Skip directly to search Skip directly to A to Z list Skip directly to page options Skip directly to site content
Scientific Data Documentation
Data From 1987 NMES Household Survey Projected to 1995

Data From the 1987 NMES Household Survey Projected to 1995:
Medical Expenditures by Type of Service and Payment Source, Demographics, and Poverty and Insurance Status

Center for Cost and Financing Studies
Agency for Health Care Policy and Research
Executive Office Building, Suite 500
2101 East Jefferson Street
Rockville, Maryland 20852
(301) 594-1400


This documentation describes the NMS95EXP file, issued by the Agency for Health Care
Policy and Research (AHCPR), that contains data based on the National Medical Expenditure
Survey (NMES). The file consists of person-level medical expenditure data classified into service
categories and separated into sources of payment within each category, a person weight that makes
national-level estimates possible, and person-level demographic characteristics. The NMES data
are from 1987 and have been projected to 1995 by adjusting a NMES person weight and aligning
per capita NMES expenditures to adjusted per capita expenditures in the national health accounts
(NHA) which are published by the Health Care Financing Administration (HCFA). Data on the
NMS95EXP file cannot be linked to other NMES data files that are not projected.

The NMS95EXP file is a person-level data file with one record for each of 34,459 persons,
designed to provide estimates that are representative of the civilian non-institutionalized population
of the United States in 1995. Each record is distinguished by a person identification number.

Medical Expenditure Categories

Each person record has data on expenditures made in 17 different medical service
categories. The 17 medical service categories and their abbreviated names are:

1. Hospital Room and Board Expenditures: HOSF

2. Inpatient Physician Expenditures: HOSM

3. Emergency Room Expenditures: EROM

4. Physician Office Expenditures: OUTP

5. Outpatient Hospital Expenditures (no mental health or chiropractor): PHYT

6. Chiropractor Expenditures (ambulatory plus outpatient): CHIR

7. Podiatrist Expenditures (ambulatory plus outpatient): PODI

8. Optometrist Expenditures (ambulatory plus outpatient): OPTM

9. Outpatient Mental Health (ambulatory and hospital outpatient): OMNH

10. Drug Expenditures: PMED

11. Orthodontia Expenditures: ORTH

12. Other Dental Expenditures: ODEN

13. Glasses and Contacts: GLAS

14. Durable Medical Equipment: DURE

15. Nondurable Medical Supplies: NOND

16. Home Health: HOME

17. Total Expenditures: TOT

The expenditures for each of the 17 service categories are divided into the following 10 sources
of payment (SP1-SP10):

SP1: Self or family (out of pocket)

SP2: Private insurance

SP3: Medicare

SP4: Medicaid

SP5: Other Federal sources--includes CHAMPUS, CHAMPVA, Supplemental
Security Income (SSI), Indian Health Service (IHS) facility or contract,
Intertribal Council, Alaska Native Corporation, Veteran Administration, and any
military and other Federal programs such as free government screening services
and care at the National Institutes of Health (NIH)

SP6: Other State source--includes services such as community health centers but
excludes local and State employment-related insurance and welfare programs

SP7: Workers' compensation

SP8: Other private sources--includes automobile insurance, other types of insurance
not specified, company (where the company is not the respondent's insurer or
employer), school (where the school is not the insurer or employer), union
(where the union is not the insurer or employer), friend, foreign government,
or not otherwise specified

SP9: Free from provider

SP10: Total value for each service category

The expenditure variable names consist of the abbreviated service category name (HOSF,
HOSM, etc.) and a suffix indicating the source of payment (SP1-SP10). As an example of the
variable naming convention, the 10 hospital room and board expenditure variables are: HOSFSP1,
HOSFSP10. To help clarify any confusion there may be concerning the relationship between
service categories and sources of payment, Table 1 is constructed with service categories on the
left margin and sources of payment across the top. This creates a table with 170 cells, where each
cell would contain the dollar value contributed from the source of payment for a particular service.

Analysts must be cautious when using the expenditures reported in the "free-from-
provider" source of payment variable. This is because there is a "double counting" issue that must
be considered. To the extent that "free care" is financed through higher fees charged to patients
paying through insurance or out-of-pocket, including imputed dollar values for free-from-provider
as a source of payment counts some health care "spending" twice. However, some services
rendered free from provider are financed by philanthropy or sources of non-patient revenue (such
as parking fees at a hospital). Therefore, some free-from-provider "payments" should be included
in total expenditures, if one is to obtain an accurate value. Unfortunately, we are not able to
separate the revenue from philanthropy and non-patient sources from the other free from provider
expenditures. On the file, the total expenditure values in each of the service categories includes
the imputed free-from-provider values, so there is some double counting in the totals. Analysts
have the option of using the total values on the file or subtracting free-from-provider expenditures
and using the results for totals.

