Inadequate Health Insurance: Costs and Consequences

Karen Donelan, ScD*, Catherine M. DesRoches, DrPH* and Cathy Schoen, MS†

Abstract and Introduction

Context: Changes in the healthcare marketplace have begun to test the nature and adequacy of health insurance. The complex nature of insurance is driving us away from the notion that there are 2 distinct groups - the insured and the uninsured - toward an idea that insurance is best represented along a continuum, from the very well insured to the chronically uninsured, with a wide range of quality of coverage in between.
Objective: The objective of this study was to examine the experiences of insured adults as they try to get needed healthcare and balance the payment for these services against other basic needs.
Design: Using data from the Commonwealth Fund 1999 Survey of Workers' Health Insurance, the study analyzes the cost and access problems of insured adults by a number of different variables including income, plan satisfaction, health status, and insurance stability.
Results: Bivariate results indicate that insured adults with low incomes and those reporting fair or poor health are more likely to experience problems getting and paying for healthcare. These groups are also more likely to have problems paying for basic living expenses.
Conclusions: The most essential notion of insurance is that it will provide protection against financial risk and assurance that we can get healthcare services when we are sick. Yet, we find substantial proportions of low- and modest-income, insured adults who struggle to afford insurance premiums; we also find that their insurance plans do not provide them with either access to care when needed or financial protection from the cost of that care.

The economic successes of the past decade in the United States have not stemmed the tide of increases in the number of Americans who experience periods without health insurance in the course of a year.[1] Changes in the healthcare system have begun to test the very nature and adequacy of health insurance. The managed care revolution and evolution have changed the way many insured Americans receive care and the types of services that are covered. Employers have managed increases in insurance premiums, at least in part, by shifting an increasing share of the burden to employees.[2] In addition, premium increases and cost sharing in the individual insurance market have impacted the nature of available policies.

The changing nature of insurance and the proliferation of plan choices and coverage options are moving researchers away from the notion that there are 2 distinct groups - the insured and the uninsured - toward an idea that insurance is best represented along a continuum, from the very well insured to the chronically uninsured, with a wide range of quality of coverage in between. The published literature, however, is somewhat sparse in defining underinsurance or ways to measure it.

Measures of Inadequate Insurance

The few measures of underinsurance that can be found in the literature might be broadly classified as either economic measures or experiential measures. The first economic measure was created by Pamela Farley Short in 1985 and defined underinsurance based on a ratio of expected out-of-pocket expenditures to income. Short estimated that, dependent on different ratios, anywhere from 8% to 26% of the nonelderly, privately insured population was underinsured in 1977.[3] A more recent update of this measure found that 19% of the nonelderly population would be underinsured if faced with a catastrophic illness. In this group, underinsurance was the result both of limited insurance policies and standard policies held by people with substantial financial risk in relation to income.[4]

Experiential measures of underinsurance use access to care and actual problems paying medical bills to estimate the number of underinsured Americans. A 1994 study measured Americans' problems paying medical bills while insured, finding that of the 19% of Americans who had problems paying hospital, doctor, prescription drug, nursing home, or home healthcare bills, 75% had health insurance at the time they were surveyed.[5] Overall, 17% of the insured population experienced one of these problems. Other research has focused on problems obtaining care, specifically examining the ability to obtain hospitalization, emergency, specialty, pediatric, or mental health care, prescription drugs, medical equipment and supplies, and home care. Himmelstein found that nearly 75% of those unable to obtain 1 or more of these services were insured and 46% had private insurance.[6] Others have defined underinsurance as being insured but failing to see a doctor for major medical conditions because of cost barriers. Using this estimate, 12% to 32% of Americans were found to be inadequately insured.[7] Some have targeted inadequacies of coverage even in very comprehensive insurance policies or universal health systems where implicit or explicit rationing mechanisms limit access to desired or needed services as perceived by consumers.[8]

This article focuses on the experiences of the insured adults as they try to get needed healthcare and balance the payment for these services against other basic needs. Because of the known relationship between income and access problems, the analysis has a special focus on the low- and modest-income populations. Using data from a 1999 national survey to look at the adequacy of coverage, the paper examines the following questions: (1) What are the general problems of the insured when they try to get and pay for healthcare services?, (2) How do these problems vary by income, continuity of coverage, and health status?, and (3) Are insured adults who are having difficulty paying for basic needs also having more negative experiences with their health plans?

