SRC
DATA PURCHASE AND USE AGREEMENT
Individual identifiers have been removed from the micro-data tapes available
from the Agency for Health Care Policy and Research through NTIS.
Nevertheless, under sections 308(d) and 903(c) of the Public Health Service
Act (42 U.S.C. 242m and 42 U.S.C. 299 a-1), data collected by the Agency for
Health Care Policy and Research may not be used for any purpose other than
the purpose for which it was supplied. The information on the micro-data
tapes available for purchase was supplied to the Agency for statistical
summaries and health services research. It is necessary, therefore, that
the individual ordering such micro-data tapes sign the following assurance:
The undersigned gives assurance that individual elementary unit data on the
micro-data tapes being ordered will be used solely for statistical summaries
and health services research.
BACKGROUND
General Information
This documentation describes one in a series of public use
tapes issued by the Agency for Health Care Policy and Research
with data from the National Medical Expenditure Survey. This
survey provides extensive information on health expenditures by
or on behalf of families and individuals, the financing of these
expenditures, and each person's use of services. The National
Medical Expenditure Survey (NMES) is a research project of the
Center for General Health Services Intramural Research, Agency
for Health Care Policy and Research.
Since the 1970s the intramural research program has given
particular emphasis to studies of the use and financing of health
services. The first series of studies (NMES-1) employed data
collected in the 1977 National Medical Care Expenditure Survey.
NMES-1 produced information on a broad range of issues such as
the number and characteristics of the uninsured and the
underinsured, the tax implications of excluding employer-paid
premiums for health insurance from employee income, and the
differences among socioeconomic and demographic groups with
respect to the use of health services.
A new series of studies (NMES-2) was initiated in the 1980s.
These studies also involve a major data collection effort - the
1987 National Medical Expenditure Survey. Like its predecessor,
NMES-2 provides information about the non-institutionalized
population. In addition and in contrast to the earlier studies,
NMES-2 also provides extensive information on the population
residing in or admitted to nursing homes and facilities for the
mentally retarded.
The Household Component of NMES-2 is based on a national
probability sample of the civilian, noninstitutionalized
population living in the community. The sample is designed to
provide a larger representation of population groups of special
policy interest to the Federal Government than would have been
obtained from a random sample. These groups include poor and low
income families, the elderly, the functionally impaired, and
black and Hispanic minorities. A Survey of American Indians and
Alaska Natives includes a separate sample of American Indians and
Alaska Natives living on or near Federal reservations and
eligible to receive care provided or supported by the Indian
Health Service. The Institutional Population Component includes a
sample of persons residing in or admitted to nursing and personal
care homes and facilities for the mentally retarded during 1987.
A separate Medicare Records Component provides claims data on all
Medicare beneficiaries included in the household and
institutional samples.
Together, the major components of NMES-2 contain information to
make national estimates of health status, use of health services,
insurance coverage, expenditures, and sources of payment for the
civilian population of the United States during the period from
January 1 to December 31, 1987. Oversampling of population
groups of special interest makes possible in-depth studies of
these groups. The database can also be used to assess the
implications of recent or proposed changes in public or private
health care benefits, methods of financing both health care and
insurance coverage, various public and private subsidies for
health care, and employee compensation arrangements.
Household Survey
Each family in the Household Survey was interviewed four times
over a period of 16 months to obtain information about the
family's health and health care during calendar year 1987.
Baseline data on household composition, employment, and insurance
were updated at each interview, and information was obtained on
illnesses, use of health services, and health expenditures for
each family member. A fifth round of interviews was conducted in
the spring of 1988 to obtain information on the tax filing and
medical deductions of each household. A long term care
supplement was administered during the first and fourth rounds of
interviewing to permit estimates of persons with functional
disabilities and the use of formal services or long term care
provided by family or friends.
In order to verify and supplement the information provided by
household respondents, the Household Component of NMES-2
included two additional surveys. The Medical Provider Survey
obtained information from the physicians, hospitals, outpatient
clinics, emergency rooms, and home health agencies used by the
household sample during 1987. The Health Insurance Plan Survey
obtained information on the private insurance of persons in the
household sample, including premiums paid by all sources and the
provisions of their coverage.
Survey of American Indians and Alaska Natives (SAIAN)
This component was conducted with the same data collection
instruments and interview procedures as the Household Component
and covered the same reference period, calendar year 1987. SAIAN
also included follow-up surveys to medical providers and health
insurers. Consequently, the data can be used to compare American
Indians and Alaska Natives eligible for care from the Indian
Health Service and the general U.S. population with regard to
such issues as health status, use of health services, and access
to care. Information was obtained on services provided outside
the Indian Health Service and on other sources of health care
financing available for persons eligible for care from the IHS.
Institutional Population Component
The Institutional Population Component of NMES-2 included
persons resident in or admitted to nursing and personal care
homes and facilities for the mentally retarded at any time in
calendar year 1987. This survey provides information on the
functional status, use of services, and health expenditures of
the institutionalized population. The Survey in Institutions
(SII) collected data from facility administrators and designated
staff on the characteristics of facilities and charges. The
Survey of Next of Kin (SNK) obtained data from the respondent's
next-of-kin or other knowledgeable persons in the community on
the financial status, insurance coverage, and personal history of
the institutionalized person.
Survey Samples
All survey components were designed to provide statistically
unbiased estimates that are representative of the civilian
population of the United States in 1987. The Household Survey
sample is a stratified multistage area probability design with a
total sample of roughly 35,000 individuals, in 14,000 households,
who completed all rounds of data collection. Oversampling of the
population subgroups of special policy interest was based on a
separate screening interview conducted in the fall of 1986 with a
sample of 36,000 addresses.