Projecting Expenditure Data From 1987 to 1995

To obtain the expenditure values on the file, NMES data had to be "aged" from 1987 to
1995. This aging was accomplished by aligning NMES data previously released on Public Use
Tape 18 to the most recent projections of the NHA. Alignment of NMES expenditures to NHA
expenditures requires that the types of health care services recorded in NMES match those in the
national health accounts. NMES identifies more categories of service than the national health
accounts, so the alignment process required combining several NMES service types into a smaller
number of categories consistent with those reported in the NHA. Table 2 shows how NMES
service categories from Public Use Tape 18 (which differ to some extent from the service
categories on the NMS95EXP file) were collapsed into the nine types of health care spending
reported in the NHA.

The expenditures in the service categories on the NMS95EXP file were constructed from
NMES data previously released on Public Use Tape 18. The expenditure service categories on
NMS95EXP are different from those shown in Table 2, but the total sum of the NMES
expenditures from NMS95EXP is equal to that calculated from the NMES expenditure categories
shown on Table 2. The NMES expenditures that comprise the 17 service categories on the

Table 1 Source of Payment Category

Table 1 Source of Payment Category
Self or
Other Fed
Other State

Table 2 NMES Household Survey Expenditure Categories and Corresponding National Health Account Expenditure Categories
NMES Service Category NHA Service Category Collapsed Category Number
Outpatient, Facility Hospital 1
Emergency Room Hospital 1
Inpatient, Facility Hospital 1
Outpatient, Facility Physician 2
Physician Office Visits Physician 2
Physician Phone Calls Physician 2
Inpatient, Physician Physician 2
Physician Home Visit Physician 2
Dental Visit Dental 3
Outpatient, Non-Physician Other Professional 4
Non-Physician Office Visits Other Professional 4
Non-Pysician Phone Calls Other Professional 4
Home Health Care Home Health 5
Prescribed Drugs Drugs & Nondurables 6
Nondurable Medical Goods Drugs & Nondurables 6
Durable Medical Goods Durable Medical Goods 7
Eyeglasses & Contact Lenses Durable Medical Goods 7
(No NMES Equivalent) Nursing Home 8
(No NMES Expenditures) Other Personal Health 9
Total Expenditures Total Personal Health  

Note: The NMES expenditure categories are from NMES expenditure data previously released on
Public Use Tape 18. They do not match the service categories in Table 1.
NMS95EXP file were aged by multiplying them by aging factors that were calculated by aligning
the NMES service categories in Table 2 to the NHA service categories in Table 2. A detailed
description of how these aging factors were created is provided below. In aging a NMES
expenditure variable used to construct the 17 service categories on NMS95EXP, we used the aging
factor from the NMES service category in Table 2 that most closely matches it. Because we do
not provide the link to Public Use Tape 18 data, users cannot reconstruct the NMES categories
in Table 2 with expenditure data provided on NMS95EXP. This means that users cannot replicate
our creation of the aging factors for the seven NHA expenditure categories used to produce the
1995 expenditure estimates on NMS95EXP.

Because NMES data do not include spending for administration of health care services,
certain government public health programs, research, and construction programs, we excluded
those expenditures from the national health accounts when doing the alignment. Moreover,
personal health expenditures, as measured in the national health accounts, include some
expenditures--for nursing home care and some other personal health care services--that are not
covered by NMES. Also, hospital care in the NHA includes expenditures for some hospitals that
are outside the scope of the NMES universe, and the NHA and NMES measure drug expenditures
differently. Adjustments were made to the national health accounts' expenditure data in these four
service categories--hospital, nursing home, other personal health care and drugs--to conform to
NMES expenditure categories. A description of these adjustments follows.

Adjustments to NHA Hospital Expenditures

In both the NHA and the NMES, hospital care includes inpatient services, outpatient
services, and emergency room services. The NMES universe of hospitals, however, includes only
short-term community hospitals, and the NMES expenditures include only patient revenues in
these hospitals. Data on non-patient revenues, such as parking fees or State and local subsidy
payments, are not included in the NMES. The NHA, in contrast, include Federal hospitals,
psychiatric hospitals, and other long-term hospitals and include all sources of revenue received by
these hospitals. In order to make NHA hospital expenditures conform to those in the NMES,
NHA expenditures for hospitals not in the NMES universe and non-patient revenues of community
hospitals were subtracted from the NHA. In 1987 such expenditures were $29.1 billion of the
$194.1 billion in total NHA hospital expenditures. For years after 1987, it was assumed that
these expenditures represented the same proportion of total NHA hospital expenditures as in 1987.

