If we want people to trust disease epidemiologists during the COVID-19 pandemic - and the next pandemic after that, since this was the third one of just coronavirus in 17 years - then we have to start holding suspect epidemiologists accountable.
Behavior that remains alarming is manufacturing agenda-filled papers or data dredging or p-hacking to get a result. Concern about "studies" claiming that particulate matter so tiny it takes an electron microscope to detect it will still give your child a lower IQ, and hundreds of woo papers written a lot like that, were dismissed as unimportant nit-picking. Until a lot of people who were jaded by claims that meat causes cancer or global warming, that homeopathy is science when it is a molecule of a trace chemical that anti-science activists want to sue about, and that alternatives to medicine magically worked were suddenly causing fear and doubt about real science, and that began costing lives.
Belief in horse dewormer only got attention from journalists and academics because it was Republicans. The anti-vaccine and nonsense beliefs of Democrats were rationalized for 20 years prior. That is not how science is supposed to work.
The new paper engages in a bane of medical epidemiology - it invokes a risk factor for a risk factor for a disease and then the press release suggests it is causal. In this case they note the women who eat alone are at higher risk for high blood pressure which is a risk for cardiovascular issues.
So eating alone causes heart attacks? No, the confounders are that this is a case study based on surveys of only a few hundred women - two giant confounders, and why these claims have to be placed in the "exploratory" pile rather than science - and that women who live alone reported higher depression, and these women were actually eating less. Eating alone does not cause the heart disease.
It is disastrous to tell elderly people - who will already have co-morbidities that make SARS-CoV-2 much riskier - that not social distancing may give them a heart attack.
