You're right, their definition of 'medical bankruptcy' does include that person if the $1,001 wasn't covered by insurance, in the cases where they were comparin

You're right, their definition of 'medical bankruptcy' does include that person if the $1,001 wasn't covered by insurance, in the cases where they were comparing the results to their 2001 study which used the same criteria. Their new results, however, apply a stricter standard, upping the bottom line number to $5,000 or 10% of household income, and they report a change of less than 1% if they eliminate the hard $5,000 and just go with 10% of income. You seem to be claiming that people who are going to go bankrupt even if you remove their medical expenses are somehow meaningless data points for the question being asked by the study. That makes no sense, though. The cases that make it into the study are 'Medical debt below threshold, bankrupt anyways', 'Medical debt above threshold, bankrupt regardless', 'Medical debt above threshold, bankrupt as a result'. You're saying that we shouldn't care that the first category is proportionately shrinking, because it may be the case that the second category is accounting for all of that shrinkage. Even if that were true, that's still a reliable indicator that medical costs are growing. Your "no control group" criticism makes absolutely no sense here. What are they controlling for? They aren't running an experiment, they're reporting on the proportion of bankruptcies that meet a particular definition, and how that proportion has changed over time.