• New for this release are estimates for enrollment in high deductible health plans (HDHPs) and consumer-directed health plans (CDHPs) that combine an HDHP with a health savings account, as well as estimates of participation in flexible spending account (FSA) arrangements for medical expenses.
• From January - September 2007, 17.5% of persons under 65 years of age with private health insurance were enrolled in a HDHP, 4.5% were enrolled in a CDHP, and 16.9% were in a family with a FSA for medical expenses.
•  From January - September 2007, 43.7 million persons of all ages (14.7%) were uninsured at the time of the interview, 54.5 million (18.4%) had been uninsured for at least part of the year prior to the interview, and 31.2 million (10.5%) had been uninsured for more than a year at the time of the interview.
• The percentage of children under the age of 18 years who were uninsured at the time of the interview was 9.2% in the first 9 months of 2007.
• In the first 9 months of 2007, more than 54% of unemployed adults and almost 23% of employed adults aged 18-64 years had been uninsured for at least part of the past year, and about 33% of unemployed adults and almost 14% of employed adults had been uninsured for more than a year.
•  Since 1997, there has been a decreasing trend in the percentage of children who were uninsured at the time of interview.

Introduction

The Centers for Disease Control and Prevention's (CDC) National Center for Health Statistics (NCHS) is releasing selected estimates of health insurance coverage for the civilian noninstitutionalized U.S. population based on data from the January - September 2007 National Health Interview Survey (NHIS), along with comparable estimates from the 1997-2006 NHIS. Three types of measures of lack of health insurance coverage are provided: current (uninsured at the time of the interview), intermittent (uninsured at least part of the year prior to the interview, which also includes persons uninsured for more than a year, and long term (uninsured for more than a year at the time of the interview). Different timeframes are used to measure lack of insurance coverage to reflect different policy-relevant perspectives. The measure of current lack of coverage provides an estimate of persons who at any given time may experience barriers to obtaining needed health care. The estimate of persons who were uninsured at any time in the year prior to interview provides an annual caseload of persons who may experience these barriers. This measure includes persons who have insurance at the time of interview, but had a period of noncoverage in the year prior to the interview, as well as those who are currently uninsured who may have had been uninsured for a long period of time. Finally, the measure of lack of coverage for more than a year provides an estimate of those with a persistent lack of coverage who may be at high risk of not obtaining preventive services as well as care for illness and injury. These three measures of lack of coverage are not mutually exclusive, and a given individual may be counted in more than one of the three measures.

The 2007 health insurance estimates are being released prior to final data editing and final weighting to provide access to the most recent information from NHIS. The resulting estimates for persons without health insurance are generally 0.1-0.3 percentage points lower than those based on the editing procedures used for the final data files. Occasionally, due to decisions made for the final data editing and weighting, estimates based on preliminary editing procedures may differ by more than 0.3 percentage points. Due to small sample sizes, estimates based on less than a year of data may have large variances. Caution should therefore be used in comparing estimates with large variances. Patterns described in this report may change as more data become available.

NCHS imputed income for approximately 30% of NHIS records from 1997-2006 and will also be imputing income for approximately 30% of NHIS records in 2007. However, beginning with survey year 2007, the imputation procedure will be modified to take into account the changes made to the income section of the questionnaire. Each year, the imputed income files are released within a few months after the annual release of NHIS microdata. Therefore, Early Release health insurance estimates stratified by poverty are based only on reported income and may differ from similar estimates produced later (e.g., in Health, United States (1)) that are based on both reported and imputed income.

Estimates for 2007 are stratified by age group, sex, race/ethnicity, poverty, marital status, employment status, region, and educational attainment. New for this release are estimates of three measures of consumer-directed private healthcare. Consumer-directed healthcare enables individuals to have more control over when and how they access care, what types of care they use, and how much they spend on health care services. National attention to consumer-directed health care has increased following the enactment of the Medicare Prescription Drug Improvement and Modernization Act of 2003 (P.L. 108-173), which established tax-advantaged health savings accounts (2). In 2007, three additional questions were added to the health insurance section of the NHIS to monitor enrollment in consumer-directed health care among persons with private health insurance. Estimates for 2007 are provided for enrollment in high deductible health plans (HDHPs), plans with high deductibles coupled with health savings accounts also known as CDHPs, and being in a family with a flexible spending account (FSA) for medical expenses not otherwise covered. For a more complete description of consumer-directed health care see the "Technical Notes" section of the report.

