The (True) Value of Laboratory Medicine
Laboratory medicine is often misquoted as having a role in 70 percent of clinical decisions – but how can we measure the true value, and more importantly, how can we improve it?
Mike Hallworth |
At a Glance
- The common claim that laboratory medicine has a role in 70 percent of clinical decisions may not be as accurate as many believe
- The IFCC task force, which evaluates the evidence supporting laboratory medicine’s role in healthcare, has devised principles for establishing the value of individual tests
- Inaccurate results are cause for concern, but factors like not receiving test results on time can cause even more harm
- Outcome studies are needed to better evaluate the benefits of new and existing tests
As lab medicine professionals, we are fully aware of the unquestionable importance of our profession. In the UK alone, every citizen has an average of 14 tests per year performed by a laboratory medicine specialist (1). Department heads increasingly rate quality care and value-for-money as key priorities, so a recognition of the value of lab medicine is of crucial importance, especially when it comes to ensuring appropriate allocation of resources. But is laboratory medicine falling at the last hurdle when it comes to providing improved benefits for patients? The International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) task force on the impact of laboratory medicine on clinical management and outcomes was set up in 2012 to settle this very problem, to evaluate the evidence supporting the impact of laboratory medicine, and to promote contributions from the field (2).
The misleading 70 percent claim
No doubt you will have heard the frequently cited claim that laboratory medicine plays a role in 70 percent of clinical decisions. That assertion sounds plausible, but the data on which the claim is based represent unpublished studies and anecdotal observations, and cannot be objectively veriﬁed at this stage (3). So where did it come from?
The earliest reference to the claim can be found in a 1996 paper from the Mayo Clinic in the US, where the author stated, “We know that, although the laboratory represents a small percentage of medical center costs, it leverages 60–70 percent of all critical decisions, e.g. admission, discharge and therapy “ (4). But even that paper failed to provide evidence for its statement. In the 19 years since the paper was published, in true Chinese whispers style, the statement has been taken and extrapolated upon, from 70 percent of critical medical decisions to 70 percent of all medical decisions (3).
A similar claim that 70 percent of all electronic medical records consist of laboratory data is also a likely contributor to the confusion (3). But this figure is completely separate from the medical decisions claim, and – unlike that claim – is backed by published papers, although the precise percentage varies between articles. It’s important not to confuse the two statements, because the amount of data in the record is a poor proxy for the importance of that data in the care of an individual. Rather than paying attention to the volume of information we gather, we should be focusing on what that information means to the life and health of our patients. This is consistent with the global shift from volume to value in healthcare provision.
A number of major organizations have seemingly fallen into the “70 percent” trap, including the UK Department of Health. A report from the House of Commons Select Committee on Health in 2002 stated “up to 70 percent of all diagnoses in NHS patients depend on laboratory tests, hence NHS pathology services are critical for the day-to-day evidence-based care of patients.” Although the idea behind that statement is certainly true, the 70 percent claim was unsupported by evidence in that report, and appears increasingly unlikely when you consider mental health and all the minor diagnoses made in primary healthcare. I would hazard a guess that, even from such a reliable source, the claim probably represents an unintentional misquotation of the Mayo Clinic study.
Measuring and improving the value
Use of the various “70 percent claims” should be resisted in favor of more speciﬁc and evidence-based indices of added value that require a better understanding of the mechanisms by which value is added or reduced. But how do we measure these values, and more importantly, how do we improve them?
Outcomes, which are defined as the results of medical interventions in terms of health or costs, provide the only real measure of clinical impact but are often overlooked in favor of prognostic accuracy studies. These accuracy studies ask simple questions centered on the diagnostic sensitivity and specificity of a test: “Does this test predict an outcome of interest?” Outcome studies go the extra step, accepting that a test might predict an outcome, but going on to question whether or not the application of the test in practice will make a real difference to patients.
The model proposed by the IFCC task force for measuring the net clinical value of a test involves balancing the benefits that a test delivers against any harm it may cause. For the model to work and to increase the value of a test, it’s important to first accept that testing can sometimes cause harm. In general, that harm stems from one of five possible sources, originally described by Epner et al. (5):
- An inappropriate test may be ordered
- The appropriate test may not be ordered
- The appropriate test result may not be used properly
- The appropriate test result may be delayed or missed
- The appropriate test result may be wrong or inaccurate.
An incorrect result, the area that receives most of our attention, is also the area with the lowest cause of diagnostic error – primarily because we in the lab have spent so much time focusing on this aspect (2). Now’s the time we need to get serious about some of the other factors.
Clinicians need our help
It’s clear to everyone who works in laboratories that, unless we help them, clinicians use lab tests badly. No matter how good a lab test is, if it isn’t used properly, it will never contribute to improved outcomes. One study questioning family physicians in the United States found that physicians order tests in 30 percent of all patient encounters. In almost 15 percent of these cases, physicians admitted to not completely understanding what tests they were ordering. An additional 8 percent admitted to being confused by the results that came back (6). The sheer volume of tests available and the rapid rate of increase means that physicians cannot be expected to understand optimal testing strategies for all conditions. They need help out there.