Data and Methods

The data reported here come from a national survey commissioned by the Commonwealth Fund, designed by the authors of this article and by Diane Colasanto at Princeton Survey Research Associates and conducted by Princeton Survey Research Associates.

A technical advisory panel of survey experts from universities and federal agencies was convened to discuss and advise on the contents of the survey instrument. While many of the questions in the survey instrument were drawn from other national surveys, a substantial number of new items were developed, many based on the feedback received from the technical advisory group. The survey instrument design was also coordinated with a moderator's guide for a series of focus groups with low-wage workers. The design of the 2 instruments complemented each other.

The survey sample drew from a disproportionately stratified random digit sample of telephone numbers that was designed to increase the likelihood of locating low-income respondents. In order to ensure sufficient sample size in the lower-income groups, the survey design oversampled in low-income telephone areas. This may result in an underestimation of the problems faced by low-income adults, as households without telephones were not included in the sample.

The data were weighted in the analysis to adjust for the disproportionate stratification of the sampling design. In addition to the strata weights, individual and household weights were constructed to adjust for variations in the sample relating to region of residence, sex, age, race, education, marital status, employment status, and telephone service interruption. The resulting data are representative of the US, nonelderly, telephone-owning, adult population.

Tables in the text indicate unweighted sample size and weighted distributions for various outcome measures and respondent characteristics. Significance tests for differences among groups indicate where results are significant at 95% or 99% confidence level. To account for the complex sampling design, STATA was used for the analysis and statistical tests.

A total of 5002 nonelderly adults (ages 18 to 64) were surveyed by telephone between January 1999 and May 1999. The response rate was 58%.† At least 15 attempts were made to complete an interview at all sampled telephone numbers. Calls were staggered over different times of the day and days of the week to increase the probability of contacting a potential respondent. Refusals and break-offs were recontacted at least once in an attempt to convert the refusal to a completed interview.

The analysis presented in this article focuses on the experience of those adults insured when surveyed. The resulting analysis sample, restricted to adults who were insured at the time of the survey, includes 4219 nonelderly adults.

All surveys are subject to sampling and nonsampling error. Results may differ from what would be obtained if the whole population of adults had been interviewed. The size of this error varies with the number surveyed and the magnitude of difference in the responses to each question. Possible sources of nonsampling error include nonresponse bias and question-wording and ordering effects. Nonresponse in telephone surveys produces some known biases in survey-derived estimates, because participation tends to vary for different subgroups of the population. The techniques described above, including weighting and staggered callbacks, help to ensure that the sample is representative.

An additional limitation of the study is in the measurement of interruptions in insurance coverage. Eight percent of insured adults are missing on this variable. This may be a limitation in that respondents who are missing on this variable may differ in important ways from those who answered the question.

Results

Despite having insurance, substantial proportions of insured adults reported medical cost burdens and/or went without needed healthcare due to costs. Nearly 1 in 5 insured adults said that there was a time in the past year when they did not have enough money to pay for medical bills, prescription drugs, or other healthcare costs, and 16% had been contacted by a collection agency about medical bills. As shown in Table 1, approximately 1 in 10 insured adults reported that they did not get needed care and similar proportions reported that they went without each of 3 services due to costs: (1) they did not see a doctor when sick; (2) they did not fill a prescription; or (3) they did not follow up on a recommended treatment or diagnostic test because of cost.

Many insured adults were concerned about their ability to meet their basic costs of living as well. One out of every 7 stated there had been a time when they did not have enough money to pay their rent or mortgage; a similar proportion said there had been a time they could not pay electric or heating bills, and 11% recalled a time when they did not have enough money to buy food.

To examine differences by income, we divided the sample into 4 income groups -- roughly approximating income quartiles for the US population. In tables presenting the comparisons by income, tests for statistical significance compare responses within each of the 3 lower-income groups to those of the top income group (annual family incomes over $60,000).