Data Source

The data are derived from the Family Core component of the 1997-2007 NHIS, which collects information on all family members in each household. Data analyses for the January - September 2007 NHIS were based on 53,334 persons in the Family Core. NHIS is a multistage probability sample survey of the civilian noninstitutionalized population of the United States. It is conducted continuously throughout the year for NCHS by interviewers of the U.S. Census Bureau. Visit the NCHS Website for more information on the design, content, and use of NHIS.

NHIS is a comprehensive health survey that can be used to relate health insurance coverage to health outcomes and health care utilization. It has a low item nonresponse rate (about 1%) for the health insurance questions. Because NHIS is conducted throughout the year, yielding a nationally representative sample each week, data can be analyzed weekly or quarterly to monitor health insurance coverage trends.

A new sample design was implemented beginning with the 2006 NHIS. In addition to the continued over sampling of black and Hispanic persons carried out in the NHIS according to the 1995-2005 sample design, persons of Asian descent are also over sampled according to the new sample design. The impact of the new sample design on estimates presented in this report is expected to be minimal. For a more complete description of the estimation procedures and definitions of selected terms, see the "Technical Notes" section of this report.

In the 2007 NHIS, the family income section of the questionnaire was redesigned to improve the collection of income information and to reduce the amount of income nonresponse. Questions used to collect income data from respondents who initially would not indicate the amount of their family's income in the last calendar year were changed from using a flash card approach to using a set of unfolding brackets. The unfolding bracket method asks a series of closed-ended income range questions (e.g., "is it less than $50,000?") if the respondent did not provide and answer to an exact income amount question. Currently, studies are being conducted to examine the impact of the redesigned income section. Initial evaluations of the distribution of poverty among selected demographic variables in the first quarter of 2007 suggests that poverty estimates are generally comparable to years 2006 and earlier (3). For a more complete discussion of the changes to the income section, see the "Technical Notes" section of this report. Readers should note that as a result of the changes in questions, the 2007 poverty ratio variable has fewer missing values compared to prior years. This reflects improved income item response rates.

This report is updated quarterly and is part of the NHIS Early Release (ER) Program, which releases updated selected estimates that are available from the NHIS website.

Two additional reports are published via the ER Program. Early Release Estimates Based on Data from the National Health Interview Survey (4) is published quarterly and provides estimates of 15 selected measures of health including insurance coverage.  Wireless Substitution: Early Release of Estimates from the National Health Interview Survey (5) is published in May and December and provides
selected estimates of telephone coverage.
 

Results

Lack of health insurance coverage

From January - September 2007, the percentage of persons uninsured at the time of the interview was 14.7% (43.7 million) for persons of all ages, 16.7% (43.4 million) for persons under the age of 65 years, 19.6% (36.7 million) for persons aged 18-64 years, and 9.2% (6.8 million) for children under the age of 18 years (Tables 1 and 2).  There was no significant change in the percentage of people under age 65 who were uninsured at the time of the interview between 2006 and the first 9 months of 2007.

Based on data from the January - September 2007 NHIS, a total of 54.5 million (18.4%) persons of all ages were uninsured for at least part of the year prior to the interview (Tables 1 and 2).  Working-age adults were almost twice as likely to experience this lack of coverage (23.9%) as children under the age of 18 (12.9%).  The percentage of children uninsured during at least part of the year prior to the interview decreased from 18.1% in 1997 to 12.9% in the first 9 months of 2007 ( Figure 1).  However, since 2003, the percentage of children uninsured during at least part of the year prior to interview has remained relatively stable and has ranged from 12.6% to 13.7%. The observed decrease in the percentage of children who lacked coverage for at least part of the year prior to the interview, from 13.0% in 2006 to 12.9% in the first 9 months of 2007, was not statistically significant.  Between 1997 and September 2007 the percentage of adults aged 18-64 who lacked coverage for at least part of the year prior to the interview fluctuated between 22.1% and 24.1% (Table 1).   

