Key Questions: What Are the Health Consequences of Being Uninsured?

VI. Key Issues: Financing & Delivery >> D. Health Insurance Coverage >> Uninsured >> Key Questions: What Are the Health Consequences of Being Uninsured? (last updated 1.29.17)

Mortality Risk

Returns to Medical Care

  • Douglas Almond et al., Estimating Marginal Returns to Medical Care: Evidence from At-Risk NewbornsThe Quarterly Journal of Economics 125, No. 2 (2010): 591–634, . Infants classified just below threshold for “very low birthweight” have lower mortality rates than infants classified just above the threshold because they tend to receive timely and appropriate medical care.

Uninsured vs. Private Insurance

Studies listed in chronological order.

  • Institute of Medicine (2002). IOM calculated 18, 314 excess deaths among 22.683 million uninsured adults age 25-64 in 2000, based on two earlier studies showing mortality risk was 25% higher among the uninsured compared with those with private health insurance. This implies 1 excess death for every 1,239 adult uninsured.
  • McWilliams, J. M., A. M. Zaslavsky, E. Meara, and J. Z. Ayanian. 2004. Health insurance coverage and mortality among the near-elderly. Health Aff (Millwood) 23:223-33.
  • Joseph J. Doyle (2005). Health Insurance, Treatment and Outcomes: Using Auto Accidents as Health Shocks.The Review of Economics and Statistics 87, No. 2 (2005): 256–270. Uninsured individuals receive less care and have higher mortality following auto accidents.
  • Baker, D. W.J. J. SudanoR. Durazo-ArvizuJ. FeinglassW. P. Witt, and J. Thompson. 2006. “Health insurance coverage and the risk of decline in overall health and death among the near elderly, 1992-2002.” Med Care 44:277-82.
  • Dorn, S. (2008). Uninsured and Dying Because of It: Updating the Institute of Medicine Analysis on the Impact of Uninsurance on Mortality. Using the IOM method of applying age-specific hazard ratios to 29.489 million uninsured adults 25-64 in 2006, this study estimates there were 22,211 excess deaths attributable to lack of coverage. This implies 1 excess death for every 1,328 adult uninsured. An alternative calculation using the global hazard ratio for uninsured adults rather than age-specific hazard ratios, shows the number of excess deaths was 20.5% higher (which implies a hazard ratio=1.31 or 1 excess death for every 1,205 adult uninsured).
  • Levy H and Meltzer, D. (2008). The Impact of Health Insurance on Health. Annual Review of Public Health. 2008; 29: 399-409. In this systematic review of the available evidence, authors conclude “many of the studies claiming to show a causal effect of health insurance on health do not do so convincingly because the observed correlation between insurance and good health may be driven by other, unobservable factors. Second, convincing evidence demonstrates that health insurance can improve health measures of some population subgroups, some of which, although not all, are the same subgroups that would be the likely targets of coverage expansion policies.”
  • McWilliams, JM (2009). Health Consequences of Uninsurance among Adults in the United States: Recent Evidence and Implications. Milbank Quarterly. 2009; 87: 443-494.
  • Wilper, AP, Woolhandler, S., Lasser, KE, McCormick, D, Bor, DH, and Himmelstein, DU (2009). Health Insurance and Mortality in US Adults. Am J Public Health. 2009; 99: 2289-2295. Using a similar method to IOM, a 2009 study calculates that after controlling for age, gender, race/ethnicity, income, education, self- and physician-rated health status, body mass index, leisure exercise, smoking, and regular alcohol use, the uninsured were more likely to die (hazard ratio=1.40; 95% CI=1.06, 1.84) than those with insurance (Wilper: Table 2).
    • Dranove, David (2009)Health Insurance and Mortality in US Adults: Another Myth in the Making? Code Red: Two Economists Examine the U.S. Healthcare System. October 1, 2009. This critique of Wilper (2009) concludes: “In regression and related analyses, a critical assumption is that the unobservable characteristics of the “control” and “experimental” groups are uncorrelated with the observables. Translation in this case – if the regression model does not include all possible factors that might predict mortality, and just one of these omitted factors is correlated with insurance status, then the reported coefficient on insurance status is biased. This is an onerous requirement for sure, but it must be met if bias is to be avoided. Without this full set of variables, and in the absence of a randomized experimental design, it is still possible to avoid bias by using advanced statistical techniques such as “instrumental variables” regression. But the Harvard study does not use this technique. The implication is that their results are biased. We can even guess at the direction of the bias. They started with a few control variables and the estimated impact of being uninsured was much bigger than 44,000 lives lost. When they added additional control variables, the estimated impact fell. It is plausible to suppose that with additional controls, the estimated impact would shrink further, perhaps to the point of statistical insignificance. I don’t know that for sure. No one does.”
  • Kronick, R. (2009). Health Insurance Coverage and Mortality Revisited. Health Services Research. 2009. 44(4): 1211-1231. This follow-up study was similar to IOM and Wilper except that it was able to obtain repeated measures of changes in health insurance status during the follow-up period. It found a slightly elevated mortality risk among the uninsured (compared to those with employer-based coverage) that was statistically insignificant (hazard ratio 1.03, 95 percent confidence interval, 0.95–1.12). The author concluded that “the Institute of Medicine’s estimate that lack of insurance leads to 18,000 excess deaths each year is almost certainly incorrect. It is not possible to draw firm causal inferences from the results of observational analyses, but there is little evidence to suggest that extending insurance coverage to all adults would have a large effect on the number of deaths in the United States.”
  • Goodman et al. (2009). This NCPA study includes a section on Does Lack of Health Insurance Cause Premature Death? Synthesizes and critiques available literature.
  • O’Neill, June E. and Dave M. O’Neill, Lack of Insurance and Health Outcomes in Who Are the Uninsured? Employment Policies Institute. June, 2009. Synthesizes and critiques available literature.
  • McWilliams, J. Michael (2010)Letting Perfect be the Enemy of Good? The Incidental Economist. February 15, 2010.  