The resulting analysis finds that access and financial problems experienced by insured adults occur with greater frequency among adults with low or modest incomes. Income groups were chosen for this analysis based on the stratification strategy of the sampling design, and we compare all income groups to those with family incomes over $60,000. These income groups are based on income quartiles in the general population. As shown in Table 1, adults with annual incomes below $20,000 were more likely than those with higher incomes to have problems paying medical bills, despite being insured. Respondents with such low family incomes were 9 times more likely than those with incomes of $60,000 or more (45% vs 4%) and > 3 times more likely than those with $35,000-$60,000 income (13%) to state having been unable to pay medical bills within the past year, despite being insured.

Access problems were the most prevalent among low-income insured adults. As shown in Table 1, insured adults with incomes of less than $20,000 were > 3 times more likely than insured adults with income of $60,000 or higher to report a time when they did not get the care they thought they needed. One quarter did not visit a doctor when sick, and similar proportions did not fill a prescription because of the cost or because it was extremely, very, or somewhat difficult to get care when they needed it.

Insured adults with annual incomes of at least $20,000 but less than $35,000 also reported notably high rates of healthcare and medical cost problems despite being insured. About 1 in 4 had been contacted by collection agencies about past due medical bills and more than 1 of 7 reported going without 1 of the 3 general medical or healthcare services surveyed (visits to doctor, prescription drugs, or follow-up treatment and diagnostic tests).

Both groups of insured adults with incomes below $35,000 were at high risk of going without dental care due to costs. Likely reflecting gaps in insurance benefits, 3 in 10 or more said they had gone without needed dental care during the year because of the expense.

Among this insured population, prior spells of uninsurance during the year appear to expose adults to access barriers as well as potential financial hardship. Among those insured when surveyed, 5% reported a time during the preceding 12 months when they had been uninsured. Findings indicate that a lack of continuity of coverage is related to increasing rates of access barriers and financial burdens. As shown in Table 2, respondents who reported a lapse in insurance coverage in the past year but are currently insured were more likely to experience financial problems. Approximately one half had a time in the past year when they were unable to pay their rent or mortgage, electric or heating bills, or buy food. Slightly more than half were unable to pay their medical bills in the past year, and 28% had been contacted by a collection agency about medical bills. On all measures of financial hardship, those with a lapse in insurance coverage in the past year were at least twice as likely as the continuously insured to report difficulties.

Access problems were also more prevalent among those with a gap in coverage. Compared to the continuously insured, those with a gap in coverage were twice as likely to report that it was extremely, very, or somewhat difficult to get care when needed and almost 3 times as likely to report a time when they did not get care they thought they needed.

Financial problems had a greater impact on the ability to get care for respondents with a gap in insurance coverage. As shown in Table 2, they were more likely than the continuously insured to report they had not visited a doctor when sick, had not filled a prescription, or had skipped some recommended treatment or test in the past year due to cost.

Prior studies have found a significant association between health status and income, with those on the bottom portion of the income scale being more likely to rate their health as fair or poor and to report chronic health conditions.[9] This study finds that those with a combination of poor health and low or modest incomes experience substantial problems getting and paying for medical care and the other basic needs of daily life.

To ensure sufficient sample size for the analysis of data about care experiences by health status and income, we combined the 2 upper income groups into 1 group with incomes of $35,000 or more. Table 3 illustrates the interaction between poor health and income by stratifying by income and by contrasting measures of care and cost difficulties among adults in fair or poor health to those in excellent, very good, or good health.

Insured respondents with family incomes under $20,000 and in fair or poor health were generally the most likely to report access and bill problems of all types. However, insured adults in fair or poor health with incomes in the $20,000-$35,000 range were also at notably high risk of going without needed healthcare due to costs and of not being able to pay medical bills or meet daily living expenses. Among the 2 lower-income groups, one third to one half of insured adults with health problems reported access or cost problems across an array of measures.