Data from January - September 2007 also revealed that 11.9% (31.1 million) of persons under age 65 (14.7% of adults and 5.1% of children) had been uninsured for more than a year (Tables 1 and 2).  Children were less likely than working-age adults to be uninsured for more than a year (Figure 2).  Among adults ages 18-64 years, the percentage uninsured for more than a year was significantly higher in the first 9 months of  2007 (14.7%) than 1997 (13.3%). By contrast, the percentage of children uninsured for more than a year decreased from 8.4% in 1997 to 5.6% in 2002 (Figure 1).  Since 2002, the percentage of children uninsured for more than a year has remained relatively steady and has ranged between 5.1% and 5.6%.

Public and private coverage

From January - September 2007, 18.4% of persons under 65 years of age were covered by public health plans at the time of the interview, and 66.4% were covered by private health insurance plans (Table 3). More than two-thirds (69.3%) of adults 18-64 years of age were covered by a private plan compared with 59.0% of children under 18 years (Figure 2).  One-third of children (33.3%) were covered by a public plan, compared with 12.4% of adults 18-64 years.  There was no significant change in the percentage of children or working age adults covered by public plans from 2006 to the first 9 months of 2007. However, private health care coverage rates among both children and working-age adults are now significantly lower than in 1997.

Insurance coverage by poverty status

From January - September 2007, 12.3% of poor children and 16.2% of near-poor children (see "Technical Notes" for definition of poverty) did not have health insurance coverage at the time of interview (Table 4). The percentage of poor children who were uninsured at the time of the interview generally decreased from 1997 through September 2007.  The observed changes in the percentage or poor and near poor children uninsured at the time of interview from 2006 to the first 9 months of 2007 were not significant. From 1997 to September 2007, the percentage of poor adults who were uninsured remained relatively stable.

Based on the January - September data, 77.8% of poor children and 54.2% of near-poor children were covered by a public health plan at the time of interview (Table 5).  Public coverage for near poor children increased from 49.0% in 2006 to 54.2% in the first 9 months of 2007. The observed increase in public coverage for poor children between 2006 (75.8%) and the first 9 months of 2007 (77.8%) was not significant. Among children, all poverty groups experienced an increase in public coverage between 1997 and the first 9 months of 2007.  However, the largest increase was seen among near-poor children (Table 5 and Figure 3) for whom coverage by a public plan more than doubled between 1997 and the first 9 months of 2007.

Based on data from January - September 2007, 38.2% of poor working-age adults were covered by public coverage.  However, this was not a significant increase from the 35.6% of poor working-age adults covered by a public plan in 2006 (Table 5).   From January - September 2007, 12.0% of poor children and 32.5% of near-poor children were covered by private health insurance at the time of interview (Table 6).  Private coverage decreased for near-poor children from 36.9% in 2006 to 32.5% in the first 9 months of 2007.  The rate of private coverage among near-poor children was approximately 23 percentage points lower in September 2007 than in 1997.  Among poor adults aged 18-64 years, 23.0% were covered by private health insurance in the first 9 months of 2007.   Private coverage significantly decreased among near poor working age adults from 52.6% in 1997 to 39.6% in the first 9 months of 2007. Among this group, the percentage with private coverage significantly decreased from 42.6% in 2006 to 39.6% in the first 9 months of 2007. Among not-poor adults aged 18-64 years, the percent with private coverage did not change between 2006 and the first 9 months of 2007 (83.6%). From 1997 to 2000, however, private coverage among not poor adults has decreased from 88.2% in 1999 to 83.6% in the first 9 months of 2007.

Lack of coverage, by selected demographic characteristics

Race/ethnicity

Based on data from the January-September 2007 NHIS, Hispanic persons were considerably more likely than non-Hispanic white persons, non-Hispanic black persons, and non-Hispanic Asian persons to be uninsured at the time of interview, to have been uninsured for at least part of the past 12 months, and to have been uninsured for more than a year (Table 7). Approximately one-third of Hispanic persons were uninsured at the time of interview or had been uninsured for at least part of the past year, and more than one-fourth of Hispanic persons had been without health insurance coverage for more than a year.