Nursing Home Residents and Expenditures

NMES and NHA are based on different populations, making it impossible to align both
total and per capita expenditures. The NMES Household Survey sample represents the civilian,
non-institutionalized population of the United States (note that the NMES does include a separate
survey of the institutionalized population). In contrast, the NHA includes several groups not
represented in the NMES Household Survey: residents of nursing homes; residents of personal
care homes and intermediate care facilities for persons with mental retardation (ICF/MR); persons
in the armed forces; inmates of prisons and jails; persons in psychiatric hospitals; and residents
of United States territories and possessions. We chose to align per capita NMES expenditures to
per capita NHA expenditures, so the issue of the different populations represented by the two
samples must be addressed. Because the per capita expenditures for acute health care services
used by the nursing home and ICF/MR (institutionalized) population are significantly higher than
the average for the community population, we subtracted the nursing facility and ICF/MR
populations from the NHA population, and we subtracted an estimate of acute care expenditures
on behalf of these persons from total NHA expenditures. We also subtracted all nursing home
expenditures from the NHA. Other populations, and their acute care expenditures, that are
accounted for in the NHA but not in the NMES were not subtracted from the NHA totals, because
their per capita expenditures are more likely than those of institutionalized persons to approximate
the mean expenditures of the community population.

Subtracting acute care expenditures for nursing home and ICF/MR residents consisted of
removing an estimate of hospital and physician expenditures for these populations. In order to
benchmark to previous HCFA estimates, we did not remove other expenditures from the NHA
for these populations. The estimates of spending on acute health care services for the nursing
home population in the NHA were based on the mean hospital and physician expenditures, as
measured by NMES, of persons living in the community who were age 65 and older and had
limitations in three or more activities of daily living. This value was aged from 1987 to 1995
using the consumer price index for medical care. This index indicates the price change for a
basket of medical care goods and services. For the ICF/MR population, we used the NMES 1987
mean total acute care expenditures for the entire civilian, community population, also inflated to
1995. Using the overall mean may underestimate acute care expenses for the ICF/MR population,
but it is not likely to affect the results significantly because they accounted for only about 7 percent
of the institutionalized population in 1987.

Other Personal Health Care

The expenditures in the other personal health care service category in the NHA are
predominantly (84.5 percent) paid by a variety of public programs. More than half of these
public expenditures are paid for by Medicaid, but are not included in any other NHA expenditure
category. These include health screening services, some home and community-based waiver
services, case management services, and other services. NMES has no equivalent category to the
NHA "other personal services" category, but most of these Medicaid-funded services fall into one
of the other NMES service categories. To account for these expenditures in the allocation process,
we re-allocated six-tenths of the NHA other personal health care public expenditures equally into
three other NHA categories: physician, other professional, and home health. The remaining
expenditures in this category--comprising mostly Public Health Service, Indian Health Service,
and industrial health service expenditures--are not within the scope of the NMES Household
Survey, and were subtracted from the NHA for the NMES/NHA reconciliation.

Prescription and Nonprescription Drug Expenditures

The NHA category "drugs and other medical nondurables" includes expenditures for both
prescription and nonprescription drugs as well as nondurable medical goods. This differs from
the NMES which, in addition to expenditures for nondurable medical goods, collects expenditure
data only for prescription drugs. In order to make NHA drug and nondurable medical
expenditures consistent with the NMES, we subtracted an estimated amount of expenditures for
nonprescription drugs from the NHA total. In 1993, spending for prescription drugs accounted
for 65 percent of NHA expenditures in this category. An estimate of the fraction of the remainder
that is accounted for by nonprescription drugs was not available. Moreover, the proportion
accounted for by prescription drug expenditures in this category has risen rapidly since 1987,
while the portion of total national health expenditures comprised of nonprescription drugs and
nondurable medical goods has remained relatively constant, in the range of 9.6 to 10 percent.
Based on the data available for 1993, we estimated that in 1995, 65 percent of expenditures in the
NHA drugs and nondurable medical goods category were for prescription drugs. Lacking detailed
information on the distribution of the remainder, we divided it equally between nonprescription
drugs and other nondurable medical goods, and then subtracted 17.5 percent of the total NHA
expenditures from this category to make it consistent with the NMES definition of drugs and
nondurable goods.

