Predictive Analytics Technology
Profility combines advanced profiling based on big data with proprietary algorithms to discover hidden patterns and previously hard-to-identify groups of similar patients. Big data holds significant promise, but it only works if you can turn data to knowledge, and put that knowledge to work, and that is where Profility comes in. Predictive analytics provides vast potential to improve the way we dig into existing information and apply new knowledge to deliver better healthcare.
Today, most analytics solutions solve for the “average patient” in other words, a blend of actual patients as illustrated below.
Current healthcare is delivered to the “average person”.
The “average person” at the center is a blend of its nearby portraits. BUT – this person does not actually resemble any of the other individuals.
When you deliver “average care” individual differences are lost. This leads to poor outcomes and higher costs.

Profility uses advanced profiling to construct homogenous clusters of patients from the bottom up, starting with a specific patient’s individual profile. Patients within each cluster share characteristics, and can thus be used to understand the effectiveness of different treatments and different providers.




Profility Tools





PROVIDERS
(Hospital Systems, Accountable Care, Organizations and Physicians)
- Decision making support tools for doctors at point of care
- Rehabilitation success or failure predictions
- Optimization of patient placement following hospitalization
- Outcome and cost optimization
- Quality Indicator benchmarks for Post-Acute Care facilities
- Hospital discharge, readmission reduction

PAYORS
(e.g., Insurance Companies)
- Health profiles for personalized care plans
- Predictive analytics to target preventative care
- Support pay-for-performance using costing metrics

GENOMICS / PHARMACEUTICAL COMPANIES
- Discovering homogeneous subpopulations that are relevant to the success of a new drug or treatment

RESEARCH ORGANIZATION
- Upgrading existing prediction models by adding a personalized dimension to the prediction process