User Note - Summary and Key Insurance Variables
- I. Guidance on Using IPUMS NHIS Insurance Variables
- II. Ipums NHIS Insurance Variables by Type and Source of Coverage, Over Time
- III. IPUMS NHIS Insurance Variables Harmonizing Information about Reasons for No Insurance Over Time
- IV. Additional Variables Of Interest To Health Insurance Analysts
I. Guidance on Using IPUMS NHIS Insurance Variables
One of the ways in which IPUMS has made the National Health Interview Survey data easier to use is the creation of health insurance variables that 1) summarize information about health care coverage across multiple plans and 2) are comparable over time, enabling analyses of trends in health insurance coverage. This user note is intended to act as a guide to selecting the appropriate IPUMS-created variables for specific health insurance analyses. This user note does not intend to delineate which insurance variables should be used for every type of analysis.
In general, the variables included in the "Insurance (summary and key)" group are either recoded from variables found in the "Insurance (other)" group, or included because they can be used to help monitor the Affordable Care Act. For example, a summary variable may be a recode of responses to questions on multiple private insurance plans (HIPRIVATEE), a recode indicating any public coverage from multiple questions on public insurance coverage (HIPUBCOVE), or a recoded category for why a person's coverage stopped (HINOUNEMPR-HINOTHER).
The "Insurance (other)" group includes variables that provide information for up to four private insurance plans, specific responses as to why a person's coverage stopped, and many measures that were included in surveys going back to 1963 but are no longer collected. Most of the variables discussed below can be accessed through the "Insurance (summary and key)" drop-down menu.
II. IPUMS NHIS Insurance Variables by Type and Source of Coverage, over Time
Type of Coverage
Users who are interested in tracking coverage type over time can use one of two sets of variables: one spanning the 1997-present period; the other covers selected years before 1997 in addition to the 1997-present period. The key difference between the two sets of variables, highlighted in the table below, is that the variables available for 1997-present were edited by NCHS staff to enhance accuracy. In contrast, the variables available for both the pre-1997 period and the 1997-present period were created by IPUMS staff based on variables not edited for accuracy so as to be comparable across the pre-1997 and 1997-present periods.
Unedited versions of the health insurance type variables are also available for 1997-forward. An example of the type of editing undertaken by NCHS staff includes changing the reported type of health insurance coverage based on the verbatim health insurance plan name provided by respondents. For example, respondents who indicated they had private health insurance coverage but gave the name of a state Medicaid health plan will have their response edited to reflect that they actually have Medicaid coverage. The unedited version of this variable would still indicate that the respondent had private health insurance coverage. For users who want to know more about health insurance variable editing, please refer to specific variable descriptions and to our user note on changes in health insurance data collection.
|Type of Coverage||Variable Name(s)|
|1997-forward||Selected pre-1997 years and 1997-forward|
|Any public insurance other than Medicare||Users can either use HIPUBCOVE (1) or combine the different types of public coverage: HIMCAIDE, HISTATEE, HICHIPE (2), HIOTHGOVE||HIPUBCOV|
|No insurance||HINOTCOVE (3)||HINOTCOV|
|Single service plans||Any single service plan: SINGLEE (2), and single service dental plan, SDENTALE (2)||Any single service plan, SINGLE, and single service dental plan, SDENTAL|
Notes: (1) HIPUBCOV and HIPUBCOVE do not take into account any military (TRICARE (CHAMPUS), CHAMP-VA, and VA coverage). Users can decide whether they consider Military coverage to be public or private and recode accordingly. It is common for TRICARE to be considered employer-sponsored coverage and for 2004 forward, users can incorporate HITRICARE into their recode. (2) = Available for 1999 forward. (3) The NCHS considered persons who only had coverage through Indian Health Insurance program to be uninsured. Consequently, persons with only IHS coverage are coded to "Yes" for "HINOTCOVE", has no coverage.
Source of Coverage
Researchers interested in conducting analyses of health insurance-related topics may also be interested in examining trends in the source of health insurance (e.g. whether the person is the policyholder or a dependent on another person's employer health plan). IPUMS created several summary variables to consolidate and harmonize information about how persons obtained private coverage. Summary variables on the source of health insurance coverage include: any private insurance obtained through employment (HIPWORKR), any private insurance plan paid in part or full by employer (HIPEMPAYR), and any private insurance plan purchased directly (HIPBUYOWNR). These variables are offered for selected pre-1992 samples and for every sample from 1992 to the present. Other insurance variables on the source of health insurance coverage include: type of plan (HIP1TYPE), and (HIP2TYPE), and whose name the plan belongs (HIP1WHO)and (HIP2WHO).