The Survey of American Indians and Alaska Natives adopted a
multistage area probability sample design using an IHS-
constructed frame of counties with individuals eligible for
services provided or supported by the Indian Health Service and
living on or near Federally recognized reservations or in Alaska.
An initial screening interview was completed in approximately
13,700 dwelling units to identify the eligible sample. The
screening yielded approximately 1,950 households responding for
the full year and approximately 6,500 SAIAN persons responded for
their entire period of eligibility in 1987.
The institutional population sample was based on a three stage
probability design. The first two stages were used to select
facilities; the final stage sampled facility residents present on
January 1, 1987. These facilities were also used to obtain a
sample of admissions between January 1, 1987, and December 31,
1987. Based on sampling specifications the Institutional
Population Component includes 1,500 facilities, 800 nursing homes
and 700 facilities for the mentally retarded. There is a total
of approximately 10,100 persons in the sample including both
residents and new admissions. This includes 5,700 persons in
nursing homes and 4,400 persons in facilities for the mentally
retarded. The sample frame for facilities in the Institutional
Population Component was derived from the 1986 Inventory of Long-
Term Care Places.
Taken in conjunction, the NMES-2 surveys yield comprehensive,
population-based information that will support studies of most
population groups of policy interest, including those presently
outside the scope of various public and private financing
mechanisms. In contrast to information obtained from program or
provider statistics, NMES-2 data can be used to analyze all
public and private sources of coverage for health care services
and out-of-pocket payments by individuals and families.
The Agency for Health Care Policy and Research sponsored the
NMES-2 data collection activities. A substantial part of the
support for the Survey of American Indians and Alaska Natives was
provided by the Indian Health Service. The Health Care Financing
Administration, the National Center for Health Statistics, and
the Office of the Assistant Secretary for Planning and Evaluation
provided extensive technical assistance during the development of
the survey design and instruments. Interviews were conducted by
the primary contractor, Westat, Inc., Rockville, Maryland and by
The National Opinion Research Center at the University of
Chicago; the Council of Energy Resource Tribes, Denver, Colorado;
and Stephen R. Braund and Associates, Anchorage Alaska. Data
processing during the analysis stage of the project is being
provided by Social and Scientific Systems, Inc., Bethesda,
Maryland.
The data were collected under the authorities of the Public
Health Service Act and are being edited and published in
accordance with the confidentiality provisions of that Act and
the Privacy Act.
Additional information on NMES-2 is available from Daniel C.
Walden, Ph.D., Director of the Division of Medical Expenditure
Studies; Center for General Health Services Intramural Research,
Agency for Health Care Policy and Research; Executive Office
Center, 2101 E. Jefferson Street, Suite 500, Rockville, MD
20852. (301/227-8400).
TECHNICAL AND PROGRAMMINGINFORMATION
General Information
This documentation describes one in a series of public use
tapes from the Household Survey of the 1987 National Medical
Expenditure Survey (NMES). The tape provides information and
related documentation on the use of and expenditures for
inpatient hospital stays in calendar year 1987. The data file
contains one record for each hospital stay for each eligible
person in the Household Survey who reported having had at least
one hospital stay during 1987. In addition to expenditures
related to the stay, each record contains selected person-level
demographic information for the respective patient, household
reported medical conditions and procedures associated with the
hospitalization, and information on length of stay.
The data file can be used to construct summary variables of
expenditures, sources of payment, and related aspects of hospital
inpatient care. Information on persons in the Household Survey
without a hospitalization in 1987 is provided in NMES Public Use
Tape 13 (Rounds 1-4 Household Survey: Population Characteristics
and Person-level Utilization), which also contains annual
person-level information on other health services use as well as
detailed demographic, employment, insurance, round specific
eligibility status indicators, and reference period dates for the
entire civilian noninstitutionalized population.
The data on this tape are being released as EBCDIC files only.
The tape also includes an EBCDIC file containing programming
statements required to create a SAS data set and a SAS format
library for the data file on the tape.
The records on this file can be linked to all NMES public use
tapes from the Household Survey by using the person identifier
(PIDX).
The following documentation offers a brief overview of the type
and level of data provided, the content and structure of the data
file and the codebook, and programming information. It contains
the following sections:
Data File Contents and Codebook Structure
Coding, Edits and Imputations
Variable Naming and Codebook Conventions
Sample Design, Estimation and Sampling Weights, and Variance
Estimation
Programming Information
References
Data Dictionary
Alphabetical and Position Listing of Variables
Codebook
Codebook Notes
More detailed information on NMES survey instruments and data
collection procedures, variance estimation programs, coding and
related information, and a list of released NMES Public Use Tapes
are in Attachments 1 to 5. Attachment 6 contains a catalogue of
data items released on this and other NMES public use tapes. It
is supplied to guide the user to the appropriate public use tape
for the data items of interest in the NMES Household Survey. All
attachments are provided in hard copy only.
Data File Contents and Codebook Structure
This public use tape contains one data file. To expedite
release of hospital inpatient data and ensure data processing
efficiency, the file structure generally reflects the structure
of the household questionnaire (see Attachment 3 for the
instrument used). The file contains information on hospital
stays, household reported main reason for admission and related
health conditions, length of stay, procedures performed, expenses
for basic hospital and for inpatient physician services, and
sources of payment. These data were obtained in four rounds of
interviews covering calendar year 1987 and in the NMES Medical
Provider Survey (see Appendix 1). The file contains 5,432
records, or one record for each specific hospital stay by a
sample person who responded for their entire period of NMES
eligibility during 1987. Hospital stays of zero nights (stays
with the same admission and discharge date) are included.
Obstetrical stays for deliveries are counted multiple times; one
for the mother and one for each newborn.
