Creating a Learning Healthcare System
Ten years ago, the Institute of Medicine called for healthcare organizations to become learning systems. The need was obvious. The time was overdue for healthcare to follow the lead of all other industries in adopting a culture of integrated learning, personalization, and predictive decision-making. Unfortunately, a decade later, healthcare’s version of learning is still mandated quality metrics, dashboards of the past, and one size fits all risk scores.
Cyft is making learning healthcare systems a reality with four steps:
Step 1: A Systematic Approach to Understanding
Healthcare isn’t one size fits all. Unfortunately, this is the implicit assumption underlying most population health dashboards of quality metrics or risk scores designed to predict the “average” admission.
Making value-based care work requires learning from the data as well as the people responsible for delivering care. Simply installing software cannot meet these needs, which is why we invest in understanding our partner’s unique environments and priorities.
Cyft’s Design Process:
- Organizational Assessment - Economics, contracts, and organizational priorities
- Intervention Assessment - Interventions, operational teams, and workflows
- Data Assessment - Available & required data, challenges, and opportunities
Step 2: Match Individuals to Your Interventions
Cyft identifies the individuals most likely to benefit, in rank order, from your specific interventions. Our technology is built on 10+ years of R&D into using all diverse health data to quickly build and evaluate thousands of models from millions of data points.
Move beyond generic “risk.” People have widely differing needs, from geriatric patients managing polypharmacy, to high-risk pregnancies, to serious mental illness. Knowing which older person is likely to fall, which Medicaid member is eligible for additional reimbursement resources, or which members are likely to disenroll is a distinct advance beyond traditional analytic approaches. Learn more about Precision Care Management.
- Use all of your data - leverage call center transcripts, EHR, care management notes, and more.
- Focus on what’s actionable - move from generic “risk” to precision intervention targeting
- Empirically validated with your data - know how each model will perform in your environment
Step 3: Deploy to Production
Delivering results at the right place and time. This is where our design effort pays off with carefully planned and executed deployment into existing workflows. Cyft’s API-driven approach enables full integration with your existing IT investments - from care management systems to EHRs - without having your team log into yet another screen.
Fully integrate into existing workflows and systems
- Realize full potential of existing care management systems / interventions
- Careful deployment planning to improve adoption and minimal disruption
- Use yesterday’s data to drive today’s decision making
Step 4: Continuous Learning
Your team and Cyft models continue learning., Learning is an ongoing, dynamic process requiring constant feedback, iteration, and adaptation. Our models continue learning with each data point to reflect the dynamic nature of your population and their needs. We use this same data to produce progress reports or activities and outcomes so your team can see which of their efforts is or isn’t working.
Continue Learning Over Time
- Cyft models continue to learn with experience
- Teams continuously improve based on insight into what is and isn’t working
- Measure operational and clinical performance with the metrics that matter to you