To assure analysts that the expenditures in the service categories in NMS95EXP align to
the adjusted expenditures in the NHA, we have constructed Table 3 in which the per capita
expenditure values from the NMS95EXP service categories are displayed with the per capita
adjusted expenditures from the NHA service categories. Table 3 also shows how the NMS95EXP
and NHA service categories can be collapsed into expenditure categories that comprise the same
sets of services. The table shows that the per capita expenditures in the collapsed NMS95EXP
cells are very close to the per capita expenditures in the collapsed NHA cells. Note that we
collapsed to six service categories here, whereas we collapsed to seven categories for the
alignment, as shown in Table 2.

Issues Regarding Sources of Payment

Alignment of NMES expenditures to NHA expenditures also involved allocating by source
of payment. Specifically, after partitioning expenditures by service category, they were further
partitioned by source of payment within each service category. This resulted in a matrix of
expenditures by type of service and source of payment for each of the two data sets. For each of
these cells there is a factor to align the NMES expenditure to the NHA expenditure.

Table 3 Comparison of Per Capita Uncollapsed and Collapsed NMS95 EXP and NHA Expenditures
NMS95EXP Service Category Per Capita NMS95EXP Expenditure Collapsed Service Category Per Capita Collapsed NMS95EXP Expenditure NHA Service Category Per Capita NHA Expenditures Collapsed Service Category Per Capita Collapsed NHA Expenditures
HOSF 1013.04 Hospital 1071.16 Hospital 1069.1 Hospital 1069.1
EROM 58.12 Hospital 1071.16 Physician 686.5 Phys/Oth Prof 915.6
HOSM 231.90 Phys/Oth Prof 929.63 Other Prof 229.1 Phys/Oth Prof 915.6
OUTP 552.75 Phys/Oth Prof 929.63 Dental 156.8 Dental 156.8
PHYT 71.16 Phys/Oth Prof 929.63 Durables 51.5 Durables 51.5
CHIR 28.24 Phys/Oth Prof 929.63 Drugs/Nondur 259.6 Drugs/Nondur 259.6
PODI 9.24 Phys/Oth Prof 929.63 Home Health 105.1 Home Health 105.1
OPTM 4.80 Phys/Oth Prof 929.63        
OMNH 31.53 Phys/Oth Prof 929.63        
ORTH 36.29 Dental 156.53        
ODEN 120.23 Dental 156.53        
GLAS 29.72 Durables 52.02        
DURE 22.29 Durables 52.02        
PMED 248.00 Drugs/Nondur 262.07        
NOND 14.07 Drugs/Nondur 262.07        
HOME 107.23 Home Health 107.23        

Note: The NMS95EXP expenditures do not include free-from-provider values.
As with type of service, there are more sources of payment in NMES than in the NHA. The
NMES has nine sources of payment, excluding the total category, which are defined above, and
NHA has eight sources of payment: out of pocket, private health insurance, other private source,
Medicare, Medicaid (Federal share), other Federal source, Medicaid (State share), and other State
or local sources. The NHA out of pocket payment source is essentially the same as the NMES self
or family payment source.

For this reconciliation exercise, we combined the two NHA Medicaid categories (Federal
and State) and the other Federal and other State or local categories into a single "other
Government" category. We also subtracted expenditures paid for by other private sources from
each NHA service category. In the NHA, other private sources of payments represent primarily
charity and philanthropy not directly attributable to services for specific patients and other
nonpatient revenues. There is no directly corresponding equivalent to these sources in NMES,
so this source of payment had to be subtracted from the NHA expenditures to perform the
alignment. Likewise, we subtracted from NMES expenditures in each service category all
expenditures from the free-from-provider source of payment because service values for this
category are not imputed for health care services in the NHA.

To align NMES with the remaining five NHA source-of-payment categories, we combined
the NMES other Federal, other State, and workers' compensation payments into a single category
called other Government. We also allocated the NMES other private source of payment category
among two other sources of payment: self or family and private health insurance. The NMES
other private payment source consists of payments made by automobile insurance, other
unspecified types of insurance, and miscellaneous private sources. In the NHA, payments from
these sources are allocated between out of pocket and private health insurance. The allocation of
the NMES other private source of payment between self or family (out of pocket) and private
health insurance was based on the relative proportion of each of these two payment sources in each
type-of-service category. Table 4 shows how the NMES and NHA source of payment categories
were collapsed prior to the actual alignment. From each source of payment within the NHA
hospital and physician service categories, we subtracted approximations of expenditures for the
nursing home and ICF/MR population. The total hospital and physician expenditures that were
adjusted for this population as described earlier were divided by the original totals in these service
categories, and then these ratios were used to prorate across each source of payment cell of the
hospital and physician service categories.

