Whether either or both of these suggestions can be put into practice is the challenge most clinical laboratories face

For Pathologist Ramy A. Arnaout, MD, DPhil, one of the biggest issues all pathologists face today is how to overcome the breakdown in cooperation between pathologists and referring physicians that can cause patient harm.

An Associate Director of the Clinical Microbiology Laboratories at Beth Israel Deaconess Medical Center (BIDMC), Arnaout was a panel member during a webinar in December sponsored by STAT News and T.H. Chan Harvard School of Public Health. During the webinar, “Medical Tests: Inaccuracies, Risks and the Public’s Health,” Arnaout explained that when errors occur in a lab, they usually happen during test selection and result interpretation, sometimes called the “pre-pre-analytical” and “post-post-analytical” phases. In these two phases of the lab-testing process, pathologists and ordering physicians need to collaborate more closely to help avoid errors and reduce the level of patient harm, he explained.

Medical Laboratory Tests Are Single Highest Volume of All Healthcare Activities

“Lab tests are the single highest-volume medical activity,” stated Arnaout during his webinar presentation. “We will do 10 billion such tests this year nationwide, and we have done most of these tests for so long and some are not terribly complex. So, when something goes wrong, it’s not in the performance of the test itself—it’s not in the analytical portion of the process. It’s usually in the interpretation of the result. And what might be more important is the choice of what test to order in the first place.”

In a widely reported study titled, “The Landscape of Inappropriate Laboratory Testing: A 15-Year Meta-Analysis,” and published in Public Library of Science’s multidisciplinary open-access journal PLOS ONE, Arnaout and colleagues at BIDMC, Harvard Medical School, and the Harvard TH Chan School of Public Health, examined the rate of overuse and underuse. “We studied that issue to see overall on average how well we do at ordering laboratory tests. Our study was a 15-year systematic review and meta-analysis that covered 1.6 million different performed tests that accounted for 46 of the 50 most commonly ordered tests in this country,” Arnaout explained. “It included tests such as the complete blood count, electrolytes, liver function, and iron studies. What we found surprised us.

Ramy A. Arnaout, MD, DPhil (above), is an Associate Director of the Clinical Microbiology Laboratories in the Department of Pathology at Beth Israel Deaconess Medical Center and Assistant Professor of Pathology at Harvard Medical School. A mathematical biologist, Ramy used mathematical and evolutionary analysis to study HIV and genetic regulatory networks as a visitor at Princeton University's Institute for Advanced Study. His scientific writings have appeared in Immunology Today, the Journal of Theoretical Biology, and the Journal of Virology. (Photo and caption copyright: The Paul & Daisy Soros Fellowships for New Americans.)

Ramy A. Arnaout, MD, DPhil (above), is an Associate Director of the Clinical Microbiology Laboratories in the Department of Pathology at Beth Israel Deaconess Medical Center and Assistant Professor of Pathology at Harvard Medical School. A mathematical biologist, Ramy used mathematical and evolutionary analysis to study HIV and genetic regulatory networks as a visitor at Princeton University’s Institute for Advanced Study. His scientific writings have appeared in Immunology Today, the Journal of Theoretical Biology, and the Journal of Virology. (Photo and caption copyright: The Paul & Daisy Soros Fellowships for New Americans.)

“First we found that the rate of overuse, meaning [medical laboratory] tests that we performed that shouldn’t have been performed, average around 20%. That means out of every 10 tests ordered about two tests on average should not have been ordered.

“Even more surprising was the rate of underuse [of medical laboratory tests]. This is the rate of tests that should have been ordered in a given time but weren’t ordered. That rate was over 40%. So, for every three tests performed, an additional two tests should have been ordered. That shows there’s a lot of room for improvement,” emphasized Arnaout.

At this point in his presentation, Arnaout addressed the need to improve communication during the lab-test ordering process. “In our clinical experience, it’s almost never the case of an individual person not meaning to do the right thing. It’s more about breakdowns in process and misaligned incentives than it is about people. This is about quality improvement in lab medicine that has important implications for quality and safety and patient care.”

Cooperation Breakdown Leads to Inappropriate Clinical Lab Test Ordering, Says Arnaout

STAT News Senior Editor Sharon Begley asked Arnaout if the over- and under-testing is the result of physicians failing to consult with pathologists or other physicians when ordering clinical laboratory tests. “Is the clinician making the call on his or her own?” she asked. “Isn’t it a collaborative process?”

“The question you’re asking forms one of the frontiers in laboratory medicine: which is to understand what we call inappropriate ordering and what to do about it,” Arnaout answered. “There are many individual issues that all come together in some form of a breakdown in cooperation.