Each record on the data file contains the following
information:
Unique person and record identifiers
Indicator of the round of data collection
Selected demographic variables
Variables based on questionnaire items G1-G77 in the
hospital inpatient stay booklet. These include
household reported conditions associated with the
hospital stay coded to ICD-9, as well as codes for
procedures performed, dates of admission and discharge,
expenses and sources of payment as well as associated
imputation flags.
Weight and variance estimation variables
Detailed information on length of stay is described in Section
2.1. Detailed information on coding of medical conditions and
procedures is provided in Section 2.2. The construction of the
expense and source of payment variables and associated imputation
procedures are described in Section 2.3. Miscellaneous edits and
omissions are discussed in Section 2.4. Additional details for
selected variables are also provided in the Codebook Notes, which
immediately follow the codebook.
Coding, Edits and Imputations
Length of Stay
Each record contains two edited variables for length of
hospital stay (NUMNGHTX and NUMNT87X). The variable NUMNGHTX
represents the total number of nights per hospital stay for all
hospitalizations in 1987, including stays that began in 1986 or
extended into 1988. The variable, NUMNT87X, represents only the
number of nights spent in the hospital in 1987. The method of
constructing these variables was the same.
Hospital stay data were obtained from linked records in the
NMES Medical Provider Survey (MPS), which obtained provider
information for hospital stays reported by household respondents
(see section 4.1.2 and Appendix 1.B). If MPS length of stay
information was not available, the length of stay as reported by
the household respondent was used. If neither was available,
length of stay was imputed, based on known values for similar
hospitalizations. A weighted sequential hot-deck procedure was
used to impute data for hospital stays with complete information
to cases with missing data but similar characteristics.
Variables with known values were used to form groups of donors
and groups of recipients. Within such groups, data from donors
were assigned to recipients, taking into account the weights
associated with the complex NMES household survey design. Source
data for the imputation were MPS data, when available, or the
Household Survey otherwise. Classification variables in this
imputation were household-reported reason for hospitalization,
patient residence in the West Census Region versus other regions,
household reported psychiatric conditions as reason for
admission, and the number of physicians billing for the stay.
Age, number of difficulties with activities of daily living,
perceived general health status, source of payment, and selected
other conditions were used as sort variables. The file contains
a flag (NUMNTFLG and NUM87FLG) for each length of stay variable
indicating whether length of stay was imputed or obtained from
the Medical Provider Survey or a household respondent.
Medical Conditions and Procedures
Medical Condition Coding
This file contains up to four condition codes per hospital stay
record. Records for normal deliveries and healthy newborn babies
have no condition codes attached. For each hospital stay
reported in the Household Survey, information on reasons for
hospitalization in terms of related medical conditions, medical
diagnosis, if any, related history, and parts of the body
affected by the condition were obtained. This respondent-
reported information was used to code each condition for the
reported hospitalization into one of the codes of the 9th
Revision of the International Classification of Diseases, revised
for use in the National Health Interview Survey (NHIS; NCHS,
1979). The revision takes into account the experience of the
National Center for Health Statistics, the sponsor of the NHIS,
in coding household-reported conditions. One major revision of
the ICD-9 coding procedure in surveys using the NHIS system is
the introduction of X-codes. These codes represent impairments
such as blindness, deafness, and paralysis. Coding instructions
directed coders to favor X-codes over other ICD-9 codes that
could be applied to a particular condition. Attachment 4.A
provides details of the NHIS coding scheme for X-codes. It is
important to note that the condition data on each record and the
information used to code medical conditions were provided
exclusively by the household respondent. Any information
regarding medical conditions related to the hospital stay that
was obtained through the Medical Provider Survey (MPS) was not
merged onto this file.
Procedure Coding
This file contains up to 3 procedure codes per record. For
each hospital stay reported in the Household Survey, questions
were asked about operations or surgical procedures performed
during that stay. The respondent-reported procedure was used to
code one of the codes in the 9th Revision of the International
Classification of Diseases, Clinical Modification, 3 (ICD9-CM3)
revised for use in the National Health Interview Survey (NHIS;
NCHS, 1979). The revision only uses the first two digits of the
ICD9-CM3 which indicates the body system or region on which the
procedure was performed. For miscellaneous diagnostic and
therapeutic procedures, the code represents a general procedure
class, for example, "diagnostic radiology" or "physical or
rehabilitative procedures". Attachment 4B provides the codes for
medical procedures.
Quality Control
Coding of medical conditions and procedures was conducted by
trained medical coders. Two-stage verification of the coding was
performed first by supervisors and then by trained nosologists.
Coders were required to maintain an error rate at or below 2
percent throughout the coding process. No additional editing or
imputation was performed on these variables. Each
hospitalization record can have up to four ICD-9 condition codes
(or three procedure codes) attached to it. The first condition
(or procedure) on the record does not necessarily reflect the
primary condition (or procedure) associated with the stay and,
more generally, the order of the conditions (or procedures) on
the record does not reflect importance or severity. In addition,
there is a small probability that duplicate conditions (or
procedures) may appear for the same record, since editing for
duplications was not done.
Expenses and Sources of Payment
Expenses
General Information
The file contains three expense variables per stay: basic
hospital facility expenses; expenses for physicians who billed
separately for any inpatient services provided during the
hospital stay; and total expenses, which sums facility and
physician expenses.
Expenses refer to charges with two exceptions. First, they
reflect payments rather than charges when these were reduced to
the amounts allowed by third-party payers, such as private health
insurers, Medicare, or Medicaid. Second, a dollar value was
assigned to the facility expense variable in all cases and to the
physician expense variable in some cases where no charge was
specified. For example, there is no stated total charge
(although there may be copayments) for stays by individuals
insured through an HMO, where care is financed through a prepaid
capitation payment. Also, there may have been no specified
charge for stays financed through government grants, charities,
and other non-patient related revenue sources. In these cases,
an expense value was imputed from the expenses associated with
otherwise similar stays. This approach was also used for most
flat fee arrangements. An example is a lump-sum charge applied
to a number of different services, for example, a series of
outpatient visits associated with an inpatient stay, where there
was no reasonable method of allocating that charge across each of
them. For some obstetrical stays, flat fees for physician
expenses were allocated across different services.