Table 4 NMES Household Survey Source-of-Payment Categories and Corresponding National Health Account Source-of-Payment Categories
NMES Source of Payment Category NHA Source of Payment Category Collapsed Category Number
Self or Family Out of Pocket 1
Private Insurance Private Insurance 2
Medicare Medicare 3
Medicaid Medicaid (1) 4
Other Federal Other Federal 5
Other State Other State 5
Worker's Compensation No NHA Equivalent (2) 5
Other Private No NHA Equivalent(3) 1.2
Free from Provider Other Private Source (4) ---

1/ Combines the NHA payment source categories Medicaid, Federal and Medicaid, State.
2/ Health expenditures under workers' compensation consist of medical benefits paid under
public law by private insurance carriers and self-insured firms. They are treated as
public expenditures in the national health accounts.
3/ NMES other private expenditures are allocated among self or family and private health
insurance to match their treatment in the NHA.
4/ NMES free from provider and NHA other private expenditures are subtracted from their
respective totals because neither represents payment for direct patient services.

Aging the Data

After making these adjustments, we split the population into subpopulations of people
under age 65, and age 65 and over. Using 1994 unpublished expenditures from the Congressional
Budget Office (CBO), we calculated the expenditures in each cell of the NHA expenditure matrix
that apply to the under-65 population and the expenditures that apply to the 65+ population.
This was done by multiplying the NHA expenditures in each cell by both the CBO under 65 and 65+
expenditure proportions to obtain approximations of the NHA expenditures for the two populations.
Because the NMES data includes the person's age, expenditures for these two populations can be
obtained directly from the NMES data. Per capita values were calculated for each cell in the
NMES and NHA expenditure matrices for both the under 65 and 65+ populations. Next, we divided
the cell values from the NHA matrices by the corresponding cell values from the NMES matrices
after re-weighting the NMES data to 1995 (described below). These are the factors used to
to complete the aging of the NMES expenditure data and to align the aged data to the NHA bench-
marks. We aged the NMES expenditure data that appear on the file by multiplying 136 of the
170 medical expenditure values on each person-record by the appropriate aging factor. The
34 expenditure values not aged in this fashion were composed of (1) all of the service category
expenditures that were free from provider and (2) all the service category expenditures paid by
other private sources. The free from provider and other private source expenditures were,
however, aged by factors unique to each service category. To obtain the free-from-provider
factors, for each service category we calculated sums of the NHA and NMES per capita
expenditures for the five collapsed sources of payment shown in Table 4. The NHA sum was
divided by the NMES sum, and this result was multiplied by the NMES free-from-provider value
for that service category. This result was the aged free-from-provider expenditure value. To
obtain the other private source aging factor within each service category we calculated a weighted
sum of the aging factors for self or family and private health insurance expenditures within that
service category. The weights are the relative proportions of each of these two sources of payment
in the service category.

As we did for the service categories, we have included Table 5 which shows that our aging
process does produce results consistent with the benchmark NHA data. Table 5 has the same
collapsed service categories as Table 3, but partitions expenditures in each service category by the
collapsed source of payment categories in Table 4. Table 5 shows that the per capita expenditure
values from NMS95EXP are very close to the per capita expenditures from the NHA. Because
the actual alignment was conducted for the two separate age groups, cumulative rounding effects
account for any discrepancies between the NMES and NHA estimates. Each cell in Table 5 has
the per capita expenditure from NMS95EXP for that cell, and the analogous per capita expenditure
from the NHA is in parentheses.

