Underinsurance in Primary Care: A Report from the State Networks of Colorado Ambulatory Practices and Partners (SNOCAP)

Methods

Results

A total of 1133 surveys were returned from 1258 distributed to patients (Figure 1). Eleven percent of the surveys were completed in Spanish and the rest were completed in English. Practice characteristics of the 37 participating practices are displayed in Table 2.

• Download figure
• Open in new tab

Figure 1.

Survey response diagram and underinsured sample. *Total number of patients seen on the study day was recorded by each practice except one small practice which provided an average daily patient count (40 patients per day).

Family medicine practices were more heavily represented; whereas rural and urban practices as well as community health centers, private practices, and hospital affiliates were more evenly represented.

Patient characteristics are displayed in Table 3. Patients tended to be between the ages of 18 and 64 (71.8%; mean age, 44.4 years; SD, 20.15 years); female; non-Hispanic white; and privately insured during the year with a self-reported annual income of <$25,000. Patients also tended to report they were in good to excellent health. The indicators for underinsurance reported by the 344 underinsured patients were delayed seeking care (232, 67.4%); unable to fill a prescription (211, 61.3%); unable to see regular doctor (176, 51.2%); unable to see a specialist (160, 46.5%), unable to receive some other recommended medical care (139, 40.4%); or unable to have a test (125, 36.3%) because of the inability to afford the care. We found no associations between any indicators for underinsurance and insurance type (P > .05).

Overall, 30.4% of our sample was classified as underinsured. Of the 948 patients with insurance for a full year, including Medicare or Medicaid, 36.3% were classified as underinsured (Figure 1). Of those who were underinsured, 19.2% had Medicare and 32.6% had Medicaid.

Characteristics of underinsured and insured patients are presented in Table 4. Higher proportions of women, Hispanics, lower incomes, and patients aged 18 to 39 were underinsured than adequately insured. Underinsured patients also reported higher proportions of fair or poor health status and that their health had suffered. Presented in Table 5 are odds ratios of being underinsured versus adequately insured by patient characteristics. Patients who reported their general health as fair to poor and who believe their health suffered because they were unable to afford the cost of necessary care also had higher odds of being underinsured (P < .01). When similar adjusted univariate analyses were performed comparing the underinsured with the uninsured patient population, no statistical differences among the demographic characteristics were found (P > .05) with the exception of income (data not shown). Patients reporting an annual income of <$20,000 had higher odds of being uninsured than underinsured (odds ratio, 0.28; 95% CI, 0.11–0.72). Patients who were female, younger than 65, African-American or Hispanic, or who had an annual self-reported income of <$50,000 had higher odds of being underinsured than adequately insured (P < .01). In addition, patients who reported Medicare as at least one form of insurance had a lower odds of being underinsured (P < .001) than adequately insured; whereas no statistical associations were found among insured patients and those reporting Medicaid as at least one form of insurance (P = .219).

Discussion

This study of patients in Colorado who were seen in primary care practices found that more than one third (36.3%) with insurance were underinsured and that they were very likely to report that their health had suffered because of the inability to afford care. Our finding of the percentage of underinsured is much higher than estimates of 7% to 12% found in national samples of nonelderly adults,^6,7 and higher than the 4.5% estimate for Colorado in 1995.³ The definition used in this study was based on experiential criteria as opposed to financial criteria used in many other studies. As costs shifting to patients increase over time the number of underinsured patients is expected to increase. As other studies have shown, we found that underinsured patients are more likely to report fair or poor health status than adequately insured patients; women and poorer patients are more often underinsured. Contrary to what has been reported elsewhere, we found that rural patients were not significantly more likely to report being underinsured.⁶

Unlike previous studies on underinsurance,^6,7 our study also included patients over the age of 64 and under the age of 18, and those patients with Medicare, Medicaid, or both. Although those with Medicare seem less likely to be underinsured, this group still faces financial concerns when obtaining recommended care. Importantly, those with Medicaid seem to be equally likely to report being underinsured and in poorer general health as those with no insurance. Because this is a poorer patient population, contributions to their own care, even if in smaller amounts, may still cause them to forgo care because they did not feel that they could afford it. This has enormous implications for the current efforts at health care reform that may rely heavily on expanding Medicaid without other substantial reforms.

This study suggests that the underinsured outnumber the uninsured and that underinsurance does not defer the suffering associated with the inability to pay for care. Accordingly, merely increasing access to insurance does not seem to be the much-hoped-for panacea for solving the nation's health care crisis. Consideration of the extent of underinsurance is essential in any reform effort, especially if half of underinsured patients report that their health has suffered as a result, a rate equal to that of the uninsured. Our findings are consistent with those of the Kaiser Commission, which found that as many as 18 million nonelderly adults with private insurance coverage have significant problems paying their medical bills.² The Kaiser Commission study also found that those with private insurance but medical debt were more likely to skip recommended tests, fail to fill a prescription, or postpone care because of cost.

This study has several limitations. Because of missing data, responses to the survey questions about out-of-pocket expense were difficult to interpret, which limited our ability make judgments about underinsurance based on income and medical expenses. This study was conducted in primary care offices in which patients self-reported their experiences so it is not known which delays or missed tests or examinations were of clinical significance. In addition, this sample may not be generalizable to the general population, although it does reflect reported demographics of those who seek medical care in primary care offices. A little over half of the patients seen in participating practices were offered a survey from the front office staff, which may have led to biased results. However, among those patients who received a survey, 90% returned it completed. Because this study was done on a single day in each practice, this sample of patients could be different from the practice as a whole, creating potential bias. Finally, this study likely underestimates the total number of underinsured people because some may be avoiding seeking care because of their underinsurance or because they have not yet exceeded their ability to pay for their care.^2,7

Conclusions

To determine the prevalence of patients with insurance who were underinsured, this study used the practical approach of asking patients about their decisions to delay or forgo recommended care. Using this approach, 36.3% of the patients with insurance who were surveyed in primary care practices were underinsured. It is important for the primary care provider to understand that a substantial percentage of their patients may not follow through with their recommendations because of inability to afford the recommended care, despite having insurance. This may even include patients with Medicare and Medicaid. Understanding this issue may lead to additional conversations between primary care providers and their patients, which could potentially result in a modification of physician recommendations. Although this study was not designed to extrapolate the number of underinsured patients participating in this study to the population of the state or the country as a whole, it is intended to point out that the problem of underinsurance may be significant and should be considered when exploring options for health care reform. The fact that the study only looked at patients who were willing to come to the office to see a health care provider and excluded those who may not have even been willing to do that out of concerns of the costs would suggest that this study would underestimate the total percentage of underinsured. The burden of this could be significant given our finding that about 50% of underinsured patients felt that their health suffered because of inability to afford recommended care the same level as the uninsured patients.

This study highlights underinsurance as a significant problem that requires further research to define state and national prevalence, health outcomes, and potential solutions. This is particularly important when evidence suggests that those who are underinsured have outcomes equally poor as those who are uninsured. When considering health care reform it is essential to address underinsurance if the reform is an attempt to provide adequate financial protection from health care expenditures.

Appendix: Survey Items that Determined Underinsurance Status of Insured Patients*