Zahoor Elahi is COO of healthcare technology company Health at Scale. Views are the author's own.

Employers around the country shoulder a large financial burden when it comes to employee health plans and benefits. They are responsible for offering benefits that not only meet the gamut of health care needs facing their employee population, but are also affordable and include access to trusted, high-value providers. According to a recent study from Mercer, employers expect health benefits costs to grow 4.4% from 2020 to 2021 alone. These factors, coupled with the uncertainty brought on by COVID-19, mean employers are looking for ways to reduce these costs while continuing to improve outcomes.

Large self-insured employers have served as laboratories for health plan innovation, enjoying the freedom to construct plans that promote a more value-based approach. But novel approaches do not always yield novel results. The "centers of excellence" (COE) model is a prime example. 

COEs create a narrow provider network with a high concentration of expertise and related resources centered on a particular area of medicine. It is an intuitively attractive approach that should yield better outcomes. In fact, a recent survey of U.S. employers conducted by Willis Towers Watson (WTW) found that 73% of employers intend to adopt COEs as they shift toward value-based care delivery and reimbursement models. 

However, the model has major issues. As currently designed, COEs do very little to reduce costs and improve care outcomes for national employers. Even for local or regional companies, COEs have severe drawbacks; they reduce choices available to employees and exclude providers that might be optimal for some individuals at the expense of including those who are good for the "average" person. 

For example, a National Institutes of Health study assessed whether hospitals designated as spinal surgery COEs provided higher quality care. Not only did the researchers find no difference in outcomes between hospitals in COEs and those outside of COEs, they also found that COEs did not lower complication rates, 30-day readmission rates, or 90-day costs compared to other hospitals. 

The WTW survey also found that some COEs only improve employee health and reduce costs half of the time at best. According to the survey, mental and behavioral health COEs only yield improved outcomes 32% of the time, diabetes-focused COEs were just 51% effective, and cardiovascular COEs 45%, to name but a few therapeutic areas. The true cost of a COE to employers is not just the cost of the procedure, which an employer may look to lower with volume-based discounting, but the costs of outcomes, follow-up care, and missed productivity as it continues to accrue over subsequent months to years.

To improve the COE model, employers simply need to account for the most important variable in this equation: the individual employee.

This challenge is evident at a typical COE or academic health center, where there are often dozens of cardiologists, with training in different sub-specialties and varying experience dealing with different kinds of cases. If the employer plan refers an individual to this center and they schedule the next available physician, they haven’t necessarily been guided to the cardiologist best suited to produce their optimal outcome. Even a narrow COE could wind up referring an employee to the wrong doctor. Best case scenario, the individual makes another appointment with a different doctor, requiring the employer to cover unnecessary costs. The worst case involves suboptimal health outcomes, higher risk of complications, and even greater increased costs.

Directing employees to the best facility or physician to suit their needs can be done with considerable precision. Artificial intelligence and machine learning algorithms applied to data from large, representative populations spanning multiple geographies and lines of business offer the opportunity to model how physician outcomes vary from person to person across a broad set of health factors. These algorithms and models, combined with a deeply personalized health profile, can navigate employees to providers who are well-suited to their individual health conditions and needs. This method is not just a major improvement, it is scalable: a national employer could help its employees find the right doctor at the right time at a location of their choice that allows continuity of care closer to home.

There are incentives for employers to adopt a group mentality when constructing certain aspects of their plans. What services are used most? What is the average spend on those services? But a re-think is certainly afoot, focusing on optimizing the delivery system closer to home, ensuring each employee is catered to based on their specific characteristics and, most importantly, in a manner that improves outcomes.