Of note are the wide health status differentials in the experiences of people in the $20,000-$35,000 income group. On 8 of the 12 measures of access and cost problems, insured respondents in this income group who reported fair or poor health were at least twice as likely to report problems as were those in better health. They were twice as likely as those in excellent/very good/good health to have problems paying rent or mortgage, electric, heating, medical, or grocery bills, and twice as likely to have been contacted by a collection agency about medical bills. They were also at least twice as likely to have problem getting care: 24% reported a time when they did not get needed care, 32% said it was difficult to get care, and 27% had a time in the past year when they did not visit a doctor when sick due to cost.

In the survey, 22% of all insured respondents rated their insurance as only fair or poor. Respondents facing general financial difficulties were less likely to be satisfied with their health plan. As shown in Table 4, respondents who rated their health insurance as poor were 4 times more likely than those giving their plan an excellent rating to have at least 1 problem paying for basic daily living expenses - including rent or mortgage, electric or heating bills, or food. They were also significantly more likely to have problems paying medical bills or to have been contacted by a collection agency about these bills.

Negative healthcare experiences while insured appear to contribute to lower ratings of insurance. Respondents rating their insurance as only fair or poor reported significantly higher rates of access problems: they were 4-9 times more likely to report problems on the access indicators than those giving their plans an excellent or good rating. Those rating their insurance negatively were also more likely to report a time when they were unable to pay medical bills or a time when they had been contacted by a collection agency about bills during the year.

Discussion

If this were an article about people with low or modest incomes and without any insurance, we would perhaps not be surprised to find such high rates of problems with access to care, financial crises, and collection agencies, and higher burdens among sicker patients. But this study is about people who have insurance policies. The most essential notion of insurance is that it will provide protection against financial risk and assurance that we can get healthcare services when we are sick.

We find substantial proportions of adults who have low and modest incomes who are doing what we would like them to do, buying insurance. A major share of these adults struggle to afford these premiums, get the healthcare they need when sick, and balance the potential and actual cost against other basic living expenses. At the same time, their insurance policies apparently are not offering the guarantee of protection when people are sick. While we do not have information on the benefits and coverage limits of the respondents' health plans, the data on health plan ratings appear to reflect this lower quality, with low-income adults notably more likely to give their plans negative ratings. Given their experiences and perceptions of their quality of coverage, it might not seem unreasonable for them to choose to go without coverage.

A number of efforts at the state and federal level are focused on the expansion of coverage lives, without sufficient consideration for the scope of benefits or the needs of more vulnerable populations. If we are to think about expanding access to insurance for people living on low or modest incomes and often living with higher rates of chronic disease, we would do well to consider the comprehensiveness, affordability, and quality of insurance available. The findings indicate that those with annual incomes below the national average, as well as the poor, need better, more comprehensive coverage than do those with adequate savings to absorb unexpected costs.

References

  1. Copeland C. Characteristics of the nonelderly with selected sources of health insurance and lengths of uninsured spells. Employee Benefits Research Institute. Issue Brief. 1998; p 198.

  2. Gable J, Whitmore H, Pickreign J, et al. Employer Health Benefits, 1999. Report to the Kaiser Family Foundation.

  3. Farley PJ. Who are the underinsured? Milbank Mem Fund Q Health Soc. 1985;63:476-503.

  4. Short PF, Banthin JS. New estimates of the underinsured younger than 65 years. JAMA. 1995;274:1302-1306.

  5. Blendon RJ, Donelan K, Hill CA, Carter W, Beatrice D, Altman D. Paying medical bills in the United States: why health insurance isn't enough. JAMA. 1994;271:949-951.

  6. Himmelstein DU, Woolhander S. Care denied: US residents who are unable to obtain needed medical services. Am J Public Health. 1995;85:341-344.

  7. Cook J, et al. State-specific prevalence estimates of uninsured and underinsured persons -- behavior risk factor surveillance system, 1995. MMWR Morb Mortal Wkly Rep. 1998;47:51-55.

  8. Donelan K, Blendon RJ, Schoen C, Davis K, Binns K. The cost of health system change: public discontent in five nations. Health Affairs. 1999;18:206-216.

  9. Weissman J, Epstein A. Falling Through the Safety Net. Baltimore, Md: Johns Hopkins University Press; 1994.