Integrating Clinical Prediction Rules into EHRs to Improve Care, Reduce Waste

“Very early on, AHRQ understood the language and the work.”

As a young physician in the Bronx, New York, Thomas McGinn, M.D., M.P.H., noticed the tremendous amount of wasteful spending in the U.S. healthcare system. At the same time, he observed firsthand that large numbers of patients and would-be patients in his community couldn’t afford healthcare.

The contradiction confounded him. How could people go without care when the overall system was awash in spending? “This was an inherently frustrating concept to me,” Dr. McGinn recalls. “I started asking, ‘Can we do a better job of allocating resources in an evidence-based fashion?’” This led Dr. McGinn to study clinical prediction rules, or decision aids that clinicians use at the point of care to make evidence-based, cost-effective care decisions, such as whether to prescribe antibiotics or order a scan. 

Today, Dr. McGinn is the Executive Vice President of Physician Enterprise at CommonSpirit Health. A practicing internal medicine physician, he is a nationally recognized researcher with appointments at Baylor College of Medicine and Creighton University School of Medicine. With AHRQ support, Dr. McGinn has developed clinical prediction rules and integrated them into electronic health record (EHR) systems, accelerating their adoption and use.

He credits AHRQ with truly “getting it” from the beginning. “Very early on, AHRQ understood the language and the work,” he notes. “AHRQ is a kindred spirit. They understand why clinical prediction rules are important and how technology intersects with them.” 

Dr. McGinn was no stranger to AHRQ when he came across clinical prediction rules. He received his first AHRQ grant in 2000 to enhance the research infrastructure of New York’s Mount Sinai Health System’s Primary Care Practice-Based Research Network. Working in East Harlem’s urban, underserved communities, his 3-year grant provided the opportunity to strengthen the system’s ties to community partners, develop a computerized data management infrastructure, and identify research relevant to minority and underserved patients.

Dr. McGinn started working on clinical prediction rules to estimate the probability of a disease or clinical outcome. He explains the concept as something of a checklist: “Consider it a five-finger rule. If a patient meets zero of five criteria, stop. Go home. You’re fine. If you meet four out of the five, you should probably get an antibiotic, so let’s write it up. If you meet two out of the five, if you’re in that intermediate range, we’ll test you a little more. But if that low range is 30 percent of patients, we can reduce 30 percent of unnecessary testing and treatment—and that benefits everybody.”

In 2009, Dr. McGinn received a 4-year AHRQ grant to integrate two clinical prediction rules, for strep throat and pneumonia, into Mount Sinai’s new EHR system. This led to significant declines in inappropriate antibiotic and diagnostic test ordering. Analyzing the results, Dr. McGinn and colleagues found widespread clinical and patient acceptance of the integrated clinical prediction rule.

Under this grant, he tested the rules’ usability, which is the interface of culture, workflow, evidence, and technology. “What this taught me was that if I do the usability testing appropriately and thoroughly, I get adoption. If I get adoption, then I can see if this tool is going to have an impact on outcomes. And this was clearly the way to go.”

Building on this work and next working at Northwell Health, a large Long Island not-for-profit system, Dr. McGinn received another 4-year AHRQ grant in 2013 to establish a usability lab at Northwell’s Feinstein Institute for Medical Research. He developed a team of experts in evidence-based medicine, informatics, usability testing, and integration of clinical prediction rules, leading to the identification of factors in overtesting for pulmonary embolism and publishing lessons learned for the field.

In 2019, Dr. McGinn was a co-investigator on another AHRQ grant. This 3-year grant, which built on his earlier work in EHRs, allowed the researchers to develop, test, and measure the impact of a clinical decision support system at Northwell that would function across multiple EHR platforms—a key consideration given the system’s nearly 2 dozen hospitals and more than 700 outpatient facilities. This work also enabled the team to develop a publicly available survival calculator for hospitalized COVID-19 patients.

“AHRQ planted the seeds early on, which I think influenced how other funders are thinking about projects,” he says. “Ten or 15 years ago, I had trouble getting funders to look at a grant. Now, when I write in ‘dissemination implementation’ on a grant application, funders are impressed. Fifteen years ago, nobody was thinking about this—but AHRQ was.”

Principal Investigator: Thomas McGinn, M.D., M.P.H.
Institution: CommonSpirit
Grantee Since: 2000
Type of Grant: Various

Related AHRQ Resources

• Practice-Based Research Networks.
• Clinical Decision Support.

Return to AHRQ Grantee Profiles