Age Isn’t the Full Picture in Covid-19 Death Risks, Population Study Finds

Adding age to social inequalities + comorbidities doesn’t really improve predictive power for Covid-19 death.

This study, “Unequal impact of structural health determinants and comorbidity on COVID-19 severity and lethality in older Mexican adults: Considerations beyond chronological ageing,” is published on 29 June in Gerontology. The lead author is Omar Yaxmehen Bello-Chavolla, MD, PhD, and specialist in Applied Statistics at the National Institute of Geriatrics in Mexico City. As the title says, despite what is known, age is not all there is in explaining Covid-19 severity and death.

The Mexican Ministry of Health documented 101,238 Covid-19 cases, of which 28,804 (20.5%) affected those over 60 years old. As expected, the older age group had higher Covid-19 death rate at 29.49% compared to younger age group at 6.95%. The older people also had a higher prevalence of medical comorbidities.

Most studies did not analyze SLI, a measure of social disadvantage and structural inequality.

The studied risk factors for Covid-19 severity include age, sex, smoking, asthma, obesity, diabetes, hypertension, chronic obstructive pulmonary disease (COPD), chronic kidney disease (CKD), immunosuppression, and social lag index (SLI). Here is the interesting part: Most studies did not analyze SLI, a measure of social disadvantage and structural inequality; for example, in access to healthcare, education, basic necessities, and living conditions. Their main findings are as follows:

• For pneumonia cases: Older age, smoking, diabetes, and obesity were independent risk factors.
• For severe cases in need of hospitalization: Older age, male sex, obesity, diabetes, COPD, CKD, and increased SLI were independent risk factors.
• For more severe cases in need of ICU admission: Obesity, male sex, and increased SLI were independent risk factors.
• For most severe cases in need of invasive ventilators: Older age, obesity, and increased SLI were independent risk factors.
• For death cases: Older age, immunosuppression, CKD, diabetes, obesity, SLI, and being treated in public (vs private) hospitals were independent risk factors. (Note: Affording treatment at private hospitals probably reflects higher SLI.)

Using Bayesian networks, they took their statistical analyses to the next level to study interactions between risk factors. They calculated that comorbidities had the most massive predictive power for Covid-19 death, followed by increased SLI and then older age.

Adding age to comorbidities + SLI (or SLI only) barely increases predictive power for Covid-19 death.

Comorbidities + SLI combined had reliable predictive power, with only slight improvement when factoring in age (i.e., c-statistics of 0.642 vs 0.661). The closer the c-statistics to 1, the more accurate the prediction. Thus, adding age to comorbidities + SLI only increases predictive power by 0.019, an insignificant effect size.

Even in older people without comorbidities, SLI was a risk factor for Covid-19 death with c-statistics of 0.676. Factoring age only increased this number by 0.011 — from 0.676 to 0.687. “In older adults without comorbidities, age was not a predictor of disease severity with structural factors [i.e., SLI] playing a larger role,” Dr Bello-Chavolla and colleagues wrote.

If existing studies measured social inequalities, the effect of age in Covid-19 mortality might not be so substantial after all.

Social inequalities might be a more tangible contributor to Covid-19 mortality than initially thought. “Whilst age is a significant independent predictor of Covid-19 lethality, comorbidities and structural health determinants likely play a larger role in increasing disease severity and conditioning risk of Covid-19 lethality in older patients,” the Mexican authors stated.

If existing studies measured social inequalities — such as with SLI (social lag index) — the effect of age in Covid-19 mortality might not be so substantial after all. It could have been blunted from the impact of better healthcare. Older people might be dying from Covid-19 as a result of social inequalities, at least as shown in the Mexican population study. Or perhaps as a result of increased medical comorbidities following ageing due to inadequate access to healthy foods, healthcare, health education, etc.

It has been said Covid-19 reveals the existing social inequalities in society. Essential workers like those working in counters, deliveries, gas stations, transportations are relatively low-paying; they are also at increased risk of viral exposure. A lower household income also means more crowding of family members, more reliance on public transportations, and fewer opportunities for remote work. In the United States, such social disparities are driven by racism and affect Black communities the most, as Emily Mullin pointed out in Medium Coronavirus Blog.

Age barely tells anything. The dynamics between social inequalities (or racism) and comorbidities best explain who is at risk for Covid-19 mortality, as the Mexican population study shows.

“Pandemics rarely affect all people in a uniform way,” an April comment in the Lancet Public Health said. “The inequitable response to Covid-19 is already evident. Healthy life expectancy and mortality rates have historically been markedly disproportionate between the richest and poorest populations.” Likewise, a June paper in the Journal of National Medical Association quoted the words of Plutarch (circa 46–120 CE), “An imbalance between rich and poor is the oldest and most fatal ailment of all republics,” and wrote, “In 2020’s crisis du jour, we remain hostage of our refractoriness.”