Facility Expense
Both unedited (EXPFAC) and edited (EXPFACX) facility expense
variables are included on this file. Hospital facility expenses
are all expenses for direct hospital care, including room and
board, diagnostic and laboratory work, x-rays and similar
charges, as well as any physician services included in the
hospital charge. These facility expense data were primarily
obtained from a linked record in the Medical Provider Survey
(MPS). If MPS data were not available, facility expenses as
reported by the household respondent were used. Expenses for
each hospital stay were imputed on a per diem basis where neither
was available, or for extreme values reported by household
respondents. Missing values for facility cost per day were
imputed, based on known values for similar hospital stays where
expenses were derived from the Medical Provider Survey (MPS).
Length of stay, (NUMNGHTX), source of payment, number of
physicians billing for inpatient services, and SMSA residence
were used as classification variables. Sort variables were:
expected sources of payment, surgery performed during the stay,
patient age and sex, household reported reason for admission,
discharge status (live or deceased), residence in the West Census
Region versus other regions, and admission from the hospital
emergency room. The imputed expense per day was then multiplied
by nights of stay (if NUMNGHTX > 0) to obtain total facility
expense for that stay. If NUMNGHTX was zero, then the imputed
expense per day was used.
Inpatient Physician Expense
Each record contains one edited variable for the sum of all
expenses for physicians who billed for inpatient services linked
to the hospital stay (EXPDRX). Because household respondents
generally do not accurately report expenses for physicians seen
in conjunction with a hospital stay, charges for these services
were derived entirely from the Medical Provider Survey (MPS). A
zero dollar value for inpatient physician expense was assigned
for a number of records; this reflects one or more of the
following: all physician charges were included in the facility
expense; the stay was for less than 24 hours with no separate
physician charges for inpatient visits or for post-surgical
follow-up; or the record relates to a normal newborn for whom
physician charges are included on the record for the mother's
stay.
In the Medical Provider Survey, physicians who were listed in
the medical records for a given hospitalization were enumerated
by specialty and contacted to determine the charges rendered for
the hospital episode. When only partial physician charges could
be obtained for a given hospital stay, the missing elements were
imputed using the median MPS charge for a physician of that
specialty. In cases where no charges were obtained but the
physicians could be enumerated, a median MPS charge was also used
to impute physician expenses. When no MPS data were available
for a stay, the total physician charge was imputed using the same
procedures as those described for the facility expense variable.
An imputation flag indicates for each record whether the
physician expenditure data were obtained from the MPS or imputed
using the MPS as the source of data.
Hospital Expense
The total hospital expense variable (EXPTOTX) is the sum of
facility expense plus physican expense for the hospital inpatient
stay. Excluded from this total are any expenses for private duty
nursing in cases where private duty nurses billed separately from
the hospital.
Sources of Payment
Sources of payment were obtained from the Household Survey or
imputed. Each record on the file contains constructed variables
showing the percent of the expense paid by various sources.
Again, there are two sets of source of payment variables: one for
hospital facility expense and one for associated inpatient
physician expense. Each set contains 9 variables which sum to
100 percent, corresponding to the percent paid by each of the
following sources:
Out of pocket by user or family;
Private insurance;
Medicaid;
Medicare;
Other Federal (CHAMPUS, CHAMPVA, Supplemental
Security Income (SSI), Indian Health Service facility
or contract, Intertribal Council, Alaska Native
Corporation, Veteran's Administration, any military and
other federal programs such as free government
screening services and NIH care);
Other State and local medical assistance (such as
community health centers but excluding local and state
employment related insurance and welfare programs);
Workman's compensation;
Free from provider, including professional courtesy and
bad debt; and
Other, which includes automobile and car insurance,
other kinds of insurance not specified, company (where
the company is not the insurer or employer), school
(where school is not the insurer or employer), union
(where union is not the insurer or empoyer), charity,
friend, foreign government or not otherwise specified.
Based on household reports of the percent or amount paid by
each of these sources of payment, the source of payment variables
were edited and, where necessary, imputed to correct for the
following: (1) the household reported payer was incompatible
with enrollment in public and private insurance programs reported
for the person; (2) the person was not billed for the
hospitalization and/or associated physician services so that no
expense or sources of payment were reported; (3) the sum of the
reported sources did not equal 100 percent; or (4) the sources of
payment or the amounts or proportions of the payment were
partially or completely missing.
The inpatient physician source of payment variables were
constructed based upon Household Survey reports of the number of
physicians providing care. Assuming that all physicians
associated with a stay would have similar source of payment
distributions, the physician source of payment variables for each
stay were constructed as the average for all physicians
associated with the stay for whom source of payment data were not
missing. When source of payment data were missing for all
physicians associated with a stay the source of payment data were
imputed.
The general imputation strategy used a weighted sequential hot
deck procedure. The classification variables used in the source
of payment imputation for facility and physician included
insurance coverage, region of residence, month of stay, length of
stay, whether the respondent had been able to report the charge
for the hospitalization, and union membership status of the
primary insured. In addition, stays missing source of payment
information and for which the Veteran's Administration was the
site of care (HINVA = 1 or question G20 = 4), were imputed 100
percent to the other federal category. Each record with imputed
values contains corresponding imputation flags.
Other Edits and Omissions
Many of the data items from the hospital inpatient stay booklet
were omitted from this file because they were components or
probes used to construct the summary variables provided. Omitted
variables were not considered to be of independent analytic
interest but rather were methodological probes to insure that the
respondent had provided complete information. The summary
variables included on this file reflect all of the components
collected in the hospital stay section of the questionnaire.
