Do you know what a nomogram is? They might just help your doctor predict how severe your COVID-19 coronavirus symptoms might become and whether you will end up hospitalized. Researchers at the Cleveland Clinic have believe they’ve come up with a game-changer in treating the virus.
Prediction models have been used pretty broadly during the spread of COVID-19. Remember the initial death prediction models that said over a million people would die from the virus in the US alone? When they were wrong by a wide margin, we were told it was because we had socially distanced and shut down the economy.
Here’s something you may not know. Social distancing, shutdowns and other mitigation methods were “baked into” the original models. Yes, they were that wrong. To be fair, predicting death rates across a nation for a virus we knew little about is tough sledding.
Some models, however, can be used with confidence in these situations. The nomogram developed and validated by the Cleveland Clinic researchers is likely to be that kind of tool. Their model shows promise in predicting which recently positive patients are at the greatest risk of being hospitalized and/or needing aggressive treatment.
This isn’t the teams first nomogram rodeo, so to speak. The team is spearheaded by Michael Kattan, Ph.D., chair of Lerner Research Institute’s Department of Quantitative Health Sciences and Lara Jehi, M.D., chief research information officer at Cleveland Clinic. Their first successful nomogram predicted the likelihood of a positive virus test result for individual patients.
“Ultimately, we want to create a suite of tools that physicians can use to help inform personalized care and resource allocation at many time points throughout a patient’s experience with COVID-19,” said Dr. Jehi, corresponding author on the study.
This particular predictive model used patient data from more than 4,500 patients who had tested positive for COVID-19 for its development and validation. Those patients had been tested at Cleveland Clinic locations in Florida and Northeast Ohio during the period spanning early March to early June. The risk prediction model was created by transforming data from the patients’ electronic medical records via statistical algorithms.
By comparing behavioral and health attributes between patients who avoided hospitalization and those who were hospitalized with COVID-19, the team uncovered some previously unrecognized risk factors.
Race was one of the key factors noted in the model. Black and African-American patients were more likely to become hospitalized than patients of other races.
Medications mattered, too. Those patients already on Angiotestin Converting Enzyme (ACE) inhibitors or angiotestin II type-I receptor blockers (ARBs) were also more likely to require hospitalization.
In what should be no surprise at all, smoking was a factor as well. Smokers ended up in the hospital at a higher rate than non-smokers.
Dr. Kattan, who is an expert in the development and validation of nomograms for medical decision-making believes more studies will be needed to further define the connection between ACE inhibitors and ARBs.
“In our study, taking these drugs was only found to confer increased risk for hospitalization when run through univariable analysis, which means the observed association could be the result of other, confounding variables, like a preexisting condition.”
Certain symptom combinations were also connected to higher rates of hospitalization. Patients presenting with fever and shortness of breath, along with vomiting and fatigue were more likely to end up in the hospital than those without this combination.
They also reinforced some factors already known to increase hospitalization risk. Comorbidities like hypertension and diabetes increased the risk of hospitalization, as did being over 60 years of age. They also noted that lower socioeconomic background increased the risk as well. This was measured by zip code.
“Hospitalization can be used as an indicator of disease severity,” said Dr. Jehi. “Understanding which patients are most likely to be admitted to the hospital for COVID-19-related symptoms and complications can help physicians decide not only how to best manage a patient’s care from the time of testing, but also how to allocate beds and other resources, like ventilators.”
The nomogram is available free as an online risk calculator. In practice, it has shown itself to be well calibrated and performing well. This tool should improve physicians ability to predict which patients will need advanced treatments and resources to avoid a negative outcome.
There’s more research to be done, to be sure. But as we investigate treatments and vaccines, it’s also important to develop tools like this that can be used by doctors to ensure patients get what they need before it’s critical.
More and better tools means better treatment and a higher likelihood of reducing death tolls significantly. And that is an outcome everyone can get behind.
Keep the faith and keep after it!
Journal Reference – Lara Jehi, Xinge Ji, Alex Milinovich, Serpil Erzurum, Amy Merlino, Steve Gordon, James B. Young, Michael W. Kattan. Development and validation of a model for individualized prediction of hospitalization risk in 4,536 patients with COVID-19. PLOS ONE, 2020; 15 (8): e0237419 DOI: 10.1371/journal.pone.0237419