17 May 2012

Is the ER biased against uninsured kids?


This study was flagged widely in the press recently. It's a good study, based on my cursory review, that addresses an important point:

Insurance Status and the Care of Children in the Emergency Department (full text link)
Usual disclaimers about the validity of the underlying database apply, but overall I can't disagree with their findings. Privately insured kids who go to the ER are more likely to receive diagnostic tests or interventions than those who are uninsured or on public insurance (i.e. Medicaid/SCHIP). The study authors, wisely, refrain from making sweeping statements about the cause of such disparity. Which doesn't prevent the media from leaping to conclusions, and going right for the salacious ones:

Study: Privately Insured Kids Get More Care In ED 
Emergency departments are required to treat everyone who comes through the doors, but that doesn’t mean they treat everyone the same way. 
Insurance coverage may play a major role in the kind of care a young patient receives, according to a study published in the most recent edition of The Journal of Pediatrics.

No, no, no, no a thousand times, no. The implication here is that the ER discriminates inappropriately, either undertreating indigent children or overtreating insured children. This is not what the study says. It is, I hasten to add, a valid question, worthy of research. Bias based on socioeconomic status is a real factor in medicine, well documented, and should be looked into. But if you look at the very abstract of the study in question, it concludes: "It is unclear whether these patterns represent appropriate utilization."

The problem is that the database does not allow the researchers to account for the different characteristics of two very different populations presenting to the ER. The acuity of the uninsured children is much lower, generally (in fact, the paper affirms that the triage acuity of the uninsured group was significantly lower). In large part this is because the access to primary care for kids on Medicaid is very poor-to-nonexistent, so they substitute the ER for a PCP, visiting with very minor illnesses and well child exams at a far higher rate than privately insured kids. So you would expect a lower rate of testing in the underinsured group, because they are not the same as the privately insured kids, who have good access to pediatricians, and tend to come into the ER when they are sicker and more in need of tests.

To properly evaluate whether insurance status leads to inappropriate disparities in treatment, it would be necessary to acuity-match the two populations. The one element in the cited study that roughly does so is in the comparison of the care provided to admitted children, and in that subgroup, there was no difference in the amount of tests provided. This is not surprising, since the admitted kids are by definition the sickest and most likely to require tests and interventions. More important, if you are looking for disparities in how kids are treated, is to compare matched groups of discharged children with comparable presenting complaints. but this study, due to the limitations in the data source, cannot do that.

The study authors are remiss, in my opinion, in under-recognizing the actual acuity differences in the two groups. That they chose to headline their own study with the disparity in treatment with barely a mention of the patient-specific factors is leading and invites readers to draw conclusions which are not warranted.


  1. I'm interested in your statement that children with medicaid have little-to-no access to primary care... I'm a medical student with a 4 month old, and exactly the opposite is true where I live. We go to a very well-reputed private pediatrician here in town who accepts all medicaid patients. We've had no problem at the local night clinic when we've had to go (albeit that this is run through the medical school which is affiliated with a large public regional medical center). Is this idea demonstrated in the literature? And is the lack of access really a lack of access or a lack of appropriate utilization? That certainly changes the acuity question.

    I'm also not 100% clear why the study could not have done worked to correct for the acuity issue... would a different method of study have been more appropriate (i.e. case-controlled, matching for acuity or presenting complaints)? Interested in your perspective...

  2. @anonymous, medicaid access is hugely variable depending on region. Where shadowfax and I work, it's pretty decent. I am a general ped and I take medicaid happily. That said, even here the medicaid population has more barriers to care - transportation issues, language and cultural barriers, and plain old ignorance of medical issues, which makes it harder to know what's an emergency and what can wait.

    All that said, I agree with Shadowfax that the conclusions of this study are pretty useless.

  3. In my ER we tend to do more testing if we're unsure if a kid will get follow up. For example, a kid that we know the parents will be able to get in to see their pediatrician in the morning may be less likely to get a full workup if its a borderline call.

  4. 30 seconds on Google:
    Medicaid Patients Struggle to Get Primary Care, Visit ERs More

    Half of physicians not accepting new Medicaid patients (Table 2)

    25% of U.S. counties in greatest need do not have a community health center

    And on the acuity-control issue, the database is a huge national survey of ER charts, retrospective, for general purposes. The authors simply mined this data (which is the point of its existence). But the database doesn't collect detailed enough info to due true case-matching. To do that you'd need a specifically designed data collection tool or a prospective study.

  5. I work prmarily with adults and elders rather than kids, in the same community where Shadowfax works. I hear almost daily that patients cannot find primary care providers who accept Medicare and/or Medicaid. As a result, my folks try to use their specialist (nephrologist or neurologist) as primary care and end up in the ER more than they would if they had a true primary care provider.

  6. I'm not at all surprised; ER blogs are filled with stories of people bringing kids in with non-emergent cases. Yet, my middle-class peers and I would never bring our kids into the ER unless it was a real emergency (broken bone, bad croup, stitches, etc). We all know we can get into our peds practice on short notice with the merely urgent stuff. If people are bringing their kids into the ER for chronic issues or something like a fever of 102, clearly they don't have that kind of relationship with a PCP. Nobody WANTS to go to the ER. No offense ;)