December 2009: Volume 35, Number 12
Interpretive Comments on Tests Results
How Far Should Labs Go?
By Bill Malone
With today’s focus on patient safety, diagnostic errors have been put in the spotlight. Whether the result of systemic mistakes by an organization or those of individual clinicians, some reports have noted clinicians’ failure to correctly interpret diagnostic tests, pushing the problem uncomfortably close to the walls of the lab. One explanation put forward for missed diagnoses is the sheer volume of information clinicians need to process in order to do their jobs. With PubMed adding about 670,000 new entries per year to its cache of some 18,782,970 citations, clinicians face a huge challenge to stay on top of new treatments, trends, and standards, let alone simultaneously keep an eye on developments in lab medicine.
In order to help clinicians understand test results, some labs have turned to including interpretive comments on their lab reports. Although information technology has made it easier to add interpretive comments, laboratorians have debated exactly when, how much, and what kind of commenting really works. Those labs experimenting with interpretive comments say they tread a fine line between helping physicians assimilate information from the lab without contributing to the information overload so pervasive in medicine. In fact, a recent survey suggests that physicians do want and depend on interpretive comments: general practitioners in the U.K. reported that in at least 75% of cases, the lab’s interpretation influenced patient management.
As laboratorians try to figure out how best to exploit their expertise in test interpretation, they must also find the right balance when it comes to helping clinicians, said Anand Dighe, MD, PhD, director of the core laboratory at Massachusetts General Hospital, and an assistant professor of pathology at Harvard Medical School who has studied how clinicians use interpretations from his lab. “There is a very narrow view of lab medicine among clinicians. Now that is being broken down. Lab medicine doesn’t just happen within the walls of the laboratory. In fact, it’s much broader than that,” Dighe said. “Our job description is not just to turn out ten million test results per year. Our job is to help clinicians order and interpret tests. The test result is just the starting point.”
What’s In a Comment?
Interpretive comments on lab reports span a wide range, from a basic decision limit, to delta checks for serial results, to a definitive clinical diagnosis, noted Mario Plebani, MD, professor of clinical biochemistry and clinical molecular biology at the University of Padova School of Medicine and chief of the department of laboratory medicine at the University-Hospital of Padova, Italy. As a general definition, Plebani suggests that interpretive comments include “any additional information on the lab report that may help a clinician to better interpret information from the lab.” This spectrum contains everything from a canned single-sentence comment attached automatically to every result of a particular test, to individualized paragraphs that describe the clinical situation in a detailed, narrative style, as in anatomic pathology.
Dighe and his colleagues at Massachusetts General Hospital found that the lab’s interpretive comments on coagulation results were particularly successful. They surveyed physicians when the lab first included the interpretive comments. The response was resoundingly positive: physicians said the interpretations had prevented a misdiagnosis in 71% of the cases. “Anecdotally, physicians told us that they wouldn’t even look at the results of the test until the interpretation was back because it just wasn’t worth their time without it,” Dighe explained. “With coagulation tests, there are a lot of causes of false positives and false negatives, and it was much better to have a lab expert examine the results before the physician did.”
Under the Massachusetts General Hospital program, specimens arrive in the morning or the day before, and the lab performs all the tests during the day. Reflex testing is also performed if something is abnormal, allowing the lab to produce an answer by the end of the day. Special software packages gather all the results, and pathology residents work with the attending clinician who is in charge that week. They go into the software with all the data in front of them, write an interpretive paragraph, and put cases in pre-sign status. Finally, at the end of the day, residents sit down with the attending clinician, review the cases, and the attending clinician officially signs them out. With this system improving the efficiency of information flow, the Massachusetts General Hospital coagulation lab routinely performs 50 complex interpretations per day.
Using interpretive rounds for coagulation results inspired Dighe and his colleagues to work on similar systems for toxicology, hemoglobin, blood transfusion, protein electrophoresis, molecular diagnostics, and autoimmune disorders. The approach has been very well received by clinicians, but also underscores one of the boundaries laboratorians face in beefing up interpretations: only an MD can bill his or her time for such an in-depth service under current Medicare rules.
Outside the U.S., test interpretations are not billed separately but considered a routine part of the service a lab provides. For example, in Plebani’s lab in Italy, both MDs and PhDs record interpretive comments according to each person’s specific competence and training. “In particular, in the fields of hematology, coagulation and autoimmunology, MDs are very involved, while in the areas of protein and specialized clinical chemistry, PhDs are more active,” explained Plebani. “However, the main difference is the individual level of competence and responsibility achieved during and after post-graduate courses. We consider this activity a fundamental job of laboratory professionals and that time for interpretive commenting has to be considered when discussing the number and qualifications of laboratory staff with administrators.” His lab does not receive additional fees for interpretations, but the reimbursement is considered a part of the whole laboratory service.