Age and sex

For persons under 65 years of age, for both sexes combined, the percentage of persons who were uninsured at the time of the interview was highest among those aged 18-24 years (28.1%) and lowest among those under 18 years (9.2%) (Figure 4). Starting at age 18, younger adults were more likely than older adults to lack health insurance coverage. Among adults in age groups 18-24 years, 25-34 years, 35-44 years, and 45-64 years, men were more likely than women to lack health insurance coverage at the time of interview.

Other demographic characteristics

Lack of health insurance coverage was greatest in the South and West regions of the United States.  Among adults who lacked a high school diploma, 30.9% were uninsured at the time of interview, 34.6% were uninsured for at least part of a year, and 25.5% had been uninsured for more than a year at the time of the interview (Table 7).  These rates are 2 to almost 4 times as high as those for persons with more than a high school education.  Among currently unemployed adults 18-64 years old, 54.4% had been uninsured for at least part of the past year, and 32.7% had been uninsured for more than a year.  Among employed adults 18-64 years of age, 22.6% had been uninsured for at least part of the past year and 13.9% had been uninsured for more than a year.  Married adults were more likely to have coverage than those who were divorced, separated, living with a partner, or never married.

Consumer-directed healthcare

Based on data from the January - September 2007 NHIS, 17.5% of persons under 65 years of age with private health insurance were enrolled in a HDHP, 4.5% were enrolled in a CDHP, and 16.9% were in a family with a FSA for medical expenses (Figure 5).  Adults aged 18-64 with more than a high school diploma were more likely to be covered by a HDHP, more likely to be covered by a CDHP and more likely to be in a family with an FSA for medical expenses than those who had only a high school diploma or were not high school graduates (Figure 6).

References

1. National Center for Health Statistics. Health, United States, 2006 with chartbook on trends in the health of Americans. Hyattsville, MD. 2006.

2. United States Government Accountability Office. Consumer-directed health plans: Early enrollee experiences with health savings accounts and eligible health plans. GAO-06-798. Washington. 2006.

3. Plies JR and Cohen RA. Impact of income bracketing on poverty measures used in the National Health Interview Survey's (NHIS) Early Release Program: Preliminary data from the 2007 NHIS. December 2007. Available from: http://www.cdc.gov/nchs/about/major/nhis/methods.htm.

4. Schiller, J, Barnes P and Heyman KM, Early release of selected estimates based on data from the January - September 2007 National Health Interview Survey. National Center for Health Statistics. March 2007. Available from: http://www.cdc.gov/nchs/nhis.htm.

5. Blumberg SJ, Luke JV. Wireless substitution: Early release of estimates based on data from the National Health Interview Survey, January - June 2007. National Center for Health Statistics. December 2007 Available from: http://www.cdc.gov/nchs/nhis/htm.

6. Cohen RA, Martinez ME. Impact of Medicare and Medicaid probe questions on health insurance estimates from the National Health Interview Survey, 2004. Health E-stats. Hyattsville, MD. National Center for Health Statistics 2005.

7. Lamison-White L. Poverty in the United States, 1996. Current population reports, series P60-198. U.S. Bureau of the Census.  Washington. 1997.

8. Dalaker J, Naifeh M. Poverty in the United States, 1997. Current population reports, series P60-201. U.S. Bureau of the Census. Washington. 1998.

9. Dalaker J.  Poverty in the United States, 1998. Current population reports, series P60-207. U.S. Census Bureau.  Washington. 1999.

10.  Dalaker J, Proctor B. Poverty in the United States, 1999. Current population reports, series P60-210.  U.S. Census Bureau. Washington. 2000.

11. Dalaker J. Poverty in the United States, 2000. Current population reports, series P60-214. U.S. Census Bureau. Washington. 2001.

12. Proctor B, Dalaker J. Poverty in the United States, 2001. Current population reports, series P60-219. U.S. Census Bureau. Washington. 2002.

13. Proctor B, Dalaker J. Poverty in the United States, 2002. Current population reports, series P60-222. U.S. Census Bureau. Washington. 2003.

14. DeNavas-Walt C, Proctor B, Mills R.  Income, poverty, and health insurance coverage in the United States: 2003. Current population reports, P60-226. U.S. Census Bureau. Washington. 2004.