    Yet several other observational studies that controlled for an equally robust set of characteristics have consistently demonstrated a 35-43% greater risk of death within 8-10 years for adults who were uninsured at baseline and even higher relative risks for older uninsured adults with treatable chronic conditions such as diabetes and hypertension (Baker et al. 2006; McWilliams et al. 2004; Wilper et al. 2009). Because these observational studies are not sufficiently rigorous to support causal conclusions, we should look to studies that are more experimental in design for more definitive evidence.

  • Kim, Jenny and Jeffrey Milyo (2011). Health Insurance and Mortality in US Adults: A Cautionary Tale. University of Missouri in Columbia, MO: September 17, 2011. A 2009 observational study (Wilper et al.) reported that private insurance status is associated with decreased mortality risk compared to no insurance. Employing the same statistical model but with more recent data, we observe a weaker and statistically insignificant relationship (implied hazard ratio=1.18, 95% CI=0.94, 1.45). However, Medicaid coverage is associated with increased mortality risk; the adjusted hazard ratio for Medicaid compared to no insurance is 1.32 (95% CI = 1.01, 1.72). These findings bolster concerns about using observational studies to understand the health consequences of insurance.
  • Sommers BD, Long SK, Baicker K. (2014). Changes in Mortality After Massachusetts Health Care Reform: A Quasi-experimental Study. Ann Intern Med. 2014;160:585-593. doi:10.7326/M13-2275. Concludes 1 death is averted for every 830 non-elderly adults who attain coverage. This is based on changes in county-level rates of all-cause mortality relative to statistically matched control counties, as opposed to tracking individuals over time. The authors used propensity scores to determine which counties were most similar to MA counties; the authors estimated propensity scores with a population-weighted logistic regression model using age distribution, sex, race/ethnicity, poverty rate, median income, unemployment, uninsured rate, and baseline annual mortality as predictors. For each of the 14 MA counties, there were approximately 36 matching counties used for comparison (which collectively represented about one-fifth of the country). Several criticisms of this study have been raised:
    • Can the Methodology Plausibly Detect An Effect So Small? Jim Manzi notes that the study concludes that Romneycare was associated with a reduction in non-elderly adult mortality of 0.0082 percentage points (8.2 fewer deaths per 100,000 persons). The non-elderly adult population of Massachusetts in 2010 was about 4.2 million. 4.2 million x 8.2 / 100,000 = 344. The authors are asserting that they are detecting a reduction of a few hundred deaths per year in Massachusetts associated with Romneycare.
    • All Mortality Gains Are Attributed to Romneycare. There are two potential problems.
    • All Mortality Gains Are Attributed to Gains in Coverage. To arrive at the estimate of 1 death averted for each 830 newly covered, the authors compared the net reduction in mortality in Massachusetts counties (after statistically accounting for differences between each Massachusetts county and matching counties in other states selected because they were comparable) relative to the 6.8 percentage point reduction in the uninsured rate among non-elderly adults (i.e., 6.8%/ (8.2/100,000)=829.3). Again, there are two issues:
      • Romneycare Improved Quality of Existing Coverage. While its major focus admittedly was on expansion of coverage, workers reported that the scope and quality of their coverage had improved between 2006 and 2010 (i.e., range of services, choice of doctors and quality of care available) (Exhibit VII.3). The methodology provides no way of differentiating reductions in mortality attributable to such improvements in coverage relative to gains in the number covered.