Similarly, in the U.K., interpretations are “taken as a given,” said Danielle Freedman, MB, ChB, consultant chemical pathologist and associate physician in clinical endocrinology at Luton and Dunstable Hospital NHS Foundation Trust in Luton. “This is just part of the service we offer. We have clinical scientists who also give advice, which I know is a contentious issue in the states. We may be different here, but in the U.K, we see this as a high priority of our job. There is more and more testing going on, and it’s more complex, so in particular you’re not going to expect primary care physicians to always understand the clinical utilization of those tests.” Freedman also noted that it seems new doctors coming out of medical school know less and less pathophysiology and less about lab testing.
Despite the hurdles inherent in the U.S. payment system, it’s still no excuse for laboratorians not to play a role in interpretation, said Dighe. “We get the lab staff involved and they have input because they have a lot of knowledge,” he said. “I think they work best when it’s the technologist working with the pathologist figuring out what a result really means and what the next steps are for the patient. But how to compensate them for that time is challenging. It’s really extra time. But I think you just call that part of the cost of doing business and providing a service to your customer.”
To Comment or Not to Comment?
That is the Question
Should every lab result on every report carry an interpretive comment? Probably not, say experts who have studied this issue. Physicians have little time to comb through comments that do not add value to the lab report, and in many situations the lab does not have enough patient data to provide a true interpretation.
However, for some testing areas, particularly new tests or complex panels of tests, physicians say a lab’s interpretation makes a big difference in patient care, said Mario Plebani, MD, professor of clinical biochemistry and clinical molecular biology at the University of Padova School of Medicine and chief of the department of laboratory medicine at the University-Hospital of Padova, Italy. “The extensive development of specialized sectors in clinical laboratories and the correlated increase in the number of tests and their complexity have highlighted the difficulties in data interpretation encountered by general practitioners and physicians receiving laboratory tests, particularly outside their own specialty area,” he said. “The autoimmunology laboratory represents a prototype for interpretive comments. It has undergone considerable growth in recent years thanks to new discoveries in physiopathology, target antigens, and diagnostic tests.” Plebani also noted that the literature on the subject shows that physicians have difficulties interpreting new cardiac makers, such as cardiac troponins, hs-CRP, and in vitro allergy tests with recombinant antigens.
Walking the Line
Beyond economics, laboratorians face a frustrating dilemma even with basic comments that the laboratory information system (LIS) includes automatically. The fear is always that even if backed by thoughtful editing and literature references, interpretive comments can become just so much background noise to clinicians as they slog through myriads of other details throughout the day, said Corinne Fantz, PhD, co-director of the core laboratory at Emory Crawford Long Hospital and assistant professor of pathology and laboratory medicine at Emory University School of Medicine in Atlanta. “Sometimes when we talk to physicians they don’t even see the comments or know that they’re there,” she said.
Ironically, part of the predicament lies with the electronic systems that made ubiquitous automatic interpretive comments possible in the first place. For example, in Emory’s LIS, abnormal results are bolded on the electronic report the clinician sees. Interpretive comments reside in footnotes, called out by an asterisk next to the result. Due to the nature of the system, the clinician then has to click through two more pages before seeing the added information. Unfortunately, some clinicians don’t get that far.
Even when getting to an interpretive comment means just a few more mouse clicks, when it becomes routine, clinicians often see it as a hassle or waste of time, explained Jay Jones, PhD, director of the chemistry and toxicology laboratories at Geisinger Medical Laboratories in Danville, Pa. “We’ve found that here with our electronic health record, since we use it so extensively, physicians don’t want a lot of extra text, they just want a quick answer,” he said. “They don’t like reminders or pop up boxes, they basically say, ‘get them out of my face.’”
In fact, Jones and his colleagues found that clinicians don’t want to go beyond about seven mouse clicks to produce an order. “They are very sensitive to that because it’s repetitive clicks, especially in primary care, where they see the same patients in the same sequence, and if they can find a shortcut, if they can do an encounter in five clicks instead of seven clicks, they see that as a big gain,” Jones explained. “It really becomes automatic after a while. They don’t like slow screens and additional information that comes up that they can blow right by anyway.”
