Skip to main content
Premium Trial:

Request an Annual Quote

Firms See Opportunity in Lab Test Utilization Business

NEW YORK – Laboratory testing is one of the most commonly performed medical activities with, according to the American Clinical Laboratory Association, some 7 billion tests ordered each year.

A substantial number of those tests are likely unnecessary, however. A 2013 meta-analysis published in Plos One, for instance, found that roughly 21 percent of lab test orders represented over-utilization, meaning they didn't add information that was useful to patient care. At the same time, the study put the test underutilization rate at almost 45 percent, indicating that in a substantial proportion of cases, potentially useful tests are never used.

As clinical testing is one of the main factors in physician decision-making, ensuring appropriate testing is key to patient health. Additionally, both overutilization and underutilization contribute to increased healthcare costs, making improving test utilization an area of growing importance and industry interest.

Some leading health systems like the Cleveland Clinic began implementing test utilization management systems nearly a decade ago. In recent years, a number of outside vendors have developed solutions that have begun making inroads into hospital and other labs.

One of the major players in the space is Madison, Wisconsin-based National Decision Support Company (NDSC), a Change Healthcare company, which currently has 11 customers live with its lab test utilization tool, called CareSelect Lab, and 20 customers signed for the system overall, according to Benjamin Gold, product manager at the firm.

The CareSelect utilization system, brought to market in the fall of 2017, integrates with healthcare provider's electronic health record systems and allows them to monitor the appropriateness of tests being ordered by its physicians. Appropriateness is based on guidelines developed by Mayo Clinic, which NDSC has partnered with to provide content for its system.

CareSelect was first launched in the imaging space, offering a utilization tool to help systems track and enforce the appropriateness of imaging studies being ordered. This, Gold noted, became a particular priority during the runup to and passage of the Protecting Access to Medicare Act, which mandates that physicians must consult appropriate use criteria and document this consultation in order to receive reimbursement for advanced imaging procedures.

"That accelerated the growth of the CareSelect business from 2014 to 2017," Gold said. Around 2015 and 2016, NDSC began exploring other areas within healthcare where it might help with utilization management.

"Lab and meds were the two areas where we heard most frequently from our customers that they were looking for support," he said. "We eventually found our way to Mayo Clinic and we basically recreated the structure that we had for [imaging] for diagnostic clinical labs."

In fact, Mayo was at that time a decade or more into an internal effort to track and manage lab test utilization, said Anne Wiktor, a medical policy manager at Mayo Clinic.

The effort was moving slowly due in part to the challenge of integrating with Mayo's electronic health record system.

"We wanted to move to a system that was much more scalable, especially as we were going to be moving to a new medical record system, EPIC," Wiktor said. "It seemed like a natural fit for Mayo Clinic to partner with an IT company rather than to try to build this type of system on our own."

NDSC partnered with Mayo to combine Mayo's medical expertise with NDSC's IT capabilities, said Wiktor, who is part of a team of six Mayo employees who work with the system's clinicians to develop guidelines for determining appropriate testing. These guidelines are reviewed and updated annually.

Currently, Mayo's utilization system consists of around 800 guidances or rules around test utilization, Wiktor said. The clinic is running the system in what she called "surveillance mode," meaning that it is not actually recommending usage patterns to ordering physicians but is simply collecting data on usage patterns to identify where improvements might be made.

"We want to make sure we are only targeting those areas or provider groups where we are identifying opportunities, rather than having these interventions or alerts and pop-ups across the board," she said. She added that Mayo planned to move beyond exclusively using the system for surveillance and turn on utilization alerts sometime next year.

Gold agreed with Wiktor that interventions need to be carefully tailored to avoid issues like alert fatigue.

"We want to be very prescriptive and the institution needs to buy in and say we have built the empirical case to try to change this physician's ordering decisions and we have the full institutional weight behind making this happen," he said.

Gold said that most of the CareSelect Lab customers are, like Mayo, still using the tool in surveillance mode to assess where the opportunities for better test utilization lie.

"People know, qualitatively and anecdotally, that specimens are flowing through their system and they know deep down that a lot of those [tests] never needed to be ordered," he said. "But they don't know which ones, they don't know who is causing the problems. They'll have some very specific examples, usually around high-cost send-outs, but they really have no idea among tests under $500 what is waste and where it is coming from."

Gold said that one of the company's customers began using the tool to intervene in physician test ordering this summer and that it will probably have data on the impact of those interventions towards the end of the month.

Another company in the test utilization space, Medical Database, Inc, recently published a study in the Journal of Clinical and Laboratory Medicine highlighting the potential for such tools to improve lab testing usage and to address certain lab billing challenges.

Like NDSC's CareSelect Lab, Medical Database's tool, which it calls its Laboratory Decision System (LDS), uses guidelines curated by pathologists and physicians to established what test is appropriate in what contexts. The LDS content can be integrated into providers' EMR and EHR systems as well as labs' LIS. Each test is scored by the system on a scale of one to 10, with a score of five or above meaning a test meets medical necessity.

In the JCLM study, the Irvine, California-based company partnered with the University of California, San Francisco department of laboratory medicine and Van Nuys, California-based Universal Diagnostic Laboratories to analyze how clinical lab ordering might be impacted by the LDS tool.

Looking at 96,170 lab orders comprising 374,423 test claims from a reference laboratory, the researchers found that 342,699 tests were scored by the LDS tool and that 48 percent of the total test claims were deemed appropriate by LDS while 44 percent were deemed inappropriate.

They also looked at 294,870 test claims "from a PPO provider managing self-pay insurers," and found that of the 259,840 tests covered by the LDS scoring regime, 52 percent were appropriate and 48 percent were inappropriate.

