Quoting a factoid about the distribution of actual medical _spending_ doesn't change the incorrectness of projecting CPI onto medical _costs_. There are indices - the BLS maintains one and a number of actuarial firms publish others - that track the costs of various medical services and they all outpace CPI.
You're right, there are a lot of factors at play here, its not a simple issue, and both myself and OP were somewhat oversimplifying things on opposing ends of the spectrum. My point though is that simply looking at the increase in the cost of medical services across the board is different than looking at how much the average American family is actually spending on medical costs year over year, and that is a key differentiation.