At this point it’s been made to seem like common sense: Larger-bodied people are at higher risk from this pandemic. “Those who are overweight really need to be careful,” France’s chief epidemiologist declared last week. “That is why we’re worried about our friends in America.”

In recent weeks, many news outlets—and a few scientific journals—have sent the same frightening message. A study posted on a preprint server last weekend by researchers at New York University provided fodder for the latest round of this reporting: “Obesity appears to be one of the biggest risk factors related to Covid-19 hospitalizations and critical illness,” Newsweek claimed on Tuesday. Yet this rhetoric is based on flawed and limited evidence, which only exacerbates the stigma that larger-bodied people already face in society and our health care system. That stigma is what truly jeopardizes their health, not weight itself—a fact that’s only more important to consider in the midst of this pandemic.

To date, the most plausible research pointing to higher BMI as a risk factor includes three preliminary reports that have been released since April 8: a Centers for Disease Control and Prevention report with descriptive statistics on people who’ve been hospitalized for Covid-19, showing that 48 percent of those with available BMI data are in the “obese” category (a slightly higher percentage than the 42 percent in the US as a whole); a small French study that found people with a BMI of 35 and above are at higher risk of being put on a ventilator; and a letter to the editor of the journal Clinical Infectious Diseases from researchers at NYU’s School of Medicine (including one of the authors of last week’s preprint), sharing a preliminary finding that people with a BMI of 30 or above appear to be at higher risk for hospitalization and intensive-care admission, if they’re less than 60 years old. (Among people who are 60 or older, weight status did not seem to be important.)

All of these reports are flawed in similar ways. Most important, none of them control for race, socioeconomic status, or quality of care—social determinants of health that we know explain the lion’s share of health disparities between groups of people. Structural racism and other forms of inequality in our society have long been linked to worse health outcomes, including higher rates of diabetes and hypertension (two likely Covid-19 risk factors) among people in oppressed groups. Now, those health disparities are on full display in the Covid-19 pandemic, which is disproportionately impacting black communities—not because of biology, but because of systemic inequalities like higher rates of exposure to the virus and less access to medical care.

As it happens, that recent preprint from NYU did take race into account, in its finding that having a very high BMI was a major risk factor for hospitalization. But the same analysis also found that BMI was only marginally important at predicting which hospitalized patients would go on to have “critical” illness. It also seemed to indicate that being African American was in some way significantly protective against Covid-19: Black patients admitted to the hospital were only half as likely as white patients to develop the most serious symptoms, according to the study. Needless to say, few if any outlets touted this dubious result.

Another glaring issue with the three published reports about BMI and Covid-19: They don’t control for known individual health risks that may be associated with worse outcomes for this virus, including asthma and other chronic respiratory conditions, cancer, and immunosuppressive medication use. The CDC report and NYU letter to the editor don’t control for diabetes or cardiovascular conditions, either, although these two likely risk factors for Covid-19 happen to be associated with higher BMI. The French study of 124 patients does control for diabetes and hypertension, as well as dyslipidemia, but not for other risk factors—even though in the study’s introduction, the authors themselves acknowledge that cardiovascular disease, chronic respiratory disease, and cancer are also likely to raise the risk of Covid-19.

Moreover, these reports all fail to control for the particular ways in which clinicians’ biases and beliefs about body size are likely to be influencing Covid-19 care decisions for higher-weight people. In 2013, the American Medical Association labeled “obesity” as a disease (against the recommendation of the AMA’s internal committee devoted to studying the matter), and the CDC has included a BMI of 40 or above on its list of risk factors for severe Covid-19 illness since mid-March. So higher-weight people may be more likely to get hospitalized simply because they’re viewed as unhealthy and deemed higher-risk patients. The April 8 CDC report only includes BMI measurements for 10 percent of the patients in the sample, and while it’s understandable that they’d have so much missing data amid the chaos of a global pandemic, it’s also possible that higher-weight people are more likely to be weighed because their weight is assumed to be clinically relevant. Thus, people with a high BMI could just be overrepresented in the data.

Read all of our coronavirus coverage here.

Where did the CDC get the idea that people with a BMI of 40 or above are at greater risk in the first place? It’s unclear. A CDC press contact didn’t respond to a request for comment, but the peer-reviewed evidence that was available at the time the agency made that pronouncement generally indicated weight was not a risk factor. Nearly all published data from China (where Covid-19 has been studied since first being discovered in December 2019) shows that high BMI alone isn’t associated with developing the disease or with having a critical outcome. In most Chinese studies, high BMI doesn’t even make the list of preexisting conditions among Covid-19 patients—despite the fact that one-third of China’s population has a BMI in the “overweight” or “obese” categories, and that China considers weight management a public-health priority. Early US reports from public health departments also seemed to indicate that higher BMI isn’t a risk factor: In New York State, for example, “obesity” hasn’t been on the state’s list of the top 10 preexisting conditions associated with Covid-19 fatality as of this writing. The “obesity” rate documented among some of the earliest Covid-19 cases in Washington’s King County matched that of the county as a whole.