15. DeNavas-Walt C, Proctor B, Lee C.  Income, poverty, and health insurance coverage in the United States: 2004. Current population reports, P60-229. U.S. Census Bureau. Washington. 2005.

16. DeNavas-Walt C, Proctor B, Lee C.  Income, poverty, and health insurance coverage in the United States: 2005. Current population reports, P60-231. U.S. Census Bureau. Washington. 2006.

17. DeNavas-Walt C, Proctor B, Smith J.  Income, poverty, and health insurance coverage in the United States: 2006. Current population reports, P60-233. U.S. Census Bureau. Washington. 2007.

18. Pleis JR, Dahlhamer JM, Meyer PS. 2007. Unfolding the answers? Income nonresponse and income brackets in the National Health Interview Survey. Proceedings of the 2006 Joint Statistical Meetings [CD-ROM]. American Statistical Association, Alexandria, Virginia.

Technical Notes

Estimation procedures

NCHS creates survey weights for each calendar quarter of the NHIS sample.  The NHIS data weighting procedure has been described in more detail elsewhere (view/download PDF). Estimates were calculated by using the NHIS survey weights, which are calibrated to census totals for sex, age, and race/ethnicity of the U.S. civilian noninstitutionalized population.  The weights for the 1997-1999 NHIS data were derived from 1990-census-based population estimates.  Weights for the 2000-2007 NHIS data were derived from 2000-census-based population estimates.

Point estimates and estimates of their variances were calculated using SUDAAN software to account for the complex sample design of NHIS. The Taylor series linearization method was chosen for variance estimation.  All estimates shown meet the NCHS standard of having less than or equal to 30% relative standard error unless otherwise noted.  Differences between percentages or rates were evaluated using two-sided significance tests at the 0.05 level. Logistic regression was used to evaluate the significance of time trends in the percentage of health insurance coverage.  Terms such as "greater than" and "less than" indicate a statistically significant difference. Terms such as "similar" and no difference indicate that the estimates being compared were not significantly different.  Lack of comments regarding the difference between any two estimates does not necessarily mean that the difference was tested and found to be not significant.  Due to small sample sizes, estimates based on less than a year of data may have large variances, and caution should be used in analyzing these estimates.  Patterns for such estimates may change as more data become available.

Definitions of selected terms

 Health insurance coverage - The "private health insurance coverage" category excludes plans that paid for only one type of service such as accidents or dental care.  The "public health plan coverage" category includes Medicaid, State Children's Health Insurance Program (SCHIP), state-sponsored or other government-sponsored health plan, Medicare, and military plans.  A small number of persons were covered by both public and private plans and were included in both categories.  A person was defined as "uninsured" if he or she did not have any private health insurance, Medicare, Medicaid, SCHIP, state-sponsored or other government-sponsored health plan, or military plan at the time of the interview.  A person was also defined as uninsured if he or she had only Indian Health Service coverage or had only a private plan that paid for one type of service such as accidents or dental care.  The analyses excluded persons with unknown health insurance status (about 1% of respondents each year).

Data on health insurance status were edited using an automated system based on logic checks and keyword searches.  For comparability, the estimates for all years were created using these same procedures. The resulting estimates of health insurance coverage are generally within 0.1-0.3 percentage points of those based on the editing procedures used for the final data files. However, occasionally differences greater than 0.3 percentage points have been observed between preliminary estimates and estimates based on final files.

In 2004, the NHIS instrument recorded a much larger than expected proportion of respondents with a family income for the previous calendar year of "$2," and those responses were used in calculating some estimates in previous releases of this report. Based on a subsequent extensive review, these "$2" responses were coded to "not ascertained" for the final microdata files. For a more complete discussion, see the 2004 Survey Description Document. For this report, a decision was made to re-run the 2004 estimates to reflect this editing decision.  Estimates impacted by this decision are found in those tables that present estimates stratified by poverty (Tables 4, 5, and 6). The estimates primarily impacted were those for poor persons and may differ from those released in previous reports.

The terms HIKIND, MCAREPRB, and MCAIDPRB refer to questions on NHIS. The data on type of health insurance are collected through the HIKIND question "What kind of health insurance or health care coverage does _______ have? INCLUDE those that pay for only one type of service (nursing home care, accidents, or dental care), exclude private plans that only provide extra cash while hospitalized." Respondents can indicate private and public plans or indicate that they or family members are not covered by insurance.

