Most healthcare organizations are drowning in data but are challenged to gain reliable, actionable insights from this information. It is in the form of physician notes, registration forms, discharge summaries, documents and more. This data is doubling every five years. Different from machine-ready data, this content lacks structure and is arduous for healthcare enterprises to include in business analysis and therefore is routinely left out. As a result, millions of patient notes and records often sit unavailable in separate clinical data silos. This content contains valuable information, but there’s historically been no easy way to analyze it. While there’s no silver bullet for preventing all re-admissions, hospitals and health providers can take action to significantly decrease the occurrence.
More than 80% of the client’s data was unstructured. It was in the form of physician notes, registration forms, discharge summaries, echocardiograms, and other medical documents. The client knew they had to leverage their wealth of unstructured information to discover new, population-specific clinical and operational insights and significantly reduce the occurrence of high cost CHF re-admissions by proactively identifying patients likely for re-admission and introducing early interventions to reduce cost, mortality rates, and improve patient quality of life.
How We Solved It
Trillium managed the data mining, modeling and implementation team that introduced a healthcare industry software offering that combined content analytics and natural language processing technology that helped both the health care providers and payers improve patient care and reduce costs.
Leading the entire project and predictive analytics solution team while working directly with a world leader in natural language processing, Trillium was able to provide a Content and Predictive Analytics solution that allowed the client to extract relevant clinical information from vast amounts of patient data to better analyze the past, understand the present, and predict future outcomes.
Combining natural language technology with predictive analytics allowed the client to identify the root causes of hospital re-admissions, and the ways it could decrease preventable multiple hospital visits.
- By predicting readmission candidates and introducing mitigating strategies, the client reduced costly and preventable readmissions by 7%
- Eliminated the need for traditional analysis a resource intensive task and transformed raw information into healthcare insight quickly.
- Revealed trends, patterns, deviations, and the probability of outcomes. Enabled client to derive insight in minutes versus weeks or months.