

InsightRX's architecture is fundamentally different. Patient data from dosing encounters flows continuously from participating institutions, where it is aggregated and used to retrain pharmacokinetic models using the same rigorous methods applied in population PK research.

The optimization cycle: four phases, running continuously as data flows through the network.

Most popPK models are based on small datasets from just one or two institutions and don't necessarily generalize well to unseen patient populations. By developing models based on very large datasets at hundreds of hospitals just like yours, we can better estimate what contributes to PK variability between patients and sites.

The bedside decision support your team relies on improves with every cycle, automatically, with no additional burden on clinical or IT staff.

01 Data Ingestion 02 Aggregation 03 Model Optimization 04 Deployment

Phase 01

Data Ingestion

Pharmacological data from dosing encounters (drug concentrations, lab values, biomarkers, and patient covariates) flows continuously from participating institutions. Each record is tagged with a drug identifier and model identifier. Data is de-identified before leaving the originating institution.

The only patented model optimization process in clinical dosing software

InsightRX is the first and only healthcare technology company to hold a US patent for pharmacology model optimization based on distributed data acquisition. The patent protects the core methodology by which patient data is aggregated across institutions and used to continuously retrain and validate pharmacokinetic models.

US Patent 11,075,010 B1

Pharmacology model optimization based on distributed data acquisition. Granted July 27, 2021.

View on Google Patents

The model science that makes the optimization cycle credible

Continuous Learning continuously refines InsightRX's pharmacokinetic models. These posts explain how those models are built, selected, and validated before improvement begins, documenting the scientific foundation the optimization cycle depends on.

Case Study:

Holyoke Medical Center partnered with InsightRX to enhance vancomycin dosing practice through the Apollo clinical analytics application, fostering a culture of continuous improvement and improving target attainment by nearly 30%.
29.7% Increase in target attainment from baseline
Read the case study

The system dropped our rate of kidney injury for patients on certain antibiotics. Our estimate is that the system ultimately saves us hundreds of thousands of dollars a year, and that may change more as we use it more.

- Pharmacy Manager

KLAS Spotlight Report

Maximize therapeutic success with precision medicine.

Contact Us