Two additional questions were added to the health insurance section of NHIS beginning with the third quarter of 2004. One question, MCAREPRB, was asked of persons aged 65 years and over who had not indicated that they had Medicare. The MCAREPRB question is: "People covered by Medicare have a card which looks like this. {Are/Is} {person} covered by Medicare?" The other question, MCAIDPRB, was asked of persons under the age of 65 years who had not indicated any type of coverage. The MCAIDPRB question is: "There is a program called Medicaid that pays for health care for persons in need. In this state it is also called {state name}. {Are/Is} {person} covered by Medicaid?"

Respondents who were considered insured at the time of interview were asked about periods of noncoverage in the past year. For persons who did not have health insurance at the time of interview, a question concerning the length of time since the respondent had coverage was asked. These questions were used in the estimates of intermittent and long-term uninsurance.

Respondents who originally classified themselves as uninsured, but whose classification was changed to Medicare or Medicaid on the basis of a "yes" response to either probe question, subsequently received appropriate followup questions concerning periods of noncoverage for insured respondents.

Method 1 estimates are based solely on one question (HIKIND). Method 2 estimates are based on responses to three questions (HIKIND, MCAREPRB, and MCAIDPRB). Prior to 2004, estimates in earlier releases of this report were generated using Method 1. Estimates for 2004 are presented using Method 2 in the figures and both Method 1 and Method 2 in the tables. Estimates using Method 1 for the "uninsured for more than a year" measure excluded persons whose classification was changed from uninsured to either Medicare or Medicaid coverage by either additional probe question. As a result, these respondents did not receive the followup question concerning how long had it been since they had coverage. However, they were asked the questions concerning noncoverage in the past 12 months for insured persons. All estimates beginning in 2005 are calculated using the two additional questions (Method 2).  Of the 892 people (unweighted) who were eligible to receive the MCAREPRB question in the third and fourth quarters of 2004, 55.4% indicated that they were covered by Medicare. Of the 9,146 people (unweighted) who were eligible to receive the MCAIDPRB question in the third and fourth quarters of 2004, 3.0% indicated that they were covered by Medicaid.

From July through December 2004 (the third and fourth quarters combined), the estimates (weighted) for the "uninsured at the time of interview" measure decreased from 10.4% to 9.9% for persons under 18 years of age, from 19.7% to 19.5% for adults 18-64 years of age, and from 1.7% to 1.2% for persons 65 years of age and over with the use of Method 2 as compared with Method 1. The estimates for public coverage increased from 28.1% to 29.6% for children under 18 years of age, from 11.3% to 11.4% for adults 18-64 years of age, and from 89.5% to 93.3% for persons 65 years of age and over with the use of Method 2 as compared with Method 1. The tabulation of "uninsured for more than a year" using Method 1 excludes respondents whose classification was changed to Medicare or Medicaid by either probe question because these individuals did not receive the question concerning duration of noncoverage for persons who are uninsured. There is no impact of the two additional questions on the estimates for private coverage. Additional information on the impact of these two probe questions on health insurance estimates can be found in Impact of Medicare and Medicaid probe questions on health insurance estimates from the National Health Interview Survey, 2004 (6).

Consumer-directed health care - In 2007 three additional questions were added to the health insurance section of the NHIS questionnaire to monitor consumer-directed health care. For persons with private health insurance, a new question regarding the annual deductible of each private health insurance plan was added. For plans considered to be High Deductible Health Plans (HDHPs), a follow-up question was asked regarding special accounts or funds used to pay for medical expenses: a health savings account (HSA) or a Health Reimbursement Account (HRA). Lastly, a new question about enrollment in a flexible spending account (FSA) for medical expenses in the family was added. Three measures of consumer-directed health care are presented in this report: estimates for enrollment in high deductible health plans (HDHPs), enrollment in CDHPs, and being in a family with a flexible spending account (FSA) for medical expenses.

High Deductible Health Plan (HDHP) - A HDHP was defined as a private health plan with an annual deductible of not less than $1,100 for self-only coverage or $2,200 for family coverage.

