studies, rather than separately to each age group within this range. For 2000–06, this alternative approach raises the esti- mated number of excess deaths by an a v e r a g e o f 2 0 . 5 p e r c e n t a y e a r . 3 T a b l e shows the results of those calculations. 3
New research confirms the link between insurance status and risk of death
S i n c e p u b l i c a t i o n o f t h e I O M s t u d y , a g r o w i n g b o d y o f r e s e a r c h h a s c o n t i n u e d t o d o c u m e n t a s t r o n g relationship between health coverage and health outcomes, including mortal- ity. For example, several studies have used Health and Retirement Survey data to analyze the impact of insurance sta- tus on older adults. Examining data for adults age 55–64 from 1992 through 2000, one study found that, based on the kind of observational data employed by the studies on which IOM relied, providing all such adults with insur- ance coverage would have lowered the number of deaths by 27 percent. After controlling for the impact of health sta- tus on insurance coverage, the mortality reduction reached 42 percent (Hadley and Waidmann 2006). Using a broader measure of health outcomes, another study examining Health and Retirement Survey data found similar results (Dor, Sudano, and Baker 2006). Not only did these studies show the impact of insur- ance status on morbidity and mortality, they discovered that, after adjusting for the effect of health on the likelihood of having insurance, insurance was found to have a substantially more pronounced effect on morbidity and mortality. Because the studies on which the IOM relied did not compensate for this rela- tionship, they may have understated the impact of insurance on mortality.
Another study using Health and Retirement Survey data for adults age 55–64 found that, after controlling for socioeconomic status and other factors, uninsurance increased such older adults’ risk of dying over an eight-year period from 7.5 percent to 10.5 percent. The study thus estimated that, among such near-elderly adults alone, more than 13,000 people die every year due to
uninsurance, “plac[ing] uninsurance third on a list of leading causes of death for this age group, below only heart disease and cancer” (McWilliams et al. 2004).
Using a different data source encom- passing a broader age range of survey respondents, other researchers analyzed data for 15,792 adults age 45–64 from the Atherosclerosis Risk in Communities Study, a prospective cohort study spon- sored by the National Heart, Lung and Blood Institute. After controlling for multiple factors, the study concluded that uninsurance increased mortality rates by 26 percent (Fowler-Brown et al. 2007)—a result strikingly similar to the 25 percent mortality rate differen- tial found by the studies on which IOM relied. As with the earlier, longitudinal studies cited by IOM, the Fowler-Brown research may have underestimated the impact of insurance coverage on mor- tality because it did not control for the relationship between health status and likelihood of obtaining insurance.
Studies of particular health conditions have likewise continued to find a strong relationship between uninsurance and mortality. One analysis examined stroke, the country’s third-leading cause of death. Based on 2002 hospital discharge data for adults age 18 and older, researchers found that, after controlling for socioeconomic status and other confounding variables, the absence of insurance increased the risk of death by 24 percent or 56 percent, depending on the type of stroke involved (Shen and Washington 2007). Another study examined records of all cancer cases diagnosed in Kentucky from 1995 through 1998. After controlling for demographic factors, stage of diagnosis, and initial treatment, the study found that uninsurance increased risk of death from lung and female breast cancer by 19 percent and 44 percent, respectively (McDavid et al. 2003).
On the other hand, Kronick (2003) raised questions about the earlier studies on which IOM relied, suggesting that unobserved variables such as obesity, use of tobacco and alcohol, wealth, and the value placed on health could have played a role inflating the apparent impact of insurance on mortality. If it
had been possible to control for such variables, a less robust effect may have been observed, Kronick suggested.
However, since those earlier studies criticized by Kronick, additional research controlling for many previously unobserved factors has continued to confirm a strong link between insurance status and mortality risk. Among the articles cited above, for example, Hadley and Waidmann controlled for alcohol use, tobacco use, disability, self-reported health status, and chronic health conditions; Fowler-Brown and colleagues controlled for obesity, smoking, self- reported health status, cholesterol levels, and chronic medical conditions; and the study by McWilliams and colleagues controlled for alcohol use, obesity, exercise habits, marital status, disability, chronic medical conditions, job stress, and wealth. The latter research team further conducted a sensitivity analysis showing “that the confounding effect of unmeasured variables would have to be even greater than the impact of smoking on mortality in our study for the increased mortality of uninsured adults to become statistically nonsignificant” (McWilliams et al. 2004).
More broadly, even if unobserved vari- ables mean that lack of insurance increases the risk of death by less than 25 percent, the consequences of uninsurance could still be serious. For example, applying a 15 percent increased mortality risk to all adults age 25–64, without distinction by age cohort, yields an estimate that uninsurance caused 101,000 excess deaths since the start of the decade, including 16,000 deaths in 2006.
limitations of the current analysis
A t t h e m o s t b a s i c l e v e l , t h e a b o v e e s t i m a t e s a r e n o t p r e c i s e “ b o d y c o u n t s . ” R a t h e r , t h e reader should view them as reasonable indicators of the general magnitude of excess mortality that results from uninsurance.
Uninsured and Dying Because of It: Updating the Institute of Medicine Analysis on the Impact of Uninsurance on Mortality