Improving Collaboration within the Hospital to Improve Ordering of Laboratory Tests

“To answer the question about how to improve collaboration inside a hospital we have to ask, what’s the purpose of ordering a laboratory test or any test? It’s almost always to answer a question of something like, are the signs and symptoms of my patient explained by an infection or whatever it is that the test is trying to answer?” Arnaout said. “If so, what, if anything, can I do about it?

“When you cast clinical decision making in these terms you see that different people—such as the patient, physicians, a nurse, or a pharmacist—are experts on different parts of that question. The clinician sees the patient and best knows the question,” Arnaout said. “The laboratorian, or the clinical pathologist, best knows the test and the test parameters and is trained to think most about issues of probability, the prevalence of a particular condition in a population, what that means for positive and negative predictive values, and the specificity and sensitivity values of the test.

“In the hospital we all have some passing acquaintance with these terms. But largely they fall by the wayside when a doctor is caring for a patient and that can lead to errors,” Arnaout explained. “That suggests the need for collaboration between the clinician who is seeing the patient—the patient obviously is at the center of this—and the laboratorian to help decide which tests to order.

Involving Pharmacy Appropriately When Interpreting Medical Laboratory Test Results

“On the other end—how do you interpret those results?” Arnaout asked. “Do you work with someone from Pharmacy perhaps to help? Especially if one of the options under consideration is pharmacotherapy?

“All of these things happen in the hospital,” he continued. “It’s the clarity and the efficacy with which we have turned these issues into processes—into using all the interactions that we have around the hospital; using the information we have in the electronic health record; and what technology can do for us—that’s where we fall down and where we have the most room for improvement.”

Adopting a Team Approach to Healthcare

This is where collaboration or a team-based approach to patient care can improve outcomes, he suggested. “We let the clinician ask the question, informed by the laboratorian choosing the test that is best to answer the question,” Arnaout explained. “Then, the patient, the physician, and the pharmacist, can decide what’s best to do about it.

“We can move toward a team-oriented approach as other hospitals have done,” he said. “Such an approach is in place in a few vanguard institutions such as Vanderbilt University and the [University of] Texas Medical Branch. But this approach is not used widely. Some might say, the future is here but it’s unevenly distributed.”

Financial Incentives Not in Place to Support Team-based Healthcare

If team-based care is the best approach, why isn’t it more widespread? Arnaout answered this question with a response that is common to pathologists: the financial incentives do not support such an approach to care delivery.

“Imagine that you are the head of the pathology department or laboratory medicine division. In this role, you control budget and see that it costs money every time you run a laboratory test,” he explained. “Now we know from research that there are two tests that ought not to be performed and four additional tests that ought to be performed. So, if you correct the overuse and underuse in the average laboratory, then you are doing more testing and [you] put your lab into the red.

“But, what does that testing accomplish? By virtue of these tests being the right testing for the right patient at the right time, the definition of that testing is that it helps the patient. Put aside all manner of human kindness and think of this as a bean counter would think: That translates to less additional costs, meaning fewer lawsuits and shorter lengths of stay, which are measured in many thousands of dollars. And lab tests costs are measured in tens and maybe hundreds of dollars,” Arnaout said.

“After having right-sided all of the tests that shouldn’t be ordered, and all of those that should be ordered, you end up as the pathologist running a department in the red,” he said. “That’s bad for you because you get chewed out by the front office. At the same time the front office is praising other departments for demonstrating improved patient outcomes and being in the black.

“The front office doesn’t see that by improving test ordering, costs rose in one department, but costs went down more in many other departments,” he explained. “That’s one improvement we can make—to recognize that all budgets go together.

“That’s easier said than done,” Arnaout added. “Because to improve test ordering is to determine which [tests] needed to be ordered and not ordered. We already have 4,000 different commonly ordered tests, and 10,000 ICD-9 codes, and many more ICD-10 codes. It’s a lot of work to figure out what rules to put into clinical decision support tools to help clinicians figure out which tests to order and which ones not to order for every patient.

No Easy Answers to Improve Medical Laboratory Test Ordering

“Who will do that work? It has to be someone with medical training, meaning one of the MDs we’re already paying in the pathology department to review biopsies and run tests. How do we reimbursable that activity? Right now, there is no easy answer to that question and there is no easy way to bring all the budgets together either,” Arnaout concluded.
Related Information:

Medical Tests: Inaccuracies, Risks and the Public’s Health

Large-Scale Analysis Describes Inappropriate Lab Testing Throughout Medicine

Unnecessary Testing? Study Finds Inappropriate Laboratory Testing Throughout Medicine

Sin of Omissions: When Tests Fly Under the Radar

The Landscape of Inappropriate Laboratory Testing: A 15-Year Meta-Analysis