Technology Tackles a Persistent Clinical Problem: Unwarranted Variation

Remember when almost everyone accepted a doctor’s clinical decision without question? It wasn’t long ago that diagnosis and treatment were considered a medical art, and every decision a doctor made was considered the right one. John (Jack) Wennberg, MD, MPH, deserves considerable credit for revealing that it has not – and maybe never has.

More than 50 years ago, Wennberg first described four different types of unexplained variations in clinical practice: underuse of effective care, variable outcomes attributable to quality of care, abuse of preference-sensitive treatment, and overuse of finite resources. As we entered the 21st century he observed that there are still significant differences in the use of health services that cannot be explained by corresponding differences in illness or patient preferences.

This is not about the occasional mistake made by a surgeon in an urgent situation or an important finding that was inadvertently overlooked by an overworked doctor. The concern is that treatment decisions vary unjustifiably – that is, different doctors act differently in very similar clinical circumstances. These differences are by no means trivial, but can increase the risk of physical and financial harm to the patient.

Despite extensive research on this disturbingly widespread problem, it persists and efforts to contain it have met with modest results at best. Although decades of research have improved our understanding of the underlying causes of unjustified deviations, we have found no “cure”. Some aspects of the problem could be addressed by promoting compliance with existing standards and published professional guidelines. What we lack are effective ways to prevent wrong decisions before they are made.

What could work? “Guard rails” to regulate clinical practice have been successfully used to prevent unwarranted prescriptions of certain drugs (e.g., or perform an angiogram.

Once again, technology comes to the rescue! This time with platforms and systems designed to support good clinical decision-making. I’ve followed reports from a few companies that are at the forefront of developing these new technologies: As a nationally recognized leader in this field since its inception in 2016, QURE uses artificial intelligence to make healthcare more accessible and affordable. The focus is on defining clinically relevant problems and developing practical solutions that improve the quality of care and the costs of treating certain diagnoses.

Example: A study was carried out to reduce unnecessary differences between primary care clinicians participating in a responsible care organization. An application should support practicing physicians with evidence-based feedback. This standardization initiative measured care decisions by providers and individual feedback to improve individuals and groups. The results showed a significant increase in the evidence-based quality scores (+ 27%) and a corresponding reduction in unnecessary tests (-55%).

Agathos: This relative newcomer to the field has developed technology to give doctors direct insight into practice variation. The aim is to normalize value-based care and make it more visible. In addition to incorporating current guidelines and best practices from colleagues, the platform also collects data from clinical records. The application offers real-time feedback based on clinical action patterns as well as current case studies with an actionable clinical context. Action-level feedback accelerates clinician engagement and improves collective performance.

Example: One study compared opioid prescribing rates for 15 hospital staff in a large hospital before and after receiving individualized metrics and providing insights on the move. The results showed a 12% decrease in opioid prescriptions in the first 60 days, with a further decrease thereafter.

I have been waving the flag in the fight for safer, higher quality care (without wasting precious and costly resources) for many years – but not nearly as long as the venerable Jack Wennberg. Unfortunately, many doctors have little time or interest in keeping guidelines and best practices updated. Could these new, real-time, evidence-based aids succeed where previous attempts have failed? I think they could!

David Nash, MD, MBA, is founding dean emeritus and Dr. Raymond C. and Doris N. Grandon Professor of Health Policy at Jefferson College of Population Health. He is special assistant to Bruce Meyer, MD, MBA, President of Jefferson Health. He is also the editor-in-chief of the American Journal of Medical Quality and from Population health management.

Last updated on August 30, 2021