III. IPUMS NHIS Insurance Variables Harmonizing Information about Reasons for No Insurance over Time
Reasons for No Insurance
Several summary variables were created to consolidate and harmonize the variables describing why persons do not have insurance. This consolidation collapsed multiple reasons for not having insurance into broader categories that facilitate use of variables that have changed name over time, but have the same substantive meaning. These summary variables include: employment (HINOEMPR), unemployment (HINOUNEMPR), family-related reasons (HINOFAMR), aged out of family plan (HINOAGER), cost (HINOCOSTR), was refused or denied coverage (HINOREFUSER), and some other reason (HINOTHER).
Variables capturing reasons for having no insurance are available for select samples prior to 1993 and for every sample from 1993 to the present. However, analysts should be aware of changes in the universe over time. In addition to the variables named above, WHYNOIN provides reasons for having no insurance, but the universe is substantially different: this information was collected for persons without private or Medicare coverage, but persons may have had coverage through Medicaid or a single service plan.
IV. Additional Variables of Interest to Health Insurance Analysts
There are at least an additional two sets of integrated variables related to health insurance coverage collected by NHIS and offered by IPUMS NHIS that may be of interest to researchers: 1) gaps in insurance coverage and 2) out-of-pocket expenses.
Gaps or Changes in Health Insurance Coverage
IPUMS NHIS offers a set of variables reporting whether individuals had any changes in insurance coverage-including changes from one type of insurance to another and periods of uninsurance-and if so, for how long. Unlike the health insurance type and source variables discussed above that are also available in the 1997 forward samples, these were not edited by the NCHS staff to verify accuracy.
For researchers interested in measures of discontinuous coverage, where currently insured people experienced a spell of non-coverage in the past 12 months, there are two sets of IPUMS NHIS variables available: one covering the 1993-1996 period, the other covering the 1997-forward period. For the 1993-1996 period, HINOLAPY reports whether currently insured individuals had a lapse in coverage in the past 12 months and HINOLAPYMO reports how many months the individual was without health care coverage. Prior to 1993, HINOLAPYMO refers to number of months without coverage when uninsurance was specifically due to a job loss. Similarly, for the 1997 forward period, HINOTYR indicates whether currently insured persons were without coverage in the past 12 months (note: armed forces members were specifically excluded in the 1997-1999 samples) and HINOTYRMO captures the number of months the person was without health care coverage.
For those who are interested in changes in coverage that did not involve a loss of health care coverage, (HICHANGEYR), was added to the NHIS in 2011 to monitor the effects of the Affordable Care Act. HICHANGEYR provides information on whether insured respondents with continuous coverage in the past year experienced any changes in their insurance in the previous 12 months.
Several IPUMS NHIS variables provide information about past health insurance experiences for persons who were uninsured at the time of the survey. For persons who were uninsured at the time of the survey, HINOLASTCOV reports how long ago (in years and months) individuals had health care coverage for the 1994-1996 period and HILAST reports the duration since the respondent last had coverage. This variable is available for the 1997-forward period.
IPUMS NHIS offers variables with information on out-of-pocket (OOP) expenditures separately for premium costs and health care services. Out-of-pocket premium costs were collected for each private insurance plan a person holds. IPUMS NHIS does not provide a variable that summarizes OOP premium costs across multiple plans. Users who want OOP premium cost information for each plan can use HIP1COST and HIP2COST, which report the premium cost information for plans 1 and 2, respectively. Note that for 2004 forward, the NCHS discontinued the public release of information for any third or fourth insurance plan a person may have, but these data are available for 1999-2003 (HIP3COST and HIP4COST). A very small number of persons have more than two private plans; the variable HIPRIVGT2 (available for 2004 forward) reports whether a person has more than two private plans. For 1997 and 1998, OOP premium costs are measured in interval groups, rather than as a continuous dollar amount. Intervalled costs for up to four insurance plans are included in the variables HIP1COS2, HIP2COS2, HIP3COS2, and HIP4COS2.
For information about out-of-pocket health care costs other than the cost of health insurance premiums, users can refer to the variable HCSPENDY, which reports the total out-of-pocket spending for medical and dental care, excluding over-the-counter drugs and costs for which individuals expect to be reimbursed. Expense information is offered in interval amounts ranging from less than $500 to more than $5,000. This variable is available from 1997 to the present.