But a team of Harvard researchers used a creative method for finding them: Using health insurance claims data, they looked at the Covid rates of families in the two weeks after one of them had a birthday. Overall, their paper, published in Jama Internal Medicine, found that a recent family birthday increased Covid risk by nearly a third. Their theory is that the increased risk is almost certainly explained by birthday parties. Though the insurance claims don’t show whether any individual Covid patient had actually held a birthday party, several aspects of the data strongly suggest a connection. For one, when researchers looked at other days of the year by randomly assigning birthdays instead of using actual birthdays, or examined diagnoses in weeks before birthdays, they found no such pattern. But, perhaps more significantly, they found the biggest infection risk in the weeks after the birthday of a child.