Sixty percent of high cost members in any given year are new high cost members. Yet, current analytics approaches tend to identify existing persistent high cost members; managing care and costs, after the problem appears.
Cardinal Analytx Solutions identifies next year’s new high cost members before a high cost event occurs, and works with clinician teams to precisely target interventions that improve the quality of care and contain costs.
We bring together distinguished innovators and leaders across clinical health, health management and data science, utilizing new technology and methodologies to dramatically reduce health care costs and improve clinical outcomes.
Associate Professor of Medicine, and Biomedical Data Science, Stanford University
Assistant Director of the Center for Biomedical Informatics Research, Stanford University
Elected Fellow, American College of Medical Informatics and American Society for Clinical InvestigationREAD BIO
Professor of Medicine, Stanford University
Director Clinical Excellence Research Center, Stanford University
Member, National Academy of MedicineREAD BIO
Blue Shield of California
An opportunity for you to make a substantial impact on one of the most pressing human and economic challenges of our time.
Join a diverse group of smart, passionate and driven individuals to build an amazing company. We are hiring across many roles in Software Engineering, Data Science, Product Management, Sales, Client Services, and Office Management.
Health insurers, accountable healthcare providers and other health industry stakeholders recognize that U.S. health care spending growth and shortfalls in treatment outcomes are not sustainable. The most foresightful also realize that advances in computational tools to better predict risk and select risk mitigation options, are ripe for transfer.
Sixty percent of high cost members in any given year are new high cost members and were not high cost in the prior year. Yet, current analytics approaches tend to focus on identifying existing persistent high cost members, and then managing attendant care and costs. They focus on only 40% of the problem, and address it after it appears.
At Cardinal Analytx Solutions we apply cutting edge machine learning methods developed at Stanford to predict those who aren’t yet, but are likely to become next year’s new high cost members – “cost bloomers.” Once identified, we suggest precisely targeted interventions that maintain good clinical care, and help avoid their cost increase. We work with care managers and clinical teams to effectively coordinate our recommended interventions.
CARDINAL ANALYTX COST BLOOM MODEL
A carefully constructed, iterative cycle which combines advanced machine learning methods with leading edge interventions to target and address likely high cost events before they occur.
We use a larger more complex data set than conventional approaches.
Cardinal Analytx Solutions defines hundreds of custom features (most are engineered), reflecting a deteriorating patient.
Cardinal Analytx Solutions’ proprietary methods identify next year’s likely high cost members before a high cost event occurs.
Our 1.4 million member pilot correctly discriminated 85% of cost bloomers, accounting for 57% of spend in the top decile of spending.
We recommend highly targeted, member-centric interventions, based on multiple factors and established best practices to improve outcomes and avoid cost increases.
We work closely with care managers and clinician teams to effectively coordinate recommended interventions to improve quality of care and avoid the increase in cost of care.
Given our pilot study, we expect to generate as much as 15% in year two savings (net of losses from unsuccessful interventions) on the top decile expenditure.
We tested the Cardinal Analytx Solution with a large dataset of U.S. health plan members and found that:
The majority of our predictions identified new, future high cost members not yet predicted by the plan’s own internal processes.
The majority of those identified did indeed experience a major health expense in the subsequent year, indicating that our “cost bloom” predictions were accurate.
With precise interventions designed for a sample of identified cost blooms, we estimated a reduced total spending in the first subsequent year of 15% for that group.
The promise of these findings led to further investment by Stanford University and Cardinal Partners.