Safedin Beqaj, Medical Database's president and CEO, is a pathologist and laboratory director at Biocorp Clinical Lab. He said his idea for the company stemmed from his experience as a lab director.

"I would always get questions from physicians, what test do I order for this disease, or what does this test mean, what does this result mean," he said. "I thought, I should create something for them so they won't have to call" the lab.

Though Medical Database only launched its LDS tool this year, Beqaj said the company, which is self-funded, has been developing its guidelines content since 2009. He said he is targeted both hospital systems and individual physicians. Individual doctors can subscribe to the service for $150 a month, while subscriptions for hospital systems run in the range of $1,000 to $3,000 per month, Beqaj said.

He added that he has seen interest from labs themselves, as well.

Michael Fini is laboratory manager at Biodata Medical Laboratories, which uses the LDS. He said the lab has found the tool particularly useful for making sure ordering doctors use the correct  International Classification of Diseases (ICD)-10 codes, which he said is key to getting reimbursement from payors.

Not infrequently, doctors "don't give us the right [ICD-10] codes for the tests they are ordering," Fini said. "We submit them to the insurance company and the insurance company denies it because the test isn't covered by the code they gave us."

In those cases, the lab typically bills either the patient or the doctor for the unreimbursed test, Fini said, or the doctor is forced to spend time figuring out what ICD-10 code will work for getting the test reimbursed and resubmit it.

"This is a big problem, because the doctors are really busy," Fini said. "Not too long ago, I went into a physician's office and he had a stack probably two or three inches high of requisitions [from labs] that said, your codes don't cover these tests, so either give us a new code, or we'll bill you for the test ,or we'll bill your patient."

Fini said that before he began using the Medical Database LDS, his lab was getting denied for reimbursement on around 40 percent of all the tests it ran. He added that he didn't believe this was an unusual situation for a general laboratory that runs a wide range of tests.

Fini said that the LDS allows his lab to quickly look up ICD-10 codes when receiving a test order to make sure they are correct and will be reimbursed. Ultimately, he said he hoped to integrate the tool into the lab's billing system so it would confirm the appropriateness of the codes automatically, perhaps even implementing it on the provider end "so that when they order the test and they put in the code, it will show if it isn't going to get paid under that code and they can look at alternative codes."

While firms like NDSC and Medical Database rely on manually curated guidelines, machine learning-based approaches could offer another way to tackle the test utilization challenge. Last week, researchers from Stanford University published a study in JAMA that found that machine learning could identify low-value lab tests based on information including patient demographics, normality of the most recent test of interest, numbers of recent tests of interest, history of Charlson Comorbidity Index categories, specialty teams treating the patient, time since admission, time of the day and year of the test, and summary statistics of recent vitals and lab results.

Also last week, Quest Diagnostics announced it had developed in collaboration with bioinformatics firm Hc1 a test utilization optimization system using Hc1's machine learning technology.

Called Quest Lab Stewardship, the product is part of Quest's broader move into providing lab management services for hospitals and health systems.

The system uses guidelines developed by the ABIM Foundation's Choosing Wisely initiative to assess the appropriateness of test orders combined with Hc1's data analysis and integration capabilities to help hospital systems make more effective use of that appropriateness data.

Hc1's informatics expertise comes into play at several levels, said David Freeman, general manager, Information Ventures at Quest, including enabling the integration of multiple EMR interfaces, the application of machine learning approaches to identify potential problem areas within a system, and the ability to offer providers guidance in real-time.

"The platform has been designed to really leverage the investments that hospitals and health systems have made in their EMRs and their clinical databases," Freeman said, noting that the lab stewardship tool is able to "look at a physician who places an order and then analyze, what happened with this patient; when was the last time they had this test; what is the diagnosis; where is the order coming from," and then give recommendations based on the guidelines the system's clinical leadership has decided to implement.

"If, for example, a vitamin D test was ordered on a normal, healthy patient, a pop-up might come on the screen and the physician would see a link to a protocol that says vitamin D testing isn't recommended for normal, healthy patients," he said. "Or the institution could say, this is going to be a hard stop and it requires approval [for an override]."

"There's a lot of complexity to really getting your arms around the volume of activity" in lab testing, Freeman said. "But there is real power to be had by reconciling [test data] across different systems and locations and applying common sets of rules so you can drive out clinical variation."

Indianapolis-based Hc1 currently has its platform running at more than 1,000 healthcare and lab sites, said CEO Brad Bostic.

Bostic noted that one of the issues large health systems struggle with — and which Hc1's machine learning technology is able to address — is simply putting together an accurate record of what tests they are using.

"You'll look across a health system, and they can't even get their arms around what levels of utilization they're having on a given type of test because that test is called something different in multiple different areas of their health system," he said. "And if it takes you months or even sometimes a whole year just to try to organize what testing is happening where, it's too late by the time you figure it out."

"We've trained these machine learning models across billions and billions of tests so that we can actually organize that data in a live environment," he said.

The Quest Lab Stewardship is available now to the company's lab management customers, a number of which are currently working with Quest to implement the system, Freeman said.

Looking forward, Freeman said that while much of the initial emphasis has been on overtesting, he sees test underutilization as a potential area of focus for tools like the Lab Stewardship product.

"The data suggest that underutilization is two times larger than overutilization," he noted. "If you think about giving health systems the tools to identify what are the right testing cascades and guidelines that would support patients getting diagnoses faster and less expensively … and what the downstream things are that you could avoid, whether it's medication or other kinds of interventions — we have a lot of interest in that piece of this."