Person-Level Weight

Each person-record includes a person-level sampling weight needed to produce estimates
for the civilian, noninstitutionalized population of the U.S. in 1995. This weight, SIMWT95C,
was derived from a NMES 1987 weight which was adjusted primarily through a weighting class
procedure to reflect demographic changes in the U.S. population between 1987 and 1995. The
NMES weight was first adjusted to represent the 1994 U.S. population by using data from the
March 1994 Current Population Survey (CPS). In this process, records from both the NMES and
CPS were partitioned into matrices defined by a set of eight person-level characteristics which
are included on this file: receipt of welfare cash assistance, age, race, gender, income relative
to the poverty line, primary source of health insurance, employment status of the family head,
and Census region. Each NMES weight in a given cell was then multiplied by the ratio of the sum
of the March 1994 CPS weights to the sum of the 1987 NMES weights for that cell, creating a new
weight, SIMWT94. This weight was then adjusted so that the number of persons

Table 5 Collapsed Source of Payment
Collapsed Service Category Out of Pocket
Private Insurance
Other Government
Hospital 24.24
Physician & Other Professional 178.42
Dental 81.05
69.13 (69.44) 0.00
Home Health 15.73
Drugs & Nondurables 178.13
Durables 33.34
Total 510.91

Note: The collapsed service categories are the same as the collapsed categories in Table 3,
and the collapsed sources of payment are the same as the collapsed sources in the last
column of Table 4. The values in parentheses are per capita expenditures from the NHA,
and the expenditures above those are per capita expenditures from the NMS95EXP file.
Reporting participation in Aid to Families with Dependent Children (AFDC), Supplemental
Security Income (SSI), and other Medicaid on the re-weighted NMES matched the number of
recipients of these benefits according to 1994 administrative records. This was done to
compensate for the under-reporting of these benefits on the CPS and produced a new weight,

We also needed to capture the increased enrollment in health maintenance organizations
(HMOs) since 1987. To do so, the NMES under-65 population with private coverage was
partitioned into matrices using the same eight person-level characteristics used in the first re-
weighting described above, plus an indicator of HMO participation. SIMWT94B for the people
in each cell of these matrices was post-stratified to match the HMO enrollment of the under-65,
privately insured population as reported in the 1993 National Health Interview Survey (NHIS).
Those outside this population did not have SIMWT94B adjusted. The result of this step was a new
weight, SIMWT94C. We found that among people with both Medicare and private insurance
coverage on the re-weighted NMES file, the ratio of people with employer-provided insurance to
those with individually purchased insurance did not match the ratio based on data from the
Medicare Current Beneficiary Survey for calendar year 1992 (MCBS). Experts on the Medicare
population informed us that the MCBS provides a more accurate representation of this population
than the CPS; as a result, we post-stratified SIMWT94C so that the proportion with employer-
provided coverage and the proportion with individually purchased insurance matches the
proportions found on the MCBS. No adjustment was made to the SIMWT94C values for persons
outside this population. Each SIMWT94C was aged to 1995 by using the U.S. Census Bureau
projected population figures on gender, age and race. This final adjustment produces the weight
provided on the NMS95EXP file, SIMWT95C.

There is a caveat with using SIMWT95C to obtain estimates of the number of people who
have HMO coverage because only the under-65 population with private coverage was included in
the post-stratification to account for increased HMO coverage. There were too few persons 65
and over with HMO coverage in 1987 to post-stratify this population as was done for the under-65
population. Therefore, analysts can get an estimate of the total number of people under age 65
who have private HMO coverage, but they cannot get an estimate of the total number of people
overall (including those with public coverage and over age 65) who have HMO coverage. In both
the NMES and NHIS, an HMO was defined as either a group or staff model prepaid health plan
or an independent practice association (IPA). Another caveat when using SIMWT95C to obtain
totals for a population with a particular set of characteristics is that some of the cells in the
reweighting matrices had very small numbers of observations, making estimates based on those cells
unreliable. This is a particularly important issue for the population receiving welfare cash
assistance, because when creating SIMWT94 and SIMWT94C, we divided the population into two
groups based on this characteristic, and the number of people receiving cash assistance is small.
To remedy this problem, we established a minimum cell size of 20 records for each cell, except
in some cells for the population receiving welfare cash assistance where the minimum was set to
10 records. As a result, some cells were collapsed along one or more variables. For example,
we generally used 10 age categories in the creation of SIMWT95C for people without cash
assistance, but used a variable with the 10 age categories collapsed to 3 categories for people
with cash assistance.

Demographic, Geographic, and Insurance Variables

The definition and construction of the person-level characteristics used to create
SIMWT95C: age, race, gender, income relative to the poverty line, primary source of health
insurance, employment status of the family head, region of residence, receipt of welfare cash
assistance, HMO participation status, and Medicare/private insurance status are described below.
These variables are included on the NMS95EXP file.