Unless indicated, all Hospital Stay Booklet variables are
unedited. Skip patterns associated with unedited variables have
not been reconciled.
No editing was performed on the round indicator that is on this
file. Editing was performed on all link variables except
PLACEIDX (see Codebook Notes section).
This data file contains 497 records where the length of the
hospital stay was zero nights (NUMNGHTX = 0). These records were
not used in the construction of the person-level hospital
inpatient use variables on NMES Public Use Tape 13, where they
were considered hospital outpatient department visits. (See
section 4.2.1)
Variable Naming and Codebook Conventions
The codebook provides unweighted and weighted frequencies for
all variables on the file. The codebook contains variable
information and frequency distributions for a total of 5,432
records. Weighted, these records represent 34,436,582 hospital
inpatient stays by 26,370,883 people (4,051 unweighted persons).
Complete variable listings in alphabetical order and by file
position are provided for cross-reference. Most variable
descriptions in the codebook are abbreviated versions of
questionnaire items, preceded by indicators of item number. A
copy of the hospital inpatient stay booklet is included as
Attachment 3 to this public use tape to permit a full
understanding of the content and wording of each item, the
structure of questionnaire sections, skip patterns and
administrative information.
The codebook describes an EBCDIC data set and provides the
following programming identifiers for each variable:
IDENTIFIER DESCRIPTION
NAME Variable name (maximum of 8
characters)
DESCRIPTION Variable descriptor (maximum of 40
characters)
FORMAT Number of bytes and decimal places
TYPE Type of data: numeric (indicated by
NUM) or character (indicated by CHAR)
START Beginning column position of variable in
the record
END Ending column position of variable
in the record
NOTES Indicator of an explanatory note(s)
corresponding to the variable
In general, variable names reflect the content of the variable,
with an 8 character limitation. For edited versions of original
variables, the edited variable name is identical to the original
variable with an "X" appended (and truncated when necessary to
comply with the 8 character limitation). For variables
corresponding to specific questionnaire items, the question
number is included in the variable label.
The following reserved code values are used:
VALUE DEFINITION
-1 INAPPLICABLE Question was not asked due to skip
pattern
-5 NEVER KNOW Question was asked and respondent
did not know and never will know the
answer
-7 REFUSED Question was asked and respondent
refused to answer the question
-8 DK Question was asked and respondent did
not know the answer
-9 NOT ASCERTAINED Interviewer did not record the data
Sample Design, Estimation Procedures, and Sampling Weights
Sample Design and Response Rates
Household Survey
The NMES household survey was designed to produce national
estimates representative of the civilian noninstitutionalized
population of the United States as of 1987. For sample
selection, the household component of NMES used two independent
national multistage area samples from Westat, Inc. and the
National Opinion Research Center at the University of Chicago.
To improve the quality of the data and to allow for analysis of
trends during 1987, it was conducted as a panel survey over four
core rounds of interviewing.
Sampling specifications required the selection of about 17,500
households for the first core household interview. Data were
obtained for about 86 percent of eligible households in the first
interview and 80 percent by the fourth interview. Approximately
6 percent of all survey participants provided data for only some
of the time in which they were eligible to respond. These
persons were considered total nonrespondents and a standard
nonresponse weight adjustment was used to account for possible
selection bias in this respect. For a detailed description of
the survey design and of sampling, estimation, and adjustment
methods see Cohen, DiGaetano, and Waksberg (1991).
The Medical Provider Survey
The NMES Medical Provider Survey (MPS) was primarily designed
to reduce the bias associated with national medical expenditure
estimates derived from household reported data; this bias is a
function of item nonresponse and poor quality data. By selec-
tively targeting individuals who were most likely to misreport or
not possess adequate knowledge about their medical expenditures
and medical care events that were expected to be associated with
expense data of questionable quality, optimal use could be made
of medical provider reported data to improve the accuracy of
national medical expenditure survey estimates. See Attachment 1B
for a full discussion of MPS design and response rates.
Estimation and Sampling Weights
The application of appropriate sampling weights is essential to
the derivation of estimates when using this public use file. The
weight provided for use with hospital stay data, INCALPER,
reflects adjustments for complete nonresponse to the NMES survey
and poststratification to the Census Bureau 1987 Current
Population Survey (CPS) cross-classified by age, race/ethnicity,
gender and poverty status. All persons who were eligible at any
time during 1987 and responded for the entire period of their
eligibility have positive INCALPER weights. Only the 4,051
persons who have positive INCALPER weights and who had at least
one hospital stay are represented on this file. In order to
produce national estimates related to the frequency, doctor
expenses, facility expenses, total expenses and sources of
payment for hospital care, the value in each record contributing
to the estimates must be multiplied by the weight (INCALPER)
contained on that record. It should be noted that the weight,
INCALPER, can also serve as a person-level estimation weight (see
NMES Public Use Tape 13).
For estimates involving persons in the Household Survey not on
this file (e.g., persons without a hospital stay in 1987) or for
detailed person-level characteristics of users, including round
specific eligibility status and reference period dates, the data
on this file should be merged with NMES Public Use Tape 13 (see
section 3.3) using the person-level identification variable PIDX.
Basic Estimates of Utilization and Expenditures
General Information
This file is constructed for efficient estimation of
utilization and expenditures for hospital care. Such estimates
include the total number of and expenses for hospital stays.
Basic Estimates of Utilization
In order to estimate the total number of hospital stays,
regardless of the length of the hospital stay, simply sum the
weight INCALPER across all records.
