Tuberculosis remains a huge global problem. Diagnostic tests are imperfect, and it can be difficult for clinicians to weigh up their pros and cons to decide which to use. A PLOS Medicine article aims to simplify the decision by comparing the clinical impact of the standard Xpert diagnostic assay and the new Xpert Ultra assay in differing clinical contexts.
The study has just been awarded the 2018 PLOS Tuberculosis Channel Prize, so I interviewed author Emily Kendall, assistant professor in infectious diseases at Johns Hopkins University, to find out more about her work.
What drew you to study infectious disease?
EK: I always loved to solve problems and absorb new information, so science of some sort was an obvious fit. University courses about the global challenges of HIV and tuberculosis sparked an interest that ultimately led me to change my career plans from research in theoretical physics to training as a medical doctor.
What is tuberculosis (TB), and what effect does the disease have on sufferers?
EK: TB is an infectious disease caused by a type of bacteria called Mycobacterium tuberculosis. Symptoms vary widely, but most commonly include a progressive pneumonia-like illness with cough and inflammatory symptoms such as fever and night sweats. Unusually, people can be infected by the causative mycobacteria without any symptoms, and then can develop TB years or decades later.
What is the global impact of TB?
EK: Worldwide, about 10 million people develop TB each year, and over 1 million people die of TB each year – making it the world’s most common infectious cause of death. TB also has a large impact on the quality of life of those patients who survive.
Diagnosis of TB remains imperfect, and in your study, you examined the potential impact of switching from the standard Xpert MTB/RIF diagnostic assay to the newer Xpert Ultra assay when diagnosing pulmonary TB in adults. How did you go about this?
EK: Yes, we wanted to understand the clinical impact of switching to a newer version of this assay that does a better job of finding people who have TB but also produces more false-positive results. We developed a computer model to simulate realistic patient populations being evaluated for TB in particular clinical settings. The model predicted how each assay performed in each clinical context, and also extrapolated from positive and negative test results to meaningful clinical outcomes such as dying from TB or receiving unnecessary treatment.
What did you find?
EK: One measure we used is the number of additional people unnecessarily treated for TB as a result of Ultra (false positives) for each additional TB death that Ultra was predicted to prevent. Measured in this way, we found that the favorability of a switch from standard Xpert to Ultra would vary widely between settings. In an HIV clinic in sub-Saharan Africa, the switch is likely to add fewer than 10 unnecessary treatments per death prevented, making it favorable. However, in a context with less TB and lower rates of TB-associated mortality, such as a Chinese primary care practice, the switch could add more than 300 unnecessary treatments per death prevented.
What most interested you about your findings?
EK: While the Ultra cartridge is likely to have important clinical benefits (including reducing TB mortality substantially in settings where it is currently very high), it presents tough decisions in other contexts that ultimately come down to value judgments. Knowing that TB treatment has side effects, monetary costs, and logistical and psychological burdens for patients, how many people is it justifiable to diagnose falsely and treat unnecessarily to prevent one person from dying of TB?
Why do you believe this research is so important?
EK: As a clinician, I know that all diagnostic tests are imperfect and require thoughtful interpretation. This can be difficult, particularly for a new test. Hopefully, our model will help clinicians and health systems to evaluate the role of Xpert Ultra in clinical decision-making.
Given that you have just been awarded the 2018 PLOS Tuberculosis Channel Prize, it seems that lots of people believe in the importance of your work! How did you feel when you found out that you’d won?
EK: I was honored! The goal of this kind of modeling analysis is to help clarify thinking about real-world questions, so it’s great to hear that people are reading it and finding it useful.
What do you plan to do with your $500 prize money?
EK: I’d love to get my coauthors together in one place for dinner… or I might use it as seed money for our next research project!
Research Article: Kendall EA, Schumacher SG, Denkinger CM, Dowdy DW (2017) Estimated clinical impact of the Xpert MTB/RIF Ultra cartridge for diagnosis of pulmonary tuberculosis: A modeling study. PLoS Med 14(12): e1002472. https://doi.org/10.1371/journal.pmed.1002472
Images Credits: torange.biz, CC-BY; Marcus Westberg