The partitioning of the NMES records during the creation of SIMWT94 was defined by
the following person-level variables:

CASHASST: Indicates participation in cash-assistance welfare programs. It was created
with NMES data on income from participation in SSI, AFDC, and other public
assistance programs. The formatting of this variable is:

1 = SSI
2 = AFDC or Other Public Assistance
3 = Neither

NEWIHINS: Primary source of health insurance. This variable was constructed from NMES
data on age, coverage through employment-related insurance, coverage through
non-employment-related insurance, coverage through CHAMPUS, Medicare participation,
and Medicaid participation. The data on age, CHAMPUS coverage, Medicare
participation and Medicaid participation are from the last round of participation
in the NMES Household Survey. The data on employment-related and non-employment-related
insurance are from household- and employer-reported insurance data. Persons under 65
are hierarchically coded into one of the first four categories, unless they have both
CHAMPUS and private coverage that is not employment related, in which case they are
coded into category 2. All persons over 65 are coded into category 5. This variable
is formatted as follows.

1 = Private, employment related or CHAMPUS
2 = Private, not employment related
3 = Public (Medicare or Medicaid)
4 = Uninsured
5 = Age 65 or older

POVCAT2B: Family income relative to poverty line. This variable is based on a NMES variable
that classifies people as living in poor, near poor, low income, middle income,
or high income families, or in families having negative income. This NMES variable
is based on the income of all related persons residing in the same household. The
variable is formatted as:

1 = Less than 100% of poverty level
2 = 1.01 to 1.24 times poverty level
3 = 1.25 to 1.99 times poverty level
4 = 2.00 to 3.99 times poverty level
5 = 4.00 or more times poverty level

Families with negative income were included in category 1.

AGE3: A three-category age variable based on age data from the last round of
participation in the NMES Household Survey. All records have a value for this variable,
but it was used to post-stratify on age only for persons identified as receiving cash
assistance (CASHASST = 1 or 2). The variable is formatted as:

1 = 0-17
2 = 18-64
3 = 65 or older

AGE10: A ten-category age variable based on age data from the last round of participation
in the NMES Household Survey. All records have a value for this variable, but it was
used to post-stratify on age only for persons not getting any cash assistance (CASHASST
= 3). The variable is formatted as:

1 = 0-4
2 = 5-14
3 = 15-24
4 = 25-34
5 = 35-44
6 = 45-54
7 = 55-64
8 = 65-74
9 = 75-84
10 = 85 or older

SEX: Gender as reported in NMES.

1 = Male
2 = Female

NEWRACE: A person-level variable constructed to facilitate the post-stratification of the
sampling weight by race and ethnicity based on three mutually exclusive classifications:
Hispanic; Black, non-Hispanic; and other.

1 = Hispanic
2 = Black, Non-Hispanic
3 = Other

EMPFAMHD: A constructed variable that cross-classifies employer insurance premium
contributions and employment status for the head (eldest person) of each family.
Employment status is as of the week of the last round of participation in the NMES
Household Survey. In coding this variable, all family members were assigned the same
category as the family head. This variable is formatted as follows:

1 = No job (last week)
2 = Employed and employer makes health insurance contributions
3 = Other

REGION: Identifies the Census region (Northeast, Midwest, South, and West) in which
a person resided as of the last round of participation in the NMES Household Survey.

1 = Northeast
2 = Midwest
3 = South
4 = West

The NMES variable used in the adjustment of SIMWT94 for the under-reporting of AFDC, SSI,
and Medicaid participation on the March 1994 CPS was:

CASHAST9: A hierarchical variable that indicates whether or not a person is a participant
in SSI, AFDC, or other Medicaid at any point in the NMES survey. It was created with
NMES data on participation in those programs and the person's age as of the last round
of participation in the NMES Household Survey. The formatting of this variable is as

1 = AFDC, age < 18
2 = AFDC, age 18
3 = SSI, age < 65
4 = SSI, age 65
5 = Medicaid, no SSI or AFDC, age < 21
6 = Medicaid, no SSI or AFDC, 21 age < 65
7 = Medicaid, no SSI or AFDC, age 65
8 = No SSI/AFDC/Medicaid, age < 65
9 = No SSI/AFDC/Medicaid, age 65

Be aware that for this variable, the AFDC population does not include the AFDC recipients who
also receive other public assistance. In comparison, category 2 of the variable CASHASST does
include these persons.

SIMWT94B was created by partitioning the population on the file into the nine categories in
CASHAST9. The sum of the new weights for the people in these categories was calculated. The
sum in each category was divided by the total number of people in that category as reported on
administrative records. Each person's SIMWT94 was multiplied by the ratio from his or her
CASHAST9 category to create SIMWT94B.