The next area that we need to tackle is the fact that, when doctors get appropriate test results, they don’t always use them correctly. This could occur for a number of reasons; for instance, the recipient may simply not understand the significance of the test, or the results may be misleading, either generally or in specific circumstances. Falsely labeling normal results as abnormal can confuse physicians and lead to severe effects on patients. Such errors can result from something as simple as variation reference ranges between laboratories, and urgent action is required to improve reference range harmonization.
Perhaps the most avoidable, yet one of the least talked-about issues affecting the value of laboratory medicine is that of correct test results not reaching the right place at the right time. The Commonwealth Fund, an organization dealing with healthcare and healthcare inequality in the United States, examined a range of outcome indicators based on surveys of patients and physicians, and ranked individual countries according to quality of care. Safe care rankings were produced by a range of metrics, including both the frequency of inaccurate test results and issues with getting test results back on time. In every case except France, producing the right result but not delivering it to the right place on time caused twice as much reported harm as an incorrect result (Table 1) (7). It is relevant that, in France, patients are often custodians of their own results.
Lab personnel can work really hard to optimize a method and obtain a completely accurate result, but all of that work can be a complete waste of time if we can’t do the simple task of getting the result to where it is needed, when it is needed. How much effort do we put into getting the result right, and how much effort do we put into making sure that someone actually sees it in time and does something about it?
Asking the right questions
Until now, outcome studies for lab medicine have been infrequent because of the challenges involved in linking a diagnostic test to a clinical decision and its possible downstream effects. We’re very good at asking whether or not we can trust a test, and at asking whether or not it tells us something we want to know – but we need to go on to ask, ”Does this test help?” When developing new tests, we can significantly improve the value of laboratory medicine just by looking beyond how accurate the results are and asking whether the test helps us to make better or quicker diagnostic decisions, or to increase the overall effectiveness of the treatment.
To improve our evaluations of new lab tests, we need to put more emphasis on outcome studies. A good recent example is a paper from the European Group on Tumor Markers (8), wherein they outline tumor biomarker monitoring trials and how they are defined. The panel proposes a four-phase model for biomarker monitoring trials, similar to the one used for new drug investigations. The first and second phases involve characterizing a marker and evaluating its ability to provide a readout on disease status. The third phase looks at the effectiveness of the biomarker by monitoring patient outcomes in randomized trials, whereas the final and most important phase involves post-marketing surveillance to assess the validity of the new marker’s benefits. I think a model like that is excellent – but I don’t think that it should just apply to tumor markers. I think it could, and should, potentially be applied to any biomarker.
Progress in improving the five factors negatively affecting laboratory medicine is essential if we want to demonstrate and enhance the value of our specialty. To make that progress, we’ll need investment from governments and commissioning agencies, and we (and they) will need to place more emphasis on outcome studies. The long-awaited Institute of Medicine report on Improving Diagnosis in Health Care (released September 2015) (9) emphasizes that healthcare organizations should facilitate and support collaboration among pathologists, radiologists, other diagnosticians and treating healthcare professionals to improve diagnostic testing processes (Recommendation 1A of the report). That collaboration will require strong partners in the lab medicine community. That’s why, in the future, I hope that laboratory medicine professionals who choose to work in relevant areas will take a leading role in improving patient care – so that together, we can ensure that laboratory medicine is a true clinical specialty, rather than simply a number-generating service. The rewards are better patient care and more job satisfaction for those who work in laboratories – a real win-win situation!
- The NHS, “Pathology services explained” (2014). Available at: bit.ly/1nzep3W. Accessed September 10, 2015.
- M Hallworth, et al., “Current evidence and future perspectives on the effective practice of patient-centered laboratory medicine”, Clin Chem, 61, 589–599 (2015). PMID: 25646214.
- M Hallworth, “The ‘70 percent claim’: what is the evidence base?” Ann Clin Biochem, 48, 487–488 (2011). PMID: 22045648.
- LR Forsman, “Why is the laboratory an afterthought for managed care organizations”, Clin Chem, 42, 813–816 (1992). PMID: 8653920.
- P Epner et al., “When diagnostic testing leads to harm: a new outcomes-based approach for laboratory medicine”, BMJ Qual Saf; 22: ii6–ii10 (2013). PMID: 23955467.
- J Hickner, et al., “Primary care physicians' challenges in ordering clinical laboratory tests and interpreting results”, J Am Broad Fam Med, 27, 267–274 (2014). PMID: 24610189.
- K Davis, et al., “Mirror, mirror on the wall, how the performance of the U.S. health care system compares internationally”, The Commonwealth Fund, (2014). bit.ly/1i0iX1B
- G Sölétormos, et al., “Design of tumor biomarker-monitoring trails: a proposal by the European Group on Tumor Markers”, Clin Chem, 59, 52–59 (2013). PMID: 23034139.
- Institute of Medicine. Improving Diagnosis in Health Care. (September 2015). nas.edu/improvingdiagnosis