This file includes hospital stay records for which the length
of stay is zero nights (i.e. stay with same day admission and
discharge). There are two variables on this file that measure
the length of a hospital stay (NUMNGHTX and NUMNT87X; see Section
2.1). Either of these variables can have a value that indicates
the length of the hospital stay was zero nights. In order to
replicate hospital inpatient utilization data provided on NMES
Public Use Tape 13, records with zero night stays (specifically
records with NUMNGHTX = 0) should be excluded from the estimate
(weighted, the number of hospital stays with at least one night
in the hospital are 31,225,348, by 24,169,932 people; unweighted
number of persons are 3,719). (In order to replicate the
variable HOSPNGHT on Tape 13, the variable NUMNGHTX should be
used.)
It should be noted that on the current file, stays for
deliveries are counted as multiple stays; one for the mother and
one for each newborn. To exclude the 500 newborns (3,690,243
weighted stays) from the estimate of hospital stays, records with
RSNINHOS = 5 should be dropped. (However, the expenditure portion
associated with newborns should be added to the mother's
expenditures when estimating hospital expenses).
Basic Estimates of Expenditures
As an example of a basic expenditure estimate, the mean expense
for a hospital stay regardless of length should be calculated as
the weighted sum of the total expenses across all records in the
file (sum of EXPTOTX x INCALPER) divided by weighted sum of the
number of stays (sum of INCALPER). Thus, the numerator is the
national estimate for total expenses for hospital care and the
denominator is the population estimate for the total number of
hospital stays.
While most hospitalizations began and ended during 1987, there
are 105 records on this file with admission dates prior to
January 1, 1987 or with 1988 as the year of discharge. The
edited expense variables on this file (EXPDRX, EXPFACX and
EXPTOTX) reflect expenses for hospitalization from time of
admission to time of discharge, regardless of the year (i.e.,
based on the variable NUMNGHTX). The variable NUMNT87X
constrains the length of stay to only those hospital nights which
occurred during calendar year 1987. To approximate expenses
incurred only during 1987, one must first make new expense
variables by calculating the relevant per night expense (e.g.,
EXPTOTX/NUMNGHTX) and multiplying by the restricted number of
nights (NUMNT87X).
There are some records for newborns on this file with values
for doctor expense (EXPDRX) of zero because all doctor expenses
associated with the birth appear on the mother's record. However
in all instances the facility expense (EXPFACX) for the newborn
is on the newborn's record.
Subsetting to records based on characteristics of interest
expands the scope of potential estimates. For example, the
number of hospital stays of at least one night where the facility
expense was paid for, totally or in part, by Medicare is
estimated by summing INCALPER across all records where the length
of stay (NUMNGHTX) is greater than zero and the Medicare source
of payment variable (SOPFMCR) is greater than zero.
Person-Based Ratio Estimates
Person-Based Ratio Estimates for Persons with Hospital Stays
When calculating ratio estimates where the denominator is
persons, not hospital stays, care should be taken to properly
define and estimate this denominator. If the estimate of
interest, for example, is the mean expense for all hospital-
izations across all persons with hospital stays, the following
strategy should be considered: All expenses for a person on this
file should be summed and a person-level total expense variable
created (e.g., the variable X). The mean national estimate would
then be derived by obtaining the ratio of the weighted sum of
total expense per person across all unique persons on the file
(sum of (INCALPER times X)) divided by the weighted number of
unique persons on the file (sum of INCALPER). Only one INCALPER
value for each PIDX (i.e., each person) should contribute to the
calculation of the sums for both the numerator and the
denominator.
Person-Based Ratio Estimates Relative to the Entire Population
If the ratio relates to the entire population, this file cannot
be used to calculate the denominator, as only those persons with
at least one hospital stay are represented on it. In this case,
Public Use Tape 13 which has data for all sampled people, must be
used to estimate persons. For example, to estimate the
proportion of persons 65 and over with at least one
hospitalization for a circulatory disease, the current file is
used to calculate the numerator and NMES Tape 13 is used to
calculate the denominator.
Sampling Weights for Merging Previous Releases of NMES Household
Data with the Current Tape
There have been several previous releases of NMES Household
Survey public use data (see Attachment 5). Unless a variable
name common to several tapes is provided, the sampling weights
contained on these tapes are tape-specific. The tape-specific
sampling weights reflect minor adjustments to eligibility and
response indicators due, among other factors, to birth, death, or
institutionalization among respondents. Adjustments to the
weights have also included post-stratification adjustments to
control for the distribution of the U.S. noninstitutionalized
population by poverty status and, where appropriate, nonresponse
adjustments for round-specific supplemental questionnaires (e.g.,
the health status questionnaires).
For estimates from a NMES data file that do not require merging
with variables on other NMES files, the sampling weights provided
on that tape are the appropriate weights. When merging a NMES
household survey tape to another NMES household survey tape, the
major analytical variable (i.e., the dependent variable)
determines the correct sampling weight to use. For example, for
1987 estimates of hospital use or expenditures (from the present
household survey tape) using health status variables from NMES
Public Use Tape 9, the full-year weight, INCALPER, on the present
tape should be used. By contrast, the weight HSQACCWT from
Public Use Tape 9 should be used when the major dependent
variable is health status and hospital utilization is an
independent variable. Three exceptions to this general sampling
weight and merger rule are noted below. For details concerning
the appropriate weight specific to each tape, see the hard copy
information specific to each tape.
(1) For estimates of round one data from NMES Public
Use Tape 3 (preliminary round one person
characteristics and functional health status data), the
round one weight (WGTR1PER) provided on NMES Public Use
Tape 13 should be used. It should be noted that as a
result of further response and eligibility edits, not
all persons with positive round one weights on Tape 3
will link when merging it to Tape 13. In those
instances, an imputation or weighting strategy can be
developed to adjust for all persons with positive
WGTR1PER weights. The preferred approach is to use the
round one data and the round one weight released on
NMES Tape 13.