Next, SIMWT94B was post-stratified to match HMO enrollment of persons under age 65
with private insurance, as reported in the 1993 NHIS. In both the NMES and NHIS, an HMO was
defined as either a group or staff model prepaid health plan or an independent practice association
(IPA). This population was partitioned by the same personal characteristics used in the creation
of SIMWT94, plus:

LASTHMO: Indicates whether a person was covered by a private health insurance plan
that was classified as an HMO in the last round of participation in the NMES survey.
Multiple data sources were used to determine whether a health plan was an HMO. The
primary source was data supplied by employers and insurance companies collected in the
NMES Health Insurance Plans Survey (HIPS). For plans not included in HIPS, the NMES
Household Survey was used to ascertain HMO status. HMO data were collected in several
places in the Household Survey: the enumeration questionnaire, the core questionnaire
(plan and insurance care information), the billing sections of the events booklets (why no
bill), and the access supplement (usual source of care is an HMO). If any of the data
suggested HMO enrollment and the person was only covered by one plan, the plan was
classified as an HMO plan. If after reviewing all of the data sources for that person (and
in some cases, for other family members), no indication of HMO enrollment was found,
all plans associated with the person were classified as non-HMO. For persons covered by
more than one plan, plan names were reviewed in order to determine which plans would
be classified as HMOs. The variable LASTHMO is formatted as follows:

0 = No
1 = Yes

For people under 65 with private insurance, their values of SIMWT94B were post-stratified
according to these partitions, resulting in the weight SIMWT94C. Persons outside this population
have SIMWT94C equal to SIMWT94B. Charges imputed to health services used by the HMO
population based on charges by the fee-for-service population in the 1987 NMES were carried
forward in the aged NMES data base.

Finally, for the over-65 Medicare population with private insurance coverage, SIMWT94C
was adjusted so that the ratio of the number of people with Medicare who have employer-
sponsored insurance to the number of people with Medicare who have individually purchased
insurance matches the ratio in the Medicare Current Beneficiary Survey. The Medicare population
was partitioned by:

MEDCARE4: Indicates whether a person has Medicare coverage only, Medicare and
Medicaid coverage, Medicare coverage with employment-sponsored private insurance,
Medicare coverage with other private insurance, or no Medicare coverage. A person is
considered to be covered by Medicare if the NMES Medicare participation variable
indicates Medicare coverage in any round of the survey. If a person was classified as
having Medicare coverage and had Medicaid or other public coverage on the interview date
in any round of participation in NMES, the person was assigned to category 1. If that
person was classified as being covered by Medicare and reported having group coverage
through an employer or union in any round of NMES, they were assigned to category 2.
If classified as being covered by Medicare and in any round of NMES had group coverage
that was not through an employer or union, or had non-group coverage, that person was
assigned to category 3. If classified as being covered by Medicare and had no other
coverage, that person was assigned to category 4. If not classified as being covered by
Medicare, that person was assigned to category 5. This variable is coded hierarchically
in the order listed below.

1 = Medicare and Medicaid
2 = Medicare and employer or union group insurance
3 = Medicare and individually purchased insurance
4 = Medicare only
5 = No Medicare

SIMWT94C remained the same for the remaining sample.

The variables AGE10, NEWRACE, and SEX were used to adjust SIMWT94C to 1995
levels on age, sex, and race projected by the U.S. Bureau of the Census, creating SIMWT95C for
each record on the file.

Variance Estimation

Sample design variables are also provided (sampling stratum, and PSU within stratum),
to allow the user to estimate the variance associated with projected estimates that is a consequence
of the underlying sample design of the NMES data. Variance estimates of descriptive statistics
derived from these data, using conventional procedures that focus on measuring sampling
variability, generally understate the overall variance. This is a consequence of the influence of
additional sources of error introduced to the NMES data through aging and alignment to external
data sources. The variance estimates attributable to the NMES sample design, however, should
serve as a lower bound guidepost.

Strategies for Estimation

This file is constructed to allow for estimation of health care expenditures and insurance
coverage at the person level. In order to produce estimates related to these data, the values in each
record contributing to the estimates must be multiplied by the sampling weight SIMWT95C. We
have purposely omitted person identification numbers that would enable direct linkage of the
projected NMES data to previously released NMES data files which are not aged, to prohibit users
from making projections for population groups not statistically controlled for in the aging process.

This page last reviewed: Thursday, January 28, 2016
This information is provided as technical reference material. Please contact us at to request a simple text version of this document.