Consumer-Directed Health Plan (CDHP) - A CDHP was defined as a HDHP with a special account to pay for medical expenses; unspent funds are carried over to subsequent years. A person was considered to have a CDHP if there was a "yes" response to the following question: "With this plan, do you have a special account or fund you can use to pay for medical expenses? The accounts are sometimes referred to as Health Savings Accounts (HSAs), Health Reimbursement Accounts (HRAs), Personal Care Accounts, Personal medical funds, or Choice funds, and are different from Flexible Spending Accounts."

Flexible spending account (FSA) for medical expenses - A person was considered to be in a family with an FSA if there was a "yes" response to the following question: "Do you or anyone in your family currently have a Flexible Spending Account for health expenses? These accounts are offered by some employers to allow employees to set aside pre-tax dollars of their own money for their use throughout the year to reimburse themselves for their out-of pocket expenses for health care. For this type of account, any money remaining in the account at the end of the year, following a short grace period is lost to the employee".

The three measures presented in this report of consumer-directed health care are not mutually exclusive. Therefore, a person may be counted in more than one measure.

Education - The categories of education are based on the years of school completed or highest degree obtained for persons aged 18 years and over. Only years completed in a school that advances a person toward an elementary or high school diploma, General Educational Development high school equivalency diploma (GED), college, university, or professional degree are included. Education in other schools or home schooling is counted only if the credits are accepted in a regular school system.

Employment - Employment status is assessed at the time of interview and is obtained for persons aged 18 years and over. In this release, it is presented only for persons aged 18-64 years.

Hispanic or Latino origin and race - Hispanic or Latino origin and race are two separate and distinct categories. Persons of Hispanic or Latino origin may be of any race. Hispanic or Latino origin includes persons of Mexican, Puerto Rican, Cuban, Central and South American, or Spanish origins. Race is based on the family respondent's description of his or her own race background as well as the race background of other family members. For conciseness, the text, tables, and figures in this report use shorter versions of the 1997 Office of Management and Budget (OMB) race and Hispanic or Latino origin terms. For example, the category "not Hispanic or Latino, black or African American, single race" is referred to as "non-Hispanic black, single race" in the text, tables, and figures. Estimates for non-Hispanic persons of races other than white only, black only, Asian only, or of multiple races are combined into the "other races or multiple race" category.

Poverty status - Poverty categories are based on the ratio of the family's income in the previous calendar year to the appropriate poverty threshold (given the family's size and number of children) defined by the U.S. Census Bureau for that year (7-17). Persons who are categorized as poor had a ratio less than 1.0 (i.e., their family income was below the poverty threshold), near-poor persons have incomes of 100% to less than 200% of the poverty threshold, and not poor persons have incomes that are 200% of the poverty threshold or greater.  The remaining group of respondents is coded as "unknown" with respect to poverty status. The percentage of respondents with unknown poverty status (19.1% in 1997, 23.6% in 1998, 26.4% in 1999, 27.0% in 2000, 27.1% in 2001, 28.1% in 2002, 31.5% in 2003, 29.6% in 2004, 28.9% in 2005, 30.7% in 2006 and 17.9% in the first three quarters of 2007) is disaggregated by insurance status and age in Tables 4, 5, and 6.   See the NHIS Survey Description Documents for 1997-2006 for more information on the unknown income and unknown poverty status categories .

In the 2007 NHIS, the family income section was redesigned to improve the collection of income information and to reduce the amount of income nonresponse. Questions used to collect income data from respondents who initially would not indicate the amount of their family's income in the last calendar year were changed from using a flash card approach to using a set of unfolding brackets. The unfolding bracket method asks a series of closed-ended income range questions (e.g., "is it less than $50,000") if the respondent did not provide and answer to an exact income amount question.

