How one health system trimmed clinical spending waste by improving decision support

An effective clinical decision support (CDS) system is paramount in initiatives to curb wasteful spending in clinical settings, says Arun Mathews, MD, regional chief medical officer with MultiCare, a 13-hospital system based in Tacoma, Wash.

CDS technology incorporated in MultiCare’s electronic health record has helped save $179.14 per accepted recommendation regarding lab and medication orders, according to a report from the vendor that implemented the alert system.

A prominent application of the alert is seen in the provision of sodium bicarbonate to treat acidemia. The treatment is utilized inappropriately more than 80% of the time, according to the vendor’s report, at a cost of nearly $1 million per year at one organization. Plans are underway to expand the initiative to radiology.

“Timely decision support in the moment that a person is contemplating a particular healthcare decision — that is a very powerful moment,” Mathews said. “The challenge is that a lot of folks have realized that’s a powerful moment, and we’ve inadvertently created this notion of decision-support fatigue, or alert fatigue.”

In a recent Q&A, Mathews described both the technology and the approach the organization took with its contracted physicians from a change management perspective (responses have been edited for length and clarity).

What spurred MultiCare to consider this new alert system?

It was just recognizing the financial impact of this generational black-swan event that was the pandemic, and the financial recovery and the work toward that financial recovery. One of the pieces of that was at least trying to do an analysis of opportunities around healthcare waste down to the level of the ordering prescriber. Were there opportunities there?

Traditionally, we had done some work in terms of clinical stewardship, but [now] we partnered with this technology firm to [take] a deeper dive and actually look at normative benchmarks for opportunities around healthcare waste, tied to the Choosing Wisely database.

With that first-pass analysis came two big realizations: We had opportunities in medications and lab ordering, and the vast majority of those opportunities happened with the folks that rendered the vast majority of the care — our hospitalist physicians, who touch 80% to 90% of our inpatient encounters.

What were some of the key considerations in implementing the technology?

It wasn’t a magic bullet. We had to do a genuine outreach and engagement initiative with the physicians.

That initiative uncovered some pretty powerful insights, which all had to do with alert fatigue and a sense that every time a new notification came up, there was a significant amount of cognitive load placed on a physician.

We actually did a cognitive workflow diagram unpacking the way a practitioner thinks when they are interacting with an alert. The first thing they have to do is acknowledge the alert. The second thing they have to do is appreciate [whether] this alert is truly relevant to the decision around the medication or lab order that they’re about to move forward with.

And then they have to reconcile: “OK, do I trust this alert? Let me click the reference. OK, I trust that reference,” or “I trust the Choosing Wisely database,” and then they actually have to do the work of changing the order or stopping the order or thinking about their care a little bit differently. All of that amounts to what I consider to be something called clinical cognitive burden, and our big insight was that exists for any alert.

With that in mind, how did you communicate this initiative to physicians?

We needed to figure out a way so that they feel that their time interacting with this alert and doing this additional work has value. We partnered with their parent contracting company and developed a clinical stewardship quality metric that actually was tied to healthcare waste reduction down to the individual-physician level.

Clinical variance and waste reduction based on a practitioner’s own practice patterns is something that’s very unique. And because of the granularity of this data, we were able to define target metrics down to the level of the practice, which was kind of a wonderful breakthrough.

I bet now when the physicians came across these alerts, they knew that they were getting credit for interacting with the alert and doing the work of either supporting or negating the alert. That’s what we were looking for. The moment we released that clinical stewardship metric, that’s when we saw hockey-stick [increases in] engagement with the tool, and that’s when the flywheel of clinical cost reduction really started to turn.

[The new metric] joined the other metrics in a dashboard. It took place over five hospitals over the course of a calendar year [2023]. We were able to demonstrate about $2.5 million worth of waste reduction.

What workflow considerations factored into implementation?

The next phase of [the vendor’s] engagement with us was looking at the additional work of stopping or starting a new order associated with that alert. Within the substance of that alert, if you decide to support the recommendation, in the background it actually does the [work] so that you don’t have to get out of the alert screen and go into the ordering screen.

To me, as [a trained] informaticist, workflow design is just really sacred. And this was a perfect example of human-factors engineering and listening to the end user to make that experience a little bit better.

What balance were you attempting to strike with this technology?

The work of this was finding a balance between bubbling up only the most important, relevant opportunities, but also figuring out a way to recognize the cognitive load that a physician or an advanced-practice practitioner [encounters].

The moment we realized that was the moment that we were able to say we need to recognize that value through some sort of shared savings mechanism, so that it’s not just another pilot on the practicing physician but [instead] make it so that additional cognitive work is recognized, compensated, aligned. Building out a shared savings methodology, I think, was the real innovation here.

How did the shared-savings approach get implemented?

Naturally, when you talk about shared savings, you want to be cautious because certainly we don’t want to be accused of rationing healthcare to have physicians earn some sort of quality-metric bonus. We did a literature search to make sure that there were precedents of aligning incentives around shared savings. And we actually found that there are multiple examples of this in the literature, primarily in the value-based care space, oncology fields and orthopedics.

We took that literature search, created an SBAR [Situation-Background-Assessment-Recommendation] document, sent it to our legal and compliance teams to make sure that it would pass muster. Their feedback was that they were supportive of moving forward with this clinical stewardship quality metric that employed shared savings as a mechanism of incentive. But they recommended tracking countermeasures just to make sure that there wouldn’t be any unanticipated outcomes.

The countermeasures that we tracked were readmissions and patient experience, and I’m pleased to report that we saw no statistically significant drops in either of those two while we were measuring our clinical stewardship quality metric performance.

How does this type of initiative set you up to succeed in the current political and federal policy environment?

Clinical variation is not the enemy. It is, in fact, the cornerstone of deterministic and relational care.

However, [for] big variances, we want to uncover instances where there may be a practitioner who is not aware of a treatment that is equally efficacious but is less expensive to the healthcare system, to the patient writ large. This work helps us identify those outliers.

There are thresholds of clinical variance that I think hospitals and healthcare systems, because of the evolving financial constraints, should be aware of.

But quite frankly, even if that were not the case, I think it’s just better medicine to practice medicine that doesn’t financially become a catastrophe for your patient — recognizing that your patient is more than just a container of disease, and actually after their treatment is going to be on the hook for some amount of cost associated with that [treatment], and what that might do to their overall sense of health and well-being.

If we can be better stewards of the care that we can administer, and [we can] offer to our patients equivalent care that may be more cost-effective, I think that’s part and parcel of being a modern healthcare provider.