Press

Trilliant Health Announces Machine-Learning Model to Benchmark Hospitals with SimilarityIndex™ | Hospitals

Written by Trilliant Health | September 8, 2022

A new application that informs health system strategy and capital allocation, SimilarityIndex™ | Hospitals shows how more than 2,000 U.S. hospitals truly compare across quality, operational, and financial metrics.  

Trilliant Health launches a web-based, interactive data visualizer that reveals a hospital’s most similar peers as well as their relative distance from aspirational peers. 

BRENTWOOD, Tenn. – September 8, 2022 – Trilliant Health, the leading predictive analytics and market research firm in the healthcare industry, today announced SimilarityIndex™ | Hospitals, a machine-learning-driven application that allows every stakeholder in the health economy to understand how U.S. hospitals truly compare and benchmark against each other.  

SimilarityIndex™ | Hospitals introduces a new methodology to accurately compare and benchmark more than 2,000 acute-care hospitals in all 927 U.S. core-based statistical areas, informing evidence-based strategic planning for competitive analysis, clinical quality program management, M&A, and other strategic initiatives. 

“The 800+ ‘Top 100 Hospitals’ may be useful for marketing campaigns, but they are useless for developing evidence-based strategies,” said Hal Andrews, CEO of Trilliant Health. “Evidence-based strategies depend on accurate benchmarks against relevant peers. The healthcare industry is both capital-intensive and capital constrained, and applying mathematical rigor to benchmarking and strategic planning reduces the risk of suboptimal capital allocation for healthcare providers, suppliers, and payers. Our SimilarityIndex™ | Hospitals application allows stakeholders to understand how hospitals truly compare with respect to quality and other important metrics.” 

Combining algorithmic intelligence with the industry’s most robust data sources, SimilarityIndex™ | Hospitals includes an interactive data visualizer that uses normalized Euclidean distance in multi-dimensional space to determine the mathematical distance between hospitals across a variety of measures.  

 The Quality Similarity Score shows how similar hospitals are to each other in clinical quality based on several CMS quality metrics. The Aggregate Similarity Score combines Quality with other categories that inform the SimilarityIndex, including Outpatient Service Line, Financial Similarity, Patient Mix, and Market Share. 

Historically, healthcare decision-makers have relied on lists like U.S. News & World Report “Hospital Honor Roll,” Healthgrades and others to justify business strategies or compare themselves to “aspirational” peers. However, these approaches lack critical data inputs, important context, and adequate benchmarking elements, which can often result in arbitrary, incomplete, and even irrelevant comparisons. 

In contrast, the mathematical foundation of SimilarityIndex™ | Hospitals reveals that hospitals can be peers to one group of hospitals in quality and to a completely different group of hospitals in another category. Notably, SimilarityIndex™ | Hospitals demonstrates that the most renowned hospitals in the U.S. are, in fact, quite different from each other. For example, using the SimilarityIndex™ | Hospitals to compare quality measures reveals Johns Hopkins Hospital’s true peers are hospitals like Thomas Jefferson University Hospital and Tampa General Hospital as opposed to renowned hospitals like NYU Langone Medical Center or Cedars-Sinai Medical Center.  

“Over time, these ‘Top 100’ lists designed for consumer-marketing purposes have been incorporated into strategic planning, even though arbitrary rankings that incorporate self-reported survey data and subjective measures such as ‘perceived prestige’ are inadequate to compare hospitals. The fact that Johns Hopkins Hospital and Cedars-Sinai are both frequently in a ‘Top 10’ list does not mean that they are comparable in every respect, and neither is a relevant peer for the vast majority of U.S. hospitals because of differences in patient acuity, payer mix, market competition, and more,” said Trilliant Health Chief Research Officer and SVP of Market Strategy Sanjula Jain, Ph.D. “Our goal is that the SimilarityIndexwill fundamentally change how health economy stakeholders think about and compare hospitals to develop strategies that are mathematically based and objectively true.” 

SimilarityIndex™ | Hospitals is the second application to launch from Trilliant Health’s SimilarityEngine, the health economy’s first machine-learning-based similarity model, following SimilarityIndex™ | Markets in June.  

SimilarityEngine™ processes hyper-local healthcare supply, demand, and yield data from a curated mix of publicly available and proprietary data sources, including Trilliant Health’s all-payer claims database that informs longitudinal patient journeys for more than 300 million Americans. Stakeholders across the health economy rely on Trilliant Health and its SimilarityEngine™ applications to create customized, evidence-based strategies.  

 About Trilliant Health 

Trilliant Health combines healthcare industry expertise, market research, and predictive analytics to form Evidence-Based Strategy for Healthcare™. Trilliant Health's proprietary analytics platform produces a comprehensive understanding of local market dynamics providing exponentially better data insights to maximize returns from growth strategies.

 

Journalists can reach the Trilliant Health media team at: media@trillianthealth.com