MedCity Influencers, Artificial Intelligence

The Case for Automatic Surveillance during Dialysis

This lack of monitoring is a serious problem—both for patients and for the overall healthcare system. The data show that in a typical dialysis center, each patient will suffer an average of 1.7 access blockages per year.

For the many thousands of people with kidney disease, dialysis remains a lifesaver. But providing dialysis services has an Achilles heel—the challenge of maintaining an adequate vascular access to be able to withdraw blood for dialysis and return it to the patient. These access points are difficult to create and need to be protected from problems like clotting (thrombosis).

Maintaining an adequate vascular access is so important that the Centers for Medicare and Medicaid Services (CMS) mandates that dialysis providers monitor the health and function of the access points. Historically, the two types of such monitoring have been a physical examination of the access or the use of an ultrasound probe to check for possible clots or other blockages.

Unfortunately, in practice, these forms of monitoring are rarely used. Both take extra time and staff effort, which adds costs and increases the time patients must spend in dialysis centers. This lack of monitoring is a serious problem—both for patients and for the overall healthcare system. The data show that in a typical dialysis center, each patient will suffer an average of 1.7 access blockages per year. When these episodes of thrombosis do occur, not only can patients miss their dialysis treatment, which can put their lives at risk, they also often end up in the hospital, where doctors need to perform procedures to try to open the access point. Sometimes patients will end up with a temporary dialysis catheter threaded through the large veins in their necks into their hearts, which is risky and prone to infections. Indeed, undetected vascular access blockages in dialysis patients cause about 20% of the hospitalizations and many deaths among the half million Americans on dialysis.

The economic toll is enormous—an average of $26,000 per patient per thrombosis that leads to access loss. That adds to an estimated $2.8 billion in healthcare costs every year. Yet the fee-for-service medical system in the United States offers no incentives for providers to reduce those costs or to improve patient care by reducing the rate of access blockages.

I’ve been a practicing nephrologist for more than 25 years, and realized early on the need for a solution to this problem. So I was encouraged by the discussions I began having about 15 years ago with my colleagues at Henry Ford Hospital, where I did my training. They reasoned that dialysis machines already automatically collect a lot of potentially useful data as they filter each patients’ blood, such as the patient’s arterial and venous pressures. So they asked a key question: could these data provide clues about the risks of blockages in the patients’ vascular access?

The answer turned out to be a resounding yes. The Henry Ford physicians developed an algorithm that uses all the data collected by the dialysis machine to compute an overall score, from 1 to 10. The higher the score, the more likely it is that patients will experience stenosis and require interventions to prevent complications.

Since then, the Henry Ford team has improved the algorithm to make it more predictive. Meanwhile, once they decided to roll out the technology, the two vascular access centers in my practice were among the first to try it out. We thought that we might be able to provide a valuable service to our patients by using this new form of access surveillance.

The results exceeded our expectations. By automatically calculating a risk score for potential blockage during each dialysis session, we’ve dropped the number of thrombosis events from the average of 1.7 per patient per year to only 0.7 to 0.8—an improvement of more than 50 percent. Similarly, our rates of catheter use have fallen to only 8 percent, less than half the national average of 17-18 percent. That’s great for our patients and for the cost of healthcare since every thrombosis event avoided saves $26,000. Moreover, calculating the risk score takes no staff time or effort, and the expense is trivial.

This technology also fits into a larger trend in healthcare–the use of analytics to provide new value by improving care and reducing overall costs, particularly in kidney care, where it can help avoid high-cost vascular access complications.

Currently, however, this risk score approach is being used by only a few hundred of the thousands of dialysis centers in the United States. So, we need to dramatically accelerate the adoption of this approach. Encouragingly, there are two reasons to expect that this could happen. One is the recent publication of a study in the Journal of Vascular Access validating the technology. Not only did the study show that high risk scores successfully predict access problems, it also showed that the technology works better than the other two surveillance methods, physical examination and ultrasound. We can now convincingly show doctors that this approach is a major step forward in the treatment of patients with failing kidneys.

The second reason is the CMS is rolling out a new payment system, called value-based care, which puts dialysis providers on the hook for the risks—and costs—of vascular access issues. So far, only about 30 percent of the nation’s dialysis centers come under the new system, but for those that do, there’s a powerful financial incentive to avoid access blockages and the high costs of responding to a thrombosis event.

Now, we have a unique opportunity to drive down the rates of thrombosis and catheter use to historic lows. It’s a rare win-win for the healthcare system, improving patient outcomes while also significantly reducing costs.

Photo: Edwin Tan, Getty Images

Editor’s Note: The author is not compensated by Henry Ford Hospital but is on the board of the company whose algorithm was studied by Henry Ford and whose results were published in the Journal of Vascular Access. 

Avatar photo
Avatar photo

Robert Provenzano

Robert Provenzano, MD, FACP, FASN, FNKF, is an Associate Professor of Medicine at Wayne State University School of Medicine in Detroit, Michigan.

He has served previously as the Chief of the Division of Nephrology, Hypertension & Transplantation, Director of Nephrology Research and Director of Acute Dialysis Services at St. John Hospital & Medical Center in Detroit; Chief Operating Officer of St. Clair Specialty Physicians in Detroit; and the Vice-President of Medical Affairs for DaVita. Most recently, he has overseen all the owned and managed practices for DaVita as its CMO for Nephrology Practice Solutions and is currently CEO of InnoCura Nephrology.

Dr. Provenzano works with several professional societies and advisory boards and has received several awards and recognition as a clinician, researcher, and educator.

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.