      • Implied Reduction in Uninsured Mortality Risk is Implausibly High. Across the entire population, there was a 2.9% decline in relative mortality after adjusting for differences between Massachusetts counties and their statistically comparable counterparts. But because the authors are assuming all of these gains occurred solely among the 6.8% who gained coverage, this implies a 30.5% decline in mortality risk among the previously uninsured (if the population-wide decline in deaths is assigned to just the 6.8% gaining coverage, this decline follows algebraically). But this in turn implies that the elevated death risk facing the uninsured is 43.9% higher than for those with insurance (also derived algebraically). This elevated risk greatly exceeds that reported in observational studies (Franks=1.25; Dorn=1.31; Wilper=1.40; Kronick=1.03; Kim=1.18), especially when one considers that all of these studies all compared being uninsured to having private coverage, as opposed to Medicaid. The evidence that Medicaid reduces mortality  risk is much thinner (see below).
    • Impact of Great RecessionMegan McArdle notes that Massachusetts had a substantially more favorable response to the Great Recession, which could explain some of the findings. The authors control for unemployment and poverty. But Avik Roy argues “if the financial crisis affected Massachusetts differently from the rest of the country — and it did — then that may well show up in mortality in ways that are hard to measure.” On a related point, “Massachusetts’ response to the Great Recession may have been different than other states’ due to its wealthier economy based largely on universities and technology, in a way that could explain the difference in mortality.”
    • Contamination Due to Changes in Population.
      • General Migration Trends. Avik Roy also has noted: “The population of Massachusetts has not been static; approximately one-third of people living in the state weren’t born there, and another one-third no longer live there; around 1 million people have moved into the state and more than 1 million have moved out over the time frame of the study.”
      • Potential Changes in Low Income PopulationRomneycare also may have stimulated in-migration of low-income people looking for subsidized health care. Those able/willing to move in this fashion may have been in better health, thereby improving mortality outcomes in MA while worsening them elsewhere.
    • Massachusetts Experience May Not Be Generalizable. Massachusetts ranks #4 in the nation in America’s Health Rankings. Whether states such as Mississippi (#40) or Arkansas (#49) could achieve similar gains is an open question.

Uninsured vs. Medicaid

Studies listed in chronological order.

Uninsured vs. Medicare

  • Card, D., C. Dobkin, and N. Maestas (2004). The impact of nearly universal insurance coverage on health care utilization and health: evidence from Medicare. NBER Working Paper Series. Cambridge, MA: National Bureau of Economic Research.
  • Finkelstein, A. and R. McKnight (2008). What Did Medicare Do? The Initial Impact of Medicare on Mortality and Out of Pocket Medical Spending. Journal of Public Economics. 2008; 92: 1644-1669.
  • David CardC. Dobkin, and N. Maestas (2009). ‘‘Does Medicare Save Lives?’’ The Quarterly Journal of Economics 124, No. 2 (2009): 597–636. Individuals just old enough to qualify for Medicare have lower mortality rates (despite similar illnesses) than those just too young to qualify.

