Privately Insured Patients Have Better Outcomes
A study in the journal Cancer that found people with private health insurance fared better after prostate surgery that either patients with Medicare or Medicaid. According to Fox News:
Researchers determined that among more than 61,000 men who had their prostates removed to treat cancer, those with private insurance had fewer complications from surgery and were less likely to die in the hospital.
Men covered by Medicare or Medicaid were more likely to need a blood transfusion to treat blood loss: almost eight percent and 11 percent, respectively, had a transfusion, compared with just over five percent of men with private insurance.
A similar study by some of the same researchers also found privately insured patients who had a kidney removed also had better outcomes than those covered by Medicare and Medicaid.
This is actually pretty bad. You can’t compare an older population (Medicare patients) to a general group of private insurance patients. Similarly you can’t compare those who are enrolled in Medicaid to those in general group. That makes little to no sense. Those groups have higher mortality based on their demographics alone. It might be shocking but older patients die sooner.
Obviously Medicare and Medicaid enrollees are in poorer health than the average privately insured patient. However, researchers undoubtedly controlled for health status or severity of condition. You could argue that there are unobserved variables implicit in those enrolled in public coverage (especially Medicaid). Or maybe the doctors that accept patients covered by public coverage spend less time or are not as skilled as those who only accept privately insured patients. This is problematic because, for the most part, it’s the same doctors treating patients covered by both types of coverage. Regardless of your view, it’s still an interesting exercise to ponder the reasons behind the results.
I’m sorry there wasn’t even a control for smoking status. So how can you say that this correctly controls for confounding variables. Prostate cancer and cigarette smoking have somewhat of a relationship (http://aje.oxfordjournals.org/content/143/10/1002.short)
But again your not arguing with me on the basis that these are two different samples being investigated. I’m fine with the confounding part. It goes back to the relationship argument you were making. But if the private insurance sample is younger (its 10 years younger according to the means). Then this study is telling us that older people are dieing, which we would assume no? Fundamentally you can’t control this away statistically this is a problem with the sample being used.
I can agree its an interesting view, but really its telling me private insurance helps people more than older Medicare patients, and a small sample of medicaid individuals. Not trying to be argumentative, but this is a bad study that should NOT be getting press.
And if you disagree with me tell me what the identifying variation is here, because the samples pretty much destroy a clear answer.
Yeah sorry not to be argumentative, just frustrated we can’t get to this study with the proper identifying variation. The authors do the best they can but this is NOT the type of research that we should be able to get gross generalizations like the heading implies.
I’m with Jonathan. Judging from the abstract, this “study” doesn’t even begin to consider possible sources of variation.
They say “Multivariable logistic regression analyses, fitted with general estimation equations for clustering among hospitals, adjusted for confounding factors.” That’s supposed to make me feel all warm and fuzzy? I’d feel a whole lot better if they bothered to list some of the independent variables they put in the fancy logistic regressions.
One wonders if they’ve been spending too much time with the Dartmouth Atlas. Hospital clustering may not control for much given the extent to which both Medicare and privately insured patients travel.
It’s really bizarre that there were so few Medicaid patients. One wonders how they dealt with dual eligibles.
So the real indictment is against private health care providers, no? The implication is that they are providing a different quality of services based on the source of coverage. Damning of providers who are discriminating against an individual’s source of coverage, not damning of who or how the bills are paid.
This study is not an indictment against anyone. There is some evidence that government run programs provide worse care than private providers. This study finds worse outcomes for those who are on government programs but the question is whether it has adequate statistical controls for known variation in the groups in question, variations that are known to affect health status.
The problem is that who pays the bills determines how much of the bill is payed. The amount of payment varies by source of coverage. Standard Medicaid pays below market for most specialists. On the whole, Medicare underpays for a lot of things, too. That is the nature of government run care. The question is whether underpayment harms health as economic theory suggests is likely to be the case.
But government programs don’t “deliver” health care services, they only pay for them. Again, your statement is that private practice, not government employed, physicians are looking at how much they are getting paid for prostate cancer or kidney surgery and then provide inferior care to people on government programs.
How am i mischaracterizing your statement and the implications of the post? And if There are inadequate controls for variations in the groups examined, of what value is posting the results of the study if not to impugn government paid health care?
That was not my take either. My question is what is the identifying variation within the models that explains the differences in the outcome variable. You cannot say as things are constructed that the variation is correct. We want two similar individuals who are similar in everything but their type of insurance explains the difference in patient outcomes. This study does not do that and as a result I don’t think we are capturing the correct variation.
Plus, the title of the post is “Privately Insured Patients Have Better Outcomes.” Don’t think I’m implying something that wasn’t explicit in the title.
And my point exactly, the variation hes explaining is NOT the difference between private and public insurance. In fact its not even close. Hence my concerns about this inappropriate title.
The study abstract says that it found that privately insured patients have better outcomes is the study finding. Hence the title on this blog.
What we are discussing is the level of confidence that one should place in those results.