Some pre-existing conditions matter much more when evaluating COVID-19 outcomes

A new paper in Methods and protocols of biology, published by Oxford University Press, indicates that certain pre-existing conditions – including degenerative neurological diseases, dementia and severe disabilities – matter far more than previously thought when assessing those at risk of death from to COVID-19.

COVID-19 has dramatically changed lives. In the United States, the disease can cause a death rate up to 163 times higher than that of seasonal flu. COVID-19 may also be more likely to result in patients requiring mechanical ventilation or being admitted to intensive care.

Pre-existing conditions, or comorbidities, make severe illness or death from COVID-19 more likely. But assessing the risk of various conditions for the severity of COVID has been difficult. Researchers have proposed several mathematical models to predict death from COVID-19 based on comorbidities. Medical institutions use these models because they facilitate patient management and resource allocation.

Many diseases increase the death rate because they weaken the immune system, make the patient more susceptible to developing infections, and cause target organ dysfunction. One method of assessing the risk of various conditions is to group them into broad categories (such as “malignancy”) and predict outcomes for each category. Another method is to weigh different pre-existing conditions differently and use the sum to predict outcomes. The researchers here believe that these approaches have substantial flaws; the true impact of a specific pre-existing condition is often not well known, broadly similar diseases are often lumped together in prediction models even though COVID-19 outcomes may be very different, and rare diseases are not not well represented.

Researchers here believe that a better approach is to systematically investigate all pre-existing conditions, determine which ones impact outcomes, and then use that to generate a predicted probability of death that represents the overall risk posed. due to comorbidity. .

Using all diagnostic codes employed by the Department of Veterans Affairs, researchers developed a new prediction model to estimate the probability of death from COVID-19. This is the largest study to date tracking patients with COVID-19 to predict mortality. Starting in 1997, researchers here used diagnostics from the first time a patient sought care up to 14 days before a positive COVID-19 test, then compared that to COVID results for all 347,220 COVID patients. treated at veterans facilities in September 2021. They found their new model, which they call PDeathDx, outperformed other conventional prediction models.

Additionally, researchers have found that certain underlying conditions are much more likely to lead to death. These include degenerative neurological diseases, dementia and severe disabilities. Since doctors do not associate these pre-existing conditions with respiratory damage or weakened immunity, conventional risk assessments fail to capture the serious risk of COVID for patients with such conditions.

Source:

Oxford University Press United States

Journal reference:

Campbell, HM, et al. (2022) A new method to manage pre-existing conditions in the development of multivariate predictive models for death from COVID-19. Methods and protocols of biology. doi.org/10.1093/biomethods/bpac017.

Leave a Reply