The study’s sample of 4,066 people represents 31.25 million people ages 18-64 who did not have access to employer sponsored insurance, and among whom 11.6% have non-group health insurance and 88.4% are uninsured (see Table 1). Twenty-eight percent of individuals have no degree, 6.4% have a GED, and 44% have a high school degree as their highest degree obtained. Hispanics make up 27.7% of the study population, and 12.8% of individuals have limited ability to speak English. The racial break down of the study population is 78.5% white and 12.7% black. In comparison, in 2005 among the general civilian population under age 65, 15.8% of individuals were Hispanic, 64.6% of individuals were white, and 12.7% of individuals were black (AHRQ, 2005).
The subpopulation of those who are uninsured have lower education levels on average than those with non-group insurance, with 31% of the uninsured having no degree comparison to 9.5% of the insured (see Table 2). Thirty one percent of individuals who are uninsured are Hispanic and 14% have limited ability speaking English, while these percentages are 4% and 2% respectively for those who are insured. Differences between the racial make up for the insured and uninsured are also present. 78% of the uninsured are white and 14% are black, while 86% of those with non-group insurance are white and 2% are black.
Perceived health status differs noticeably between the two groups in spite of the fact that the proportions of individuals with selected chronic conditions do not. Among people with non- group health insurance, 39% are in excellent health, in comparison to 24% of those who are uninsured, Similarly, 6% of those who are insured are in fair or poor health while 13% of the uninsured are.
This study used logistic regression to examine whether factors beyond health status and income affect the decision to purchase non-group health insurance. The study examined if possessing knowledge about health insurance and having the ability to navigate the non-group market impact purchase decisions. These factors are not directly measurable, and this study focused on education level, ethnic and racial minority group status, and English language ability as factors that may be correlated with knowledge about insurance. The results indicate that those who are black or Hispanic, those with limited ability to speak English, and those with lower education levels are less likely to purchase non-group health insurance, controlling for other demographic characteristics and health status.
Most interestingly the results of this study indicate that language limitations are not the cause of lower rates of non-group insurance purchases by Hispanics. The results indicate that being of Hispanic ethnicity has a statistically significant negative effect on purchasing non-group health insurance, after controlling for English language limitations. The odds of purchasing non- group insurance are 85% less for Hispanics compared to non-Hispanics. When looking at someone with average characteristics, someone who is Hispanic and does not have a language limitation is 16.2 percentage points less likely to purchase non-group insurance than someone who is non-Hispanic without a language limitation.
Interestingly individuals’ attitudes about how much they value health insurance and how likely they are to take risks were not statistically significant in the regression model. As insurance is a product that gives one protection against unknown outcomes, attitudes on willingness to take risk would be expected to play a role in purchasing decisions. In addition, indicating that one values insurance would be expected to be strongly correlated with purchase decisions. However, these results may have been affected by the significant amount of missing data on these variables.
This analysis is limited as it was not able to control for several demographic characteristics that likely impact the decision to purchase non-group insurance. These characteristics include parenthood status, assets, and family members’ insurance status (if one’s spouse or child has coverage). In addition, the variables used to control for health status are imperfect measures. The analysis only included the presence of eight chronic conditions, whereas the severity of the conditions may influence decisions, and many other health conditions exist which may influence decisions. The analysis is also limited due to the lack of information on supply side factors that affect purchase decisions. Information on state level regulations that impact the availability, generosity, and price of benefits offered was not available. The study does not include information about price, including the price of premiums quoted to individuals if they shopped for insurance and the price of premiums those insured paid.
Previous studies that came to similar conclusions about the lower likelihood of ethnic and racial minorities, and those with lower education levels, to purchase non-group insurance hypothesized that unmeasured attitudes towards medical care and the health insurance system, or discrimination against minorities in health insurance marketing, may explain these results (Saver and Doescher, 2000). Research showing that among insured individuals minorities are less likely to use medical care supports this hypothesis (Fiscella et al., 2002). Another possible explanation may be that disadvantaged groups have less exposure to health insurance, and how to navigate the health insurance marketplace, and that if they had the same level of exposure as other groups they would be equally likely to purchase insurance.
The fact that this study found similar results as one using 1987 data shows the persistence of the unmeasured underlying factors that lead to differences in purchasing rates. Future research to uncover these unknown factors would provide an important foundation for policymakers to address gaps in insurance coverage between advantaged and disadvantaged groups.
The results of this study indicate that policies aiming to use the non-group market to expand insurance coverage could exacerbate existing disparities in health insurance coverage. As of 2007, 21% of black non-Hispanic individuals were uninsured, and 34% of Hispanic individuals were uninsured, in comparison to 12% of white individuals (The Henry J. Kaiser Foundation, 2008). Policies that rely on subsidies to encourage purchase of insurance in this market, and that do not address non-economic factors that contribute to purchase decisions, will be unlikely to induce disadvantaged groups to purchase insurance. Furthermore, this suggests that any efforts to extend health insurance coverage to the uninsured by utilizing the non-group market should consider special outreach to these groups of individuals who otherwise may be more likely to remain uncovered.
