For infectious disease clinicians, identifying what pathogen is making a patient sick is critical to determining appropriate treatment and care. Without input from clinical laboratories, they can be left in the dark.
Yet even as innovations in laboratory medicine revolutionize infectious disease diagnostics, the advance of molecular testing often has moved faster than the traditional pace of regulators, insurers, and other payers. These stakeholders generally expect to see obvious, linear relationships between tests, diagnoses, treatments, and outcomes—including evidence typical of pharmaceutical interventions such as randomized controlled trials. According to the National Cancer Institute, clinical utility refers to the likelihood that a test will, by prompting an intervention, result in an improved health outcome. The clinical utility of a genetic test is based on the health benefits related to the interventions offered to individuals with positive test results.
As the science of molecular testing advances, experts see an even greater need for laboratorians and clinicians to help the broader medical community understand and appreciate the increasing sophistication, power, and value these tests offer to improve patient care.
We spoke to Angela Caliendo, MD, PhD, Warren Alpert Foundation Professor of Medicine at Brown University’s Alpert Medical School in Providence, Rhode Island, about how infectious disease physicians perceive clinical utility and the evolution of evidence-based medicine in the molecular field. A former secretary of the Infectious Diseases Society of America, Caliendo also has chaired the society’s Diagnostics Task Force.
How would you describe the difference between analytic validity, clinical validity, and clinical utility for molecular testing?
Analytic validity is how the test performs in the laboratory out of clinical context. What’s the limit of detection, the linear range, or the reproducibility? These are nuances that most clinicians don’t spend a lot of time thinking about. They just assume that tests are valid, and that the results that come out of the laboratory are reliable. That doesn’t mean that with proper education and understanding they won’t learn or come to appreciate those nuances, but it’s not intuitive. COVID-19 brought these nuances to the forefront.
Clinical validity is how the test works in patient populations. A test may have a very low limit of detection, but is it low enough to be useful clinically? Or is it too sensitive? Does it differentiate people who have infection from people who don’t have an infection, or people who are at risk of severe disease from people who aren’t?
For example, transplant physicians and infectious disease physicians had already come to understand that not all cytomegalovirus (CMV) viral load tests have the same limit of detection. The important clinical question is not the limit of detection, but rather what viral load value predicts that the patient has a high risk of disease and should be treated with antiviral therapy.
How has the approach to clinical utility for molecular infectious disease tests evolved along with the science?
We’ve become more sophisticated in how we interpret test results clinically. As we began to understand that not all quantitative molecular assays give the same value, we’ve also grown to appreciate the importance of specimen type, the specific gene target, and the method used to extract nucleic acids from a clinical specimen. All these components of a test will affect its performance. It all impacts how the test is interpreted.
For example, do you treat CMV until patients are negative by a PCR test? Early CMV viral load tests had a limit of detection of 400 to 500 copies/mL of plasma, now we have assays that can quantify to 20 or 40 copies/mL. If a very low level of virus persists with no symptoms, do you need to treat until the result is below the limit of quantification? These types of test characteristics impact the clinical utility of the assay.
More generally, when I see persistent nucleic acid after someone’s symptoms have resolved, what do I do with that information? Is that pathogen still viable or not? There’s no easy way to determine that based just on a molecular assay. Again, these are nuances that, prior to COVID-19, people didn’t spend a whole lot of time thinking about.
Overall, we’re asking more sophisticated questions—how these more sensitive assays make a material impact on patient outcomes. COVID-19 really elevated people’s thinking about the clinical utility of molecular tests.
What is the infectious disease physician’s role in establishing clinical utility and working with payers, patients, and physicians given that a laboratory result is never used in isolation?
We should be bringing clinicians and laboratorians together so laboratorians can understand the clinical questions we’re trying to answer, and what test characteristics are needed to answer those questions. Clinicians can then begin to understand how tests perform and what factors impact the clinical utility of the test.
Infectious disease physicians also play a very important role in helping payers understand medical necessity. We all know there are challenges around getting some of these molecular infectious disease panels covered because the clinical utility is not immediately apparent.
In some situations, these tests provide clinically important information, while in other situations, they do not. For example, for a healthy adult who presents with diarrheal disease, letting the infection run its course for 3 to 5 days before doing any testing may be completely reasonable. But that same plan of action is not completely reasonable in someone who is immunocompromised.
There’s a sophisticated algorithm for when more expensive molecular panels should be used, but there isn’t an equivalent level of sophistication around reimbursement. Until clinicians and laboratorians come together and help provide that information, it’s easy for payers to refuse to pay for tests that do have clinical utility.
Some payers look for large or randomized studies to show utility, but can you explain why certain types of clinical studies are not always performed? How do you reconcile that?
Large randomized clinical trials are not always done for a few reasons. For example, with CMV viral load assays, we have Food and Drug Administration approved tests, but we’ve never done a clinical study to answer the seemingly simple questions of what viral load correlates with development of disease, or what viral load must be exceeded before someone needs to be treated? We don’t know the answers to those because tests are now widely available, they are used routinely in clinical practice, and the use of prophylactic and preemptive therapy has greatly reduced the occurrence of CMV disease. As a result, these clinical studies are not always practical to conduct and are very expensive.
What do you think infectious disease physicians need to be doing right now to establish clinical utility and value for testing services, and how can they work with lab medicine colleagues?
Infectious disease clinicians need to work collaboratively with their laboratories to identify the important clinical questions that need to be answered, and work together to design meaningful studies. That’s almost the easy part. Finding an agency or company to fund those studies and answer those important questions is much more challenging.
For all we’ve been through with COVID-19—all the testing and all the advances in testing—what’s the answer to the following question: What SARS-CoV-2 viral load correlates with infectivity? We’ve never done the study. These types of studies could rely on culture as a measure of infectivity, but we know culture lacks sensitivity. Alternatively, detailed disease transmission and epidemiologic data could be used but this is difficult to do in the middle of a pandemic.
Molecular tests for CMV and SARS-CoV-2 have just come into clinical use in recent history because they’re so much better than traditional culture or antigen testing. We haven’t stopped along the way to step back and ask that important question about the meaning of viral load.
One infection we did do that with was HIV. Researchers were able to show that HIV viral load was a remarkable predictor of time to development of AIDS. But generally, tests have become standard of care before the clinical studies are performed, because they’re so much better than traditional tests. We’re almost playing catch-up with some of the more basic clinical questions.
Jen A. Miller is a freelance journalist who lives in Audubon, New Jersey. +Twitter: @byJenAMiller