(2) For point in time estimates of persons with
activity of daily living and instrumental activity of
daily living difficulties (Tape 10), the round one or
four weight (WGTR1PER and WGTR4PER, respectively),
provided on NMES Tape 13 should be used in all
instances of merged data regardless of the type of
analysis.
(3) NMES Public Use Tape 4, which contains prescribed
medicine data for the Medicare beneficiary population,
should not be merged with the current tape (or any
other NMES public use file) because of adjustments to
the sampling weights made after the release of Tape 4.
It should be noted that the sampling weight provided on NMES
Tape 9 (HSQACCWT) reflects nonresponse adjustments specific to
the health status questionnaire and access to care supplement
data on that tape. This further nonresponse adjustment requires
additional considerations in merging Tape 9 with the current
tape.
(1) When making estimates for data in the current tape,
the sampling weights provided on the current tape
should be used. Since this would include persons not
on Tape 9 in the analysis, data items from Tape 9 will
have missing values for these persons.
(2) When making estimates of health status or access to
care indicators, the Tape 9 weight, HSQACCWT, should be
used. This weight adjusts for the exclusion of persons
included on the present tape.
Variance Estimation
Variance estimates of sample statistics require that the
complex nature of the NMES Household Survey design be taken into
account for hypothesis testing and for the construction of
confidence intervals. To obtain variance estimates of statistics
by means of statistical programs that use the Taylor series
method of variance estimation, variables must be used that denote
the strata and the primary sampling unit (PSU) within a given
stratum. The variables STRATUMX and SPSU are these variables,
respectively, and are included on the data file.
There are variance estimation programs that account for the
complex survey design. A list of available variance estimation
programs is produced in Attachment 2.
Programming Information
These files are contained on a standard label, 9 track, 6250
bpi tape. The specifications for each file on the tape are as
follows:
FILE 1:
Description: NMES Household Survey Hospital Inpatient Stay Data
File
Dataset Name: NMES.PUF144.DATA
Number of Observations: 5,432
Number of Variables: 74
Record Length: 400
Block Size: 16,000
Record Format: FB
FILE 2:
Description: Technical and Programming Information and Data
Dictionary for the NMES Hospital Inpatient Stay
Data File
Dataset Name: NMES.PUF144.DOC
Record Length: 133
Block Size: 19,950
Record Format: FB
FILE 3:
Description: Additional Documentation for SAS Users for the
NMES HHS Hospital Inpatient Stay Data File
Dataset Name: NMES.PUF144.SRC
Record Length: 80
Block Size: 800
Record Format: FB
File 1 was created using the SAS (Statistical Analysis System,
version 5.18) computer software, and converted to EBCDIC format.
File 2 contains the technical documentation stored as an
Operating System (OS) EBCDIC file containing ASA carriage control
characters in the first byte in each record, which will direct
the line printer to skip lines, begin a new page, etc. This
technical documentation can be copied to disk and retrieved on-
line to view, to modify with a text editor program such as
WYLBUR, or to make additional copies. File 3 is an EBCDIC file
containing the following additional documentation for SAS users:
INPUT statement to create the SAS file, including a LABEL
statement; SAS statements which assign a format name to each
variable; and SAS statements describing formats.
References
Cohen, S.B., DiGaetano, R. and Waksberg, J. (1991). National
Medical Expenditure Survey: Sample Design of the 1987
Household Survey, Methods 3. AHCPR Pub. No. 91-0037. DHHS:
U.S. Public Health Service.
National Center for Health Statistics (1979). Medical Coding
Manual: National Health Interview Survey. DHHS: U.S. Public
Health Service.
DATA DICTIONARY
Alphabetical Listing of Variables
NMES EVENTS -- HOSPITAL STAYS CODEBOOK
ALPHABETICAL AND POSITIONAL LISTING OF VARIABLES
DATE: JUNE 1, 1992
-----ALPHABETICAL LISTING OF VARIABLES-----
START END NAME DESCRIPTION
----- --- ---- -----------
142 143 ANYOPER G16 OPERATION DURING HOSP STAY
128 129 BABYBORN G11 DELIVERY DURING HOSPITAL STAY
140 141 BABYNORM G14 BABY NORMAL
