Analytics reframes decisions from bench to C-suite

Order codes are a feature that has evolved as the dashboards have been refined. For example, the team started with one order code for COVID tests but soon realized it needed separate codes for rapid tests and immediate tests. With those, “we were able to translate those order codes into actionable filters on our tools to help people make decisions,” Dr. Henricks says.

Changing the order codes was complicated, Muenzenmeyer recalls. “We did have to go back into our database and review our coding to make sure we had appropriate flagging mechanisms for the new order codes. That was challenging. Not just the order codes, but the way these tests get resulted with different English texts posed initial challenges for us. But once we got the hang of our database flagging mechanisms, it was just a matter of keeping up with the operational and IT changes.”

Similar dashboards have become standard for other laboratory projects at the institute. “That was always part of the vision Dr. Henricks and I had in mind,” Muenzenmeyer says. “We took every failure and every lesson learned during COVID and translated them into standards that we now apply to all of our lab metrics and projects and required analytical tools. So we have several dashboards for all of our groups in clinical pathology and anatomic pathology to help drive decision-making and ultimately add value.”

The COVID-19 dashboards are geared to a larger goal as well, one that Dr. Henricks sums up as “connecting the dots.”

“The question is: How does pathology and laboratory medicine provide value and document value in the health care system, beyond just the obvious role of reporting our results?” he says. “How do our activities affect other outcomes in health care and health care organizations, whether they be clinical outcomes like infection rates or other quality measures that we report to the government, or economic and financial outcomes?”

He uses the phrase “connecting the dots” to talk about the linking of laboratory data with other data sources in the organization to demonstrate, document, and find opportunities for how laboratory work affects clinical and financial outcomes that aren’t specifically laboratory numbers.

“The analytics unit has the express purpose of improving how we use our data, report our data, and visualize our data,” Dr. Henricks says. But a larger purpose is to allow pathology and laboratory medicine data to add value to the entire system. “We built an organization to improve our data analysis and visualization capabilities. It is laboratory centric, but we are very closely aligned with our enterprise analytics group and we’ve been able to integrate our analytic capabilities with theirs.”

One sign of that alignment is that the Cleveland Clinic’s executive team has been drawn to the COVID dashboards, which laboratory staff know because of the messages they get when the dashboard updates are interrupted for even a short while. “I can tell you that if we had any issue with the dashboards not being up and functional, I received phone calls and emails from across the enterprise within 30 minutes,” Muenzenmeyer says. “It’s clear they are being used to make decisions both internal to the institute on our leadership calls and at the enterprise executive level.”

“Our C-suite looks directly at some of our dashboards,” Dr. Henricks says. “They like the information they get from the laboratory, and they like the way we present it. That in itself shows the value we provide.”

In Muenzenmeyer’s view, the health care industry in general is not yet up to speed with respect to analytics, including the capability to visualize data. “During COVID, it became obvious that analytics is its own business function, like finance or IT, that requires specific skill sets and tools. And health care organizations need to make a greater investment in analytics after not being able to produce things quickly enough during the pandemic. That’s not only to help decision-making based on asking questions about what happened yesterday or why. It’s about moving up the maturity curve in analytics and being able to predict what will happen and to adjust business plans accordingly.”

Deploying these more sophisticated information tools requires expertise and commitment, which means resources and money, notes Dr. Henricks, who says that partnering with institutionwide analytic efforts might be one way that laboratories could get help in obtaining those resources.

The infection prevention group at Cleveland Clinic recently requested a seven-day rolling average of COVID positives, so rolling averages were incorporated into the institute’s dashboards. “Why did they want it? That’s the metric we turn in to the CDC that they are looking at to inform masking and protocols. The infection prevention people may want it to decide which populations we test and when we should pull the triggers for pre-procedure tests.”

Typically, Muenzenmeyer adds, the laboratory is seen as the first group to ask since the data are already curated and held in its systems. It’s also important, Dr. Henricks says, that the team was able to turn around a fairly sophisticated request in less than a day for significant decision-making about infectious disease.

Muenzenmeyer is enthusiastic about the experience and skills the analytics team has gained from the pandemic. “Now we have a solid understanding of best practices and standards in place that allow us to develop things quickly, and we have more expertise. We know how to handle different types of data. If we get a request for a moving average again, we’ll be able to show some of our new team members this dashboard as an example. So we’re using everything we’ve learned from this pandemic to keep improving our data projects.”

As the pandemic winds down or just shifts in shape, Dr. Henricks thinks the COVID dashboards will continue to be useful and says other laboratories considering an investment in analytics are well advised to make the move. “This capability we’ve built is going to be tremendously valuable and this is really the crucible in which it was developed,” Dr. Henricks says. “We’re extending all the things we’ve learned to apply these tools throughout our laboratory and other elements of our operation. And we’re testing and looking for other opportunities to use analytics as we go.”

Anne Paxton is a writer and attorney in Seattle.