Posted by tangplay on 6/15/2020 2:58:00 PM (view original):
****, I copied the wrong link.
Here ya go:
https://www.aafp.org/news/health-of-the-public/20181210lifeexpectdrop.html#:~:text=%22The%20latest%20CDC%20data%20show,from%20drug%20overdose%20and%20suicide.
https://www.bmj.com/content/bmj/suppl/2018/08/15/bmj.k3096.DC1/mid_life_mortality_v37_datasupp.pdf
https://www.advisory.com/daily-briefing/2019/12/02/middle-age-death
https://www.snopes.com/fact-check/life-expectancy-in-the-united-states-declining-thanks-to-obamacare/
Ok, so to start out, the BMJ and Advisory Board links are both exactly what I described - summaries of Woolf. You'll notice that bmj summary uses 1999-2016 data. This is somewhat misleading RE: ACA effects, but fortunately the actual study has tables in the supplementary information reporting both 1999-2016 and 2010-2016 data separately. It's important to note how the bottom right figure changes in this time window. From 1999-2010, the components listed here as "organ disease" actually had negative excess deaths, so the 2010-2016 excess deaths are closer to 50k. However, the external causes deaths trend up over the full duration of the study, so excess deaths from 2010-2016 are IIRC ~25-30k. I already alluded to this in my summary of the Woolf paper earlier - drug abuse is in fact the leading cause of the so-called excess deaths from 2010-2016, but in an aggregated sense the things here referred to as "organ disease" are still a larger driver of excess deaths during the relevant time period than "external causes" like suicide and drug-related deaths. However, nothing in this study attempts to unpack the role of drug abuse in the organ disease category, so drug-related syndromes could play an even bigger role than these data suggest at face value. Not really clear if an addict dying of a heart attack at 47 counts as heart disease or drug deaths. Or maybe it's in there and I just don't remember that detail.
I haven't read the CDC briefings discussed in the AAFP summary. Let's be honest, it's based on a 1-year sample and not longer-term trends, so it's kind of inherently pointless. I could try to make something of the fact that the biggest proportional increase in mortality rate in their sample is from the influenza and pneumonia category, but let's be honest, the biggest driver of variance in this category is the severity of the seasonal flu epidemic from year to year. Which is exactly why single year-over-year changes aren't particularly enlightening relative to longer-term studies, which do exist.
Snopes is the huge win here for my argument. So thanks for that. I actually learned something new. Note the quote from Chetty
et al. near the bottom:
Measures of health insurance coverage and spending (the fraction of uninsured and risk-adjusted Medicare spending per enrollee) were not significantly associated with life expectancy for individuals in the bottom income quartile.
This is kinda my argument in a nutshell. Just providing coverage doesn't actually make people healthier until you do something to actually increase the availability of care. How many times today has someone said "how can providing access to healthcare not improve the health of Americans," or something along those lines? Well, you just posted the answer for me. Statistically speaking, this study that made it into Snopes found that insurance does not improve healthcare outcomes for low-income Americans.