It's no secret that doctors sometimes find themselves befuddled by a laboratory test menu of at least 3,500 tests, double the number available 20 years ago, and one that is riddled with confusing nomenclature.
New research, however, has pinpointed the bottlenecks that an important subset of providers—primary care physicians (PCPs)—face in ordering and interpreting tests, as well as tangible opportunities for labs to help PCPs use tests better.
The insights come from a survey conducted by the Centers for Disease Control and Prevention's (CDC) Clinical Laboratory Integration into Healthcare Collaborative (CLIHC), an ambitious project of the agency's Division of Laboratory Science and Standards aimed at filling gaps that impede practicing clinicians from effectively utilizing lab services for better patient care. The survey was the last in a series of CLIHC initiatives to gather data about diagnostic test utilization. The survey questionnaire, which grew out of three focus groups CLIHC held, was sent in late 2011 to 31,689 PCPs, of whom 1,768 (5.6%) responded (J Am Board Fam Med 2014;27:268–74).
Off the top, participants reported ordering diagnostic testing in 31% of patient encounters per week, with uncertainty about ordering and interpreting tests 14.7% and 8.3% of the time, respectively. These figures perhaps don't sound too dramatic until one puts them in the context of total PCP visits in the U.S.—more than 300 million—which means inappropriately ordered or interpreted tests potentially affect some 23 million patients per year, according to lead author of the survey analysis, John Hickner, MD, MSc, department head and professor of clinical family medicine at University of Illinois at Chicago Medical School.
"If we think in terms of LEAN and quality improvement, even a one percent error rate is too high, and here we have, perhaps, up to 20 percent of cases with uncertainty," he said. "This is a very big problem."
Conventional Wisdom and More
In many ways the survey confirmed conventional wisdom about the diagnostic process: that the kaleidoscope of test names trips up physicians, and that they are frustrated by slow turnaround times and lack of access to prior results. PCPs also reported as problematic lab-to-lab variations in normal ranges for tests, and lab-to-lab variations in report formats.
More surprisingly, the top factors respondents found challenging in ordering and using tests all involved costs—how much tests would cost patients, whether insurers mandate specific labs to perform tests, as well as lack of comparative cost information. "That physicians think about test price so much was not in our expectations. That was a surprise," explained co-author Paul Epner, MBA, MEd, a consultant interested in strengthening the link between lab services and patient outcomes. "They're worried about patients' ability to afford tests and they're doing mental gymnastics around the issues of what the value is of a test and whether knowing the results will be worth the cost to the patient."
The survey also shed light on how PCPs deal with uncertainty in ordering lab tests. Nearly three-quarters reporting considered curbside consults, e-references, and referrals to specialists helpful in solving diagnostic ordering dilemmas. Of these, reviewing e-references was by far the most used of any tactic, with a full 57% of respondents doing so daily or at least once per week. Asking a lab professional about any uncertainty in ordering tests was rated highly, but hardly any participants reported doing so, a mere 6%—the lowest of eight possible tactics.
PCPs' strategies for dealing with uncertainty in interpreting lab test results mirrored these responses. Physicians found it most helpful to follow-up with patients, review patients' histories, and review e-references. Once again, consulting with a lab professional, although helpful when done, was the tactic physicians were least likely to use.
Hickner reflected on the gap between how helpful PCPs found lab consults and how infrequently they reported using them. He recounted his own recent experience in calling a lab to confirm the type of swab needed for a culture he was about to order. "It took several minutes before I got somebody in microbiology who knew what they were talking about to give me the answer," he recalled, emphasizing that doctors have little time to spare pursuing guidance.
Co-author Brian Jackson, MD, MS, agreed that labs can be inefficient in responding to physician calls, but suggested that limited use of this tactic might also stem from lack of knowledge. "I bet every lab in the country has a call line. The fact that only 24 percent of doctors have knowledge of it, that's troublesome," he contended. "I think we haven't actively made it as easy to contact us as possible." Jackson is vice president and chief medical informatics officer at ARUP Laboratories in Salt Lake City and a member of CLN's Patient Safety Focus editorial board.
IT to the Rescue?
Survey respondents thought information technology (IT) and systems-type solutions like reflex testing, trending, interpretive comments, and computerized physician order entry with electronic suggestions were most likely o help them. The study co-authors agreed that these are the most fruitful strategies, but all predicted that it will take some time for these strategies to be widely disseminated.
Still, some IT solutions within labs' control could be implemented in the near term. Jackson suggested that labs look at their online test menus. "Have we set those up to make it really easy to select available tests and understand the methods, the details to know which tests to order?"
Epner emphasized the value of providing trend information to PCPs and to specialists. "Reporting the ordering patterns of, say eight cardiologists in a shared practice will lead to normalization of divergent patterns. For example, if almost no one orders [creatine kinase-MB] tests, except one person, who orders as many CKMBs as troponins, this shared information tends to reduce inappropriate ordering."
Outside the electronic realm, Jackson suggests labs could do a better job not only communicating with physicians but also monitoring those communications. "What's the experience of doctors calling the labs? Are we actively managing how fast we respond to those calls, how fast we get them through to an expert?" he asked, adding that labs naturally have focused on test analytics, and perhaps not looked at how well they communicate with physicians.
A separate CLIHC workgroup assessed the thorny issue of test nomenclature. Labs, the group suggested "have filled their test menus…with bewildering nomenclature and abbreviations and have failed to appreciate the dangers of assigning perilously similar names to different tests." The authors of this report called for innovative software solutions to "help rather than hinder physicians" (J Gen Intern Med 2012;28:453–8).
CLIHC was expected to meet in July to consider potential interventions for solving the test ordering and interpreting challenge. "It's a pretty complicated, muddy world out there in clinical practice, and the major errors in the testing cycle are not so much in the analytics phase," said Hickner. "There ought to be quick mechanisms to reduce or eliminate that uncertainty."