Prior to 2007, the "Income and Assets" section in the Family Core of the NHIS instrument allowed respondents to report their family income in several ways. Respondents were first asked to provide their combined family income before taxes from all sources for the previous calendar year in a dollar amount (from $0 up to $999,995). Respondents who did not know or refused to state an amount were then asked if their combined family income in the previous calendar year was $20,000 or more or less than $20,000. If they again refused to answer or said that they did not know, they were not asked any more questions about their family income. Respondents who did reply to the "above-below $20,000" question were then handed a list (flash card) of detailed income categories (top-coded at $75,000 or more) and asked to pick the interval containing their best estimate of their combined family income. Thus, NHIS respondents fell into one of four categories with respect to combined family income information: those willing to supply a dollar amount, those who indicated their combined family income from a fairly detailed set of intervals (approximately 65%), those who said that their combined family income was either $20,000 or more or less than $20,000, and those unwilling to provide any information whatsoever. However the flash card approach had a very low item response rate (15-18%) and this led to income variables such as poverty ratio having relatively high levels of missing data (approximately 30% item nonresponse rate).

Based on results of a pilot test conducted in the second quarter of 2006, the 2007 NHIS income follow-up questions were changed from using a flash card approach to using a set of unfolding brackets. These questions utilize a series of income intervals, and respondents answering the complete path of questions would answer either 2 or 3 questions. The first follow-up income question asks a respondent if the family's income is less than $50,000. With this as a starting point, for respondents who answered "yes" to this question, additional questions are asked to determine if the family's income is below $35,000 and if the family's income is below the poverty threshold. Alternately, for respondents who answered "no" to the less than $50,000 question, additional questions are asked to determine if the family income is below $100,000 and below $75,000. A more detailed description of the pilot test is available elsewhere (18). Based on preliminary data from the first quarter of 2007, 56% of the respondents eligible for the modified income follow-up questions answered all the questions in the applicable path. A more complete analysis of the impact of bracketing on poverty measures is published elsewhere (3). Readers should note that as a result of the changes in income questions the 2007 poverty ratio variable has fewer missing values compared with prior years. This reflects the improved income item response rates.

NCHS imputed income for approximately 30% of NHIS records from 1997-2006 and will also be imputing income for approximately 30% of NHIS records in 2007. However, beginning with survey year 2007, the imputation procedure will be modified to take into account the changes made to the income section. The imputed income files are released within a few months after the annual release of NHIS microdata. Therefore, Early Release health insurance estimates stratified by poverty are based only on the reported income and may differ from similar estimates produced later (e.g., in Health, United States (1)) that are based on both reported and imputed income.

Region-In the geographic classification of the U.S. population, states are grouped into the following four regions used by the U.S. Census Bureau:

Region               States included

Northeast         Maine, Vermont, New Hampshire, Massachusetts, Connecticut, Rhode Island, New York,

                            New Jersey, and Pennsylvania

Midwest            Ohio, Illinois, Indiana, Michigan, Wisconsin, Minnesota, Iowa, Missouri, North

                            Dakota, South Dakota, Kansas and Nebraska

South                 Delaware, Maryland, District of Columbia, West Virginia, Virginia, Kentucky,

                            Tennessee, North Carolina, South Carolina, Georgia, Florida, Alabama,

                            Mississippi, Louisiana, Oklahoma, Arkansas, and Texas

West                  Washington, Oregon, California, Nevada, New Mexico, Arizona, Idaho, Utah, Colorado,

                            Montana, Wyoming, Alaska and Hawaii

Future plans

The NCHS Early Release Program will continue updating and releasing estimates of health insurance coverage 6 months after NHIS data collection has been completed for each quarter. In addition to this special report focusing on health insurance, the Early Release Program releases estimates for 15 selected measures of health, including lack of health insurance coverage and type of coverage, usual place to go for medical care, obtaining needed medical care, influenza vaccination, pneumococcal vaccination, obesity, leisure-time physical activity, current smoking, alcohol consumption, HIV testing, general health status, personal care needs, serious psychological distress, diagnosed diabetes, and asthma episodes and current asthma (3). Health insurance estimates are available in both Early Release reports. A third Early Release report that examines telephone coverage in the United States is published biannually (4). New measures may be added as work continues and in response to changing data needs. Feedback on these releases is welcome (e-mail).

Announcements about Early Releases, other new data releases, publications, or corrections related to NHIS will be sent to members of the HISUSERS Listserv. To join, visit the CDC website.

Suggested citation

Cohen RA, Martinez ME, and Free, HL. Health insurance coverage: Early release of estimates from the National Health Interview Survey, January-September 2007. March 2008. Available from: http://www.cdc.gov/nchs/nhis.htm