51 52 DATEBDD G2 DATE PID ENTERED HOSPITAL - DAY
49 50 DATEBMM G2 DATE PID ENTERED HOSPITAL - MONTH
53 54 DATEBYY G2 DATE PID ENTERED HOSPITAL - YEAR
57 58 DATEEDD G3 DATE PID LEFT HOSPITAL - DAY
55 56 DATEEMM G3 DATE PID LEFT HOSPITAL - MONTH
59 60 DATEEYY G3 DATE PID LEFT HOSPITAL - YEAR
9 11 EN EVENT NUMBER
20 31 EVENTIDX EVENT ID (ODU+PN+EN)
250 250 EXPDRFLG IMPUTATION FLAG FOR EDITED DOC EXPENSE
242 249 EXPDRX EDITED DOCTOR EXPENSE
150 158 EXPFAC G21 ORIGINAL FACILITY CHARGE
159 167 EXPFACX EDITED FACILITY EXPENSE/STAY
168 168 EXPFCFLG IMPUTATION FLAG FOR EDITED FAC EXPENSE
335 341 EXPNRS G62 TOTAL NURSE EXPENSE/STAY
324 332 EXPTOTX EDITED TOTAL EXPENSE/HOSPITAL STAY
71 72 HINVA G5 HOSPITAL IS A VA FACILITY
106 109 ICD1 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 1
110 113 ICD2 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 2
114 117 ICD3 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 3
118 121 ICD4 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 4
342 343 IHSFAC G77 HOSPITAL IS IHS OR TRIBE/ANC
344 355 INCALPER FULL-YEAR WEIGHT
34 36 LASTAGE ED PID AGE AT END OF LAST ELIGIBLE ROUND
130 131 NORMDELV G12 NORMAL DELIVERY
63 65 NUMNGHTX G4 ED NUMBER OF NIGHTS PER HOSPITAL STAY
61 62 NUMNIGHT G4 NUMBER OF NIGHTS PER HOSPITAL STAY
66 66 NUMNTFLG IMPUTATION FLAG FOR NUMNGHTX
67 69 NUMNT87X ED NUMBR OF 1987 NIGHTS PER HOSPITL STAY
70 70 NUM87FLG IMPUTATION FLAG FOR NUMNT87X
333 334 NURSCARE G58 SPECIAL/PRIV DUTY NURSE REQUIRED
1 5 ODUX ORIGINAL DWELLING UNIT OF PERSON
12 19 PIDX PERSON ID (ODU+PN)
41 48 PLACEIDX G1 ID OF FACILITY
6 8 PN PERSON NUMBER
105 105 PREGFLG STAY RELATED TO PREGNANCY
132 139 PREGIDX PREGNANCY ID (ODU+COUNTER)
122 123 PROCODE1 G8 1ST PROCEDURE CODE
124 125 PROCODE2 G8 2ND PROCEDURE CODE
126 127 PROCODE3 G8 3RD PROCEDURE CODE
144 145 PROCODE4 G17 1ST PROCEDURE CODE
146 147 PROCODE5 G17 2ND PROCEDURE CODE
148 149 PROCODE6 G17 3RD PROCEDURE CODE
39 39 RACE3 PID RACE/ETHNICITY
38 38 RACE6 ED PID RACE
32 33 ROUND DATA COLLECTION ROUND
75 104 RSNHOSOS G6 OTHER REASON FOR ADMISSION
73 74 RSNINHOS G6 MAIN REASON FOR ADMISSION
37 37 SMPSEXR PID SEX
315 322 SOPDFFP PCT PAYMENT FREE FROM PROVIDER-DR EXP
323 323 SOPDFLG IMPUTATN FLG FOR SOURC PAYM INFO-DOC EXP
275 282 SOPDMCD PCT PAYMENT FROM MEDICAID-DR EXP
267 274 SOPDMCR PCT PAYMENT FROM MEDICARE-DR EXP
283 290 SOPDOTFD PCT PAYMENT FROM OTHER FEDERAL-DR EXP
307 314 SOPDOTHR PCT PAYMENT FROM OTHER-DR EXP
291 298 SOPDOTST PCT PAYMENT FROM OTHER STATE-DR EXP
259 266 SOPDPRVT PCT PAYMENT FROM PRIVATE INSUR-DR EXP
251 258 SOPDSELF PCT PAYMENT FROM SELF OR FAMILY-DR EXP
299 306 SOPDWC PCT PAYMENT FROM WORKERS COMP-DR EXP
233 240 SOPFFFP PCT PAYMENT FREE FROM PROVIDER-FAC EXP
241 241 SOPFFLG IMPUTATN FLG FOR SOURC PAYM INFO-FAC EXP
193 200 SOPFMCD PCT PAYMENT FROM MEDICAID-FAC EXP
185 192 SOPFMCR PCT PAYMENT FROM MEDICARE-FAC EXP
201 208 SOPFOTFD PCT PAYMENT FROM OTHER FEDERAL-FAC EXP
225 232 SOPFOTHR PCT PAYMENT FROM OTHER-FAC EXP
209 216 SOPFOTST PCT PAYMENT FROM OTHER STATE-FAC EXP
177 184 SOPFPRVT PCT PAYMENT FROM PRIVATE INSUR-FAC EXP
169 176 SOPFSELF PCT PAYMENT FROM SELF OR FAMILY-FAC EXP
217 224 SOPFWC PCT PAYMENT FROM WORKERS COMP-FAC EXP
359 359 SPSU PSEUDO PSU
40 40 SREGION PID CENSUS REGION
356 358 STRATUMX SAMPLING STRATUM
Positional Listing of Variables
NMES EVENTS -- HOSPITAL STAYS CODEBOOK
ALPHABETICAL AND POSITIONAL LISTING OF VARIABLES
DATE: JUNE 1, 1992
-----POSITIONAL LISTING OF VARIABLES-----
START END NAME DESCRIPTION
----- --- ---- -----------
1 5 ODUX ORIGINAL DWELLING UNIT OF PERSON
6 8 PN PERSON NUMBER
9 11 EN EVENT NUMBER
12 19 PIDX PERSON ID (ODU+PN)
20 31 EVENTIDX EVENT ID (ODU+PN+EN)
32 33 ROUND DATA COLLECTION ROUND
34 36 LASTAGE ED PID AGE AT END OF LAST ELIGIBLE ROUND
37 37 SMPSEXR PID SEX
38 38 RACE6 ED PID RACE
39 39 RACE3 PID RACE/ETHNICITY
40 40 SREGION PID CENSUS REGION
41 48 PLACEIDX G1 ID OF FACILITY
49 50 DATEBMM G2 DATE PID ENTERED HOSPITAL - MONTH
51 52 DATEBDD G2 DATE PID ENTERED HOSPITAL - DAY
53 54 DATEBYY G2 DATE PID ENTERED HOSPITAL - YEAR
55 56 DATEEMM G3 DATE PID LEFT HOSPITAL - MONTH
57 58 DATEEDD G3 DATE PID LEFT HOSPITAL - DAY
59 60 DATEEYY G3 DATE PID LEFT HOSPITAL - YEAR
61 62 NUMNIGHT G4 NUMBER OF NIGHTS PER HOSPITAL STAY
63 65 NUMNGHTX G4 ED NUMBER OF NIGHTS PER HOSPITAL STAY
66 66 NUMNTFLG IMPUTATION FLAG FOR NUMNGHTX
67 69 NUMNT87X