Health Fidelity was one of the first technology partners that deployed a natural language processor (NLP) engine to support complete and accurate retrospective risk adjustment through the addition of otherwise nearly unreachable unstructured clinical data. Over the past decade, that engine, today known as Lumanent® Insights, has:
- Supported 9.5+ million lives,
- Processed 139.2+ million clinical records,
- Suggested the accurate risk adjustment category more than 95% of the time.
Given the success of applying Lumanent Insights to risk adjustment at 6 of the 10 largest provider sponsored plans in the US, Health Fidelity expanded the use of its NLP to add prospective risk adjustment, deriving clinical suspects, in late 2015. In the half-decade since, Health Fidelity’s clinical suspecting has yielded 25% more confirmed conditions with 10% greater RAF value over structured-data counterparts for its partners. In circumstances where NLP suspects are integrated with traditional risk analytics sources, EHR problem lists or health plan claims data, the acceptance rate further improves, deepening the utility of existing resources.
CEO Steve Whitehurst observed, “At its core, risk adjustment requires a full understanding of everything a patient is going through medically. Healthcare as an industry excels at gathering data. Putting that data to work is a different story.” He continued, “Five years ago, we began leveraging NLP technology to meet the advancing need to completely capture risk at various points of the patient continuum, so our partners can proactively find conditions that may need more rigorous care in a way that lifts, rather than burdens, providers.”
Today, Lumanent Suspects can be incorporated into various Lumanent® workflow modules, from Pre-Encounter Prep to Retrospective Review, and is currently embedded in other market leading population health and point of care platforms too. Through these diverse channels, Lumanent Suspects is delivering value at 4 out of the top 10 non-profit health systems in the United States.
The success of Lumanent Insights begins with Dr. Carol Friedman, a professor of the Department of Biomedical Informatics at Columbia University and a Morris F. Collen Award winner. Since partnering with Health Fidelity in 2011, Dr. Friedman has overseen the NLP engine’s development and maturity. It continues to benefit from her ongoing research; its robustness, when compared to other NLP engines that were rushed to market, is attributed to its blend of clinical specificity and general linguistics applicability. She notes, “We built it to benefit from, and amplify, human expertise and to recognize general medical language patterns. As a result, it understands clinical language and how it’s used, even as physician notes and health care facilities change, without having to be retrained. Lumanent Insights is still the best at what it does.”