Other Impacts on Health

Impact on Health Status

  • Ayanian JZWeissman JSSchneider ECGinsburg JAZaslavsky AM. Unmet health needs of uninsured adults in the United States. JAMA. 2000;284: 2061–2069.
  • Ayanian JZZaslavsky AMWeissman JSSchneider ECGinsburg JA. Undiagnosed hypertension and hypercholesterolemia among uninsured and insured adults in the Third National Health and Nutrition Examination Survey. Am J Public Health. 2003; 93:2051–2054
  • Institute of Medicine (2002)America’s Uninsured Crisis: Consequences for Health and Health (2009).  This report provides an updated review of the research evidence on whether being uninsured is harmful to the health of children and adults.
  • Hadley, Jack (2003). Sicker and Poorer: The Consequences of Being Uninsured. Medical Care Research and Review 60 (No. 2): 3S–75S, (2003). Lack of coverage is associated with a significant increase in morbidity.
  • Decker, S. L. and D. K. Remler (2004). How much might universal health insurance reduce socioeconomic disparities in health? : A comparison of the US and Canada. Appl Health Econ Health Policy 3:205-16.
  • McWilliams, J. M., E. Meara, A. M. Zaslavsky, and J. Z. Ayanian (2007). Health of previously uninsured adults after acquiring Medicare coverage. JAMA 298:2886-94.

  • McWilliams, J. M., E. Meara, A. M. Zaslavsky, and J. Z. Ayanian (2009). Differences in control of cardiovascular disease and diabetes by race, ethnicity, and education: U.S. trends from 1999 to 2006 and effects of Medicare coverage. Ann Intern Med 150:505-15.
  • McWilliams, JM (2009). Health Consequences of Uninsurance among Adults in the United States: Recent Evidence and Implications. Milbank Quarterly. 2009; 87: 443-494. 

Impact on Health Behavior

According to David Whelan (5.2.13), “There is mixed evidence on whether having health insurance leads to greater risk-taking behavior. Kenneth Arrow “argued that health insurance actually has the perverse effect of making you reckless with your health, much like seat belts and air bags result in reckless driving. [Gary] Becker countered that someone who makes the effort to buy health insurance is likely someone who values “self-protection”–just like a homeowner who installs lightning rods probably also uses smoke detectors and blows out candles.”

  • No Effect on Risk-Taking Behavior. The RAND Health Insurance Experiment, a randomized controlled trial conducted in 1970’s, found no evidence that having health insurance increased obesity or smoking rates.
  • Adverse Effects on Physical Activity, Smoking and Drinking. “In 2006 two economists, Dhaval Dave and Robert Kaestner, completed a study in which they looked at a group of Americans who had no insurance but then turned 65 and qualified for Medicare. This before-and-after experimental framework allowed them to control for “having insurance.” Interestingly they found that women who gained insurance by turning 65 did not change their health behaviors or outcomes. But men, upon getting Medicare, increased their risky health behaviors. Physical activity dropped by 40%, cigarette smoking increased by 16%, and drinking alcohol with regularity increased by 32%.” Although authors controlled for work status, it is possible these effects are explained by the fact that many people retire at age 65.
  • Adverse Effects on Physical Activity, Smoking and Drinking. In 2008 another economist Anderson Stanciole analyzed longitudinal data from the U.S. Panel Study of Income Dynamics. The panel tracks 8,000 families. Stanciole found that after controlling for age, employment, income, race, sex, and other factors, having insurance corresponded with drinking more, smoking more, and being more sedentary.
  • Adverse Effects on Obesity. A provocatively titled paper in 2009 by Jay Bhattacharya called “Does Health Insurance Make Your Fat?” found a tie between having insurance and obesity. The National Longitudinal Survey of Youth–which followed 12,000 teenagers over 15 years–provided the data. Bhattacharya found that teenagers’ body mass index was 2.1 points higher when they were on Medicaid and 1.3 points higher among those with commercial insurance, compared to the uninsured.

Leave a Reply

Your email address will not be published. Required fields are marked *