Mass General's EvalMpox Tool Helps Clinicians Spot Mpox Cases and Safely Rule Them Out

Knowing when isolation can end matters as much as knowing when it should begin.
Shenoy stressed that EvalMpox's high negative predictive value helps clinicians stop isolation and reassure patients.
Mark

So what does EvalMpox actually do? It sounds like a diagnostic test.

Mimi

It isn't a lab test. It sits inside the electronic medical record and prompts the clinician: exposure questions, symptom questions, and what to look for in the rash exam. Then it sorts the patient into likely or unlikely mpox.

Mark

Why build that for a disease at all? Don't doctors know what a rash looks like?

Mimi

Mpox is uncommon, so there's limited clinical experience. And catching it matters because communicable patients need to be isolated to reduce spread.

Luke

Let me slow down on the headline claim. It flagged "nearly all" cases in the system. Do we have a number?

Mimi

Not in this reporting. We have 668 visits and the phrase "nearly all," which comes from the senior author.

Luke

So that's the investigators' characterization, and the sensitivity and specificity figures aren't given here. Same for "high negative predictive value": it's stated, not quantified.

Mark

Is the negative side as important as the positive?

Mimi

Shenoy argued it is. If the tool says low risk, you can end isolation sooner and reassure the patient.

Luke

Worth noting that the tool was assessed in patients presenting with a rash. Nothing here says how it would work for anyone else.

Mark

And what's next?

Mimi

Similar tools for Ebola, Lassa fever and Marburg, plus routine infections that require isolation. Those are in development, not yet validated.

  • Patients walk in with rashes, and clinicians must decide quickly whether mpox is among the possibilities.
  • The disease is uncommon, so few doctors have the pattern recognition needed to spot it reliably.
  • EvalMpox, built into the medical record, walks clinicians through exposure questions and a targeted rash exam.
  • Across 668 visits it flagged nearly all mpox cases, and its low-risk calls were highly reliable in ruling infection out.
  • Researchers are now adapting the approach for Ebola, Lassa fever and Marburg, with the next test being whether those tools perform as well.

Across a wide web of clinics in and around Greater Boston, clinicians meeting patients with a rash often face a quiet puzzle: mpox is a possibility, yet it is rare enough that few have learned to recognize it well. Into that gap, researchers at Massachusetts General Hospital have placed EvalMpox, a guide built directly into the electronic record that asks the right questions and directs the right examination. Its tests across 668 patient visits suggest that when the tool signals danger, the danger is usually real, and when it signals calm, the calm is usually warranted. It is a modest instrument, but it reflects an older human aspiration: that careful attention, well organized, can help us know sooner who is ill and who is not.

In more than a hundred clinical locations spread across Greater Boston, Nantucket, Martha's Vineyard, western Massachusetts and southern New Hampshire, patients arrive with a rash and clinicians must consider mpox. It is a real possibility, yet an uncommon one, and many physicians have seen few cases. A team at Massachusetts General Hospital set out to narrow that gap.

The result is EvalMpox, a decision support system embedded in the electronic medical record. Instead of relying on memory, the clinician is guided through questions about possible exposures, symptoms, and a physical examination focused on the characteristic rash. The tool then sorts each patient into "likely to have mpox" or "unlikely to have mpox."

The researchers tested it across 668 patient visits and reported their findings in Infection Control and Healthcare Epidemiology. Patients flagged as likely were prioritized for testing and proved significantly more likely to test positive. The senior author, Erica Shenoy, said the program identified nearly all mpox cases diagnosed in the system during the study, showing how decision support can speed early identification and isolation, which reduces spread.

Equally important, Shenoy noted, is the tool's high negative predictive value. When EvalMpox judged a patient low risk, that patient was very unlikely to be infected. This reassurance helps clinicians decide when isolation can safely end and lets patients know they probably do not have an infection.

The team hopes the approach will extend further. Drawing on lessons from COVID-19 and the mpox outbreak, they are building similar tools for Ebola, Lassa fever and Marburg virus disease, and for routine infections that require patient isolation. Whether those tools match EvalMpox's performance is the next thing to watch.

Across more than 100 clinical locations in Greater Boston, on Nantucket and Martha's Vineyard, in western Massachusetts and in southern New Hampshire, clinicians saw patients who walked in with a rash. Mpox was one possible explanation, but it is an uncommon disease, and few clinicians have much experience spotting it. A group at Massachusetts General Hospital built a tool meant to close that gap.

The tool is called EvalMpox. It is a clinical decision support system embedded in the electronic medical record. Rather than leaving a clinician to remember what to look for, it walks them through it: questions about possible mpox exposures, questions about symptoms, and a physical examination aimed at the characteristic mpox rash. Once the answers are entered, EvalMpox sorts the patient into one of two groups, "likely to have mpox" or "unlikely to have mpox."

The researchers, from Mass General, a founding member of the Mass General Brigham system, tested the tool across 668 patient visits. Their findings appear in Infection Control and Healthcare Epidemiology. Jacob Lazarus, MD, PhD, the lead author, is a physician investigator in the Division of Infectious Diseases in Mass General's Department of Medicine and an instructor in Medicine at Harvard Medical School. He described the project as a decision support tool built into the record that guides clinicians on key physical exam features and the important questions to ask when mpox is possible.

The patients the tool judged likely to have mpox resembled, in their characteristics, the patients who were confirmed to have the disease. Those flagged patients were prioritized for testing, and they were significantly more likely to test positive than the patients judged unlikely to have it.

The senior author, Erica S. Shenoy, MD, PhD, is Chief of Infection Control for Mass General Brigham and an associate professor of Medicine at Harvard Medical School. She said the program identified nearly all of the mpox cases diagnosed in the system during the study period. In her view, that underscores what decision support can add to early identification, isolation and management of patients with communicable diseases such as mpox. Early isolation matters because it reduces spread.

The tool also showed a high negative predictive value. In practice, when EvalMpox called a patient low risk, that person was very unlikely to have the infection. Shenoy stressed that this half of the result matters too. Early identification and isolation to prevent transmission cannot be overemphasized, she said, but it is just as important to have efficient ways to decide when isolation can end, and to reassure patients that they probably do not have an infection.

The team sees the approach reaching beyond mpox. Lazarus said that, drawing on lessons from the COVID-19 pandemic and the mpox outbreak, they are developing decision support tools for other emerging infections, including Ebola, Lassa fever and Marburg virus disease. They are also working on tools for more routine infections that affect how care is delivered because they require patient isolation. The next thing to watch is whether those tools perform as EvalMpox did.

The program identified nearly all mpox cases diagnosed in the Mass General Brigham system during the study period.
— Erica S. Shenoy, MD, PhD, senior author and Chief of Infection Control, Mass General Brigham
The team is developing decision support tools for other emerging infections such as Ebola, Lassa fever and Marburg virus disease.
— Jacob Lazarus, MD, PhD, lead author
Fale Conosco FAQ