Swine flu is a special danger to the young, but the biggest danger to the young is not an infectious disease but unintentional accidents. No matter what your age accident is among the top ten causes of death, but for those between the ages of 1 and 44 it is number one. Prevention oriented accident specialists are fond of saying that "accidents are no accidents," by which they mean that many accidental deaths are in some sense avoidable, not freakish twists of fortune. So wear your seat belts and don't go golfing in lightning storms. And while you're at it, have health insurance, since there is now new evidence that not having it makes it more likely you'll die if you do have an "accident."

The finding, in a (free access) paper in the Archives of Surgery by Heather Rosen and her colleagues at Harvard Medical School (hat tip Daniel Cressey, Nature's blog The Great Beyond), is the latest evidence that being uninsured is bad for your health in all sorts of ways: you are less likely to be screened for colorectal cancer, less likely to be admitted to a specialty hospital for heart attack and heart failure and coronary by-pass, have larger breast tumors, more severe disease and lower rates of surgical treatment, more likely to be organ donors (a reflection of sudden and traumatic death) and get fewer services once in the hospital. Insurance status is an independent predictor of death from trauma in a recent study, but that work wasn't able to account for the presence of other diseases (comorbidities) that might be more common among the uninsured trauma victim and thus skew the mortality results (these previous findings are cited as references 5 through 12 in the Rosen paper and more extensively described there). The new paper was designed to address the problem that these comorbidities explained why being uninsured made it more likely you would die in a hospital after an accident, even though there is a legal requirement on hospitals meant to prevent your payer status from affecting your emergency care. The Emergency Medical Treatment and Active Labor Act of 1986 makes it illegal to transfer a patient from one hospital to another or even refuse treatment when the patient is medically unstable. Legal or not, there is evidence that your insurance status does matter, and the Rosen et al. study set out to explore what factors were involved. Here's what they did:

• Data were provided by the National Trauma Data Bank (NTDB) (version 7.0), which contains data from 2.7 million admitted patients from more than 900 US trauma centers (from January 1, 2002, through December 31, 2006). As of 2002, there were 1154 trauma centers in the United States. Trauma center participation in the NTDB is voluntary, and thus these data are a convenience sample. The NTDB contains information on patient demographics, preexisting comorbidities, emergency department care, complications, injury severity, patient outcomes, hospital charges, and hospital information.
• We included data on adult patients (aged 18 years) who had been injured via blunt or penetrating trauma (we excluded patients with burns). Patients were divided into the following categories of insurance coverage: uninsured (includes self-pay), a managed care organization, commercial indemnity insurance (including automobile insurance, Blue Cross/Blue Shield, no-fault insurance, worker's compensation, or another commercial indemnity plan), Medicare, and Medicaid, for a total of 5 categories. (Rosen et al., "The Accidental Cost of Being Uninsured," Arch Surg. 2009;144(11):1006-1011)

The data in the trauma registry enabled the research team (which included Atul Gawande, whose books and articles on medicine and surgery in The New Yorker have brought him independent fame) to account for differences in injury type and severity as well as age, sex and race. The results corroborated what has been found previously: being uninsured puts you at increased risk of dying in the hospital after an injury. The increased risk is substantial (90% increase after accounting for the other factors) and is not explained by comorbidities (age, race and severity of injury all played a part but also didn't account for the difference; the 90% increase is after taking those factors and comorbidities into account). In fact the difference was greatest in the youngest age group (18 to 30 years) where other medical conditions would be least expected.

Still, it's not clear what the mechanism for the pervasive risk of being uninsured. The authors mention some possibilities (differences in care once in the hospital, delay in treatment once in the hospital, communication difficulties), but one not mentioned seems to us worth considering (confession: it was Mrs. R. who suggested it). The uninsured may be more likely to live in poor rural areas that are a greater distance from major trauma centers. Distance to hospital apparently was not one of the variables available in the registry, so this a form of "delay of treatment" that remains unaccounted for. The remarkable record of battlefield trauma care tells us that prompt treatment makes a huge difference in outcome (although the disgraceful US record in veteran aftercare is another matter).

The mechanism is important. While we are proponents of a wholesale revamping of American medical care to eliminate effects of income on access to decent and necessary medical care, if the factor is distance to a trauma center it won't be fixed by extending insurance to more of the population. It would be nice if that is all that is involved and maybe it is. But we're not there yet.

This post first appeared on the blog Effect Measure.