The secret ingredient to making sure clinicians read interpretive comments is in the quality and utility of the information, according to Dighe. “If it’s useful content physicians will find it and look at it,” he said. “If you save them time, if you help them not make a mistake based on the result, then they’ll look at it.” Canned comments have their place, he emphasized. For example, a comment appended to every tumor marker result can help make sure clinicians understand the limitations of the lab’s method. While some of the oncologists who see these results every day might not look at it, for those who are unfamiliar with the false positives and false negatives, interpretation is essential. “For some tests, the fact that 90% of clinicians don’t look at a comment doesn’t mean it’s not valuable. There is a core of clinicians who would be lost with that result without the comment,” he added.
At the same time, Dighe stresses that there is a difference between making sure that doctors unfamiliar with a test don’t struggle to interpret the results and adding comments that will not likely be useful to anyone. “Just a comment appended to every hepatitis result that explains what the difference is between IgM and IgA in hepatitis A—clinicians aren’t going to look at that. It’s got to be something they either don’t know or aren’t confident in to really add value.”
Trying too hard can also cause problems. “If we don’t know exactly why a physician is sending the test, we are relatively vague in the interpretation,” Dighe said. “I think that’s the art of doing a good interpretation. If you have all the information, give a great interpretation. If you don’t, back off and give them just enough to help guide the next set of tests or convey that level of uncertainty.”
To go beyond the barest of basic comments, labs have found that reflex testing and enhanced interpretations go hand-in-hand. Reports have already demonstrated that reflex testing can both reduce costs and lead to better outcomes for labs, noted Plebani. The aim should be to substitute many individual tests with a clinical question. Reflex testing has become much more standard practice in the U.K. as well, according to Freedman. For example, on a total protein that comes in at the upper limits of normal, Freedman’s lab would automatically perform a serum electrophoresis to look for multiple myeloma. “Really, the ideal scenario would be for us to decide which test to perform,” she said. “We have the expertise and should select, based on evidence-based medicine, what the best tests are for each condition.”
Due to the complexity of any given case, reflex testing is really essential to offer true interpretations, Dighe emphasized. “It’s hard to do real interpretations on single values. You really have to have the reflex algorithms in place that let you get to the bottom of why a value is elevated or low. And those require going to your medical policy committee and telling them that whenever a patient has a certain abnormality that the lab needs to do certain tests in order to give a more complete picture of a patient’s status. Even if you’re not going to do an interpretation, it’s the right thing to do in many cases,” he said. At his institution, Dighe said he’s never had a problem with clinicians accepting a reflex algorithm, as long as it’s something that saves them time and adds value. “They’re happy to have the decision out of the clinician’s hands, so if there is an abnormality on a serum protein electrophoresis, automatically free light chains get ordered on the same specimen,” he said. “The patient doesn’t have to come back in, we don’t need another specimen, and the workup gets done. We’ve never had any pushback on that.”
A Collaborative Effort
Crucial in making any level of interpretation work is getting outside of the lab and collaborating with clinicians. This is the best way to ensure that comments get read and make a difference, stressed Plebani. “Because a full knowledge of, and expertise in, all aspects of lab testing cannot be realized by any individual, the interpretation of results is the paradigm of a collaborative activity with inputs both from clinicians and laboratory professionals,” he said. “In addition, because of the increasing complexity of laboratory tests, even a single laboratory professional cannot achieve expertise in all fields of the discipline. Therefore, interpretive comments should be conceived as a collective responsibility of an individual laboratory service that should be managed through a common policy. However, lab professionals who have been trained and have achieved specialized knowledge in a particular field of activity should participate in developing comments related to their expertise.”
Even when crafting basic comments for individual tests, it’s important to consult with clinicians with expertise in that particular area, said Fantz. “We get the physicians involved when we’re making those comments for their respective services. So for example, if we’re making a coagulation comment, we’ll refer to hematology and get their input, to make sure that everyone who knows and understands the literature is comfortable with what’s being said.” Similarly, if a new guideline comes out on an area of testing, Fantz will discuss the issue with physicians in that specialty and ask what should be included in the interpretive comment.
“What happens is that if you have too much written in the chart, it can be a lot of information if the test is ordered frequently on somebody that’s staying in the hospital for a long time, and if they print this out, there is a lot of that comment that they have to scroll through,” said Fantz.
For Freedman in the U.K., working closely with clinicians is what lab medicine is all about. “I think some people don’t understand what lab medicine is supposed to be. It’s not just about analyzing a sample—it’s about advising the clinician, whether it be a primary care physician or a hospital clinician, both prior to an investigation and after it’s been analyzed with an interpretation. So beforehand there needs to be a dialogue with the clinicians, or in regard to written protocols, what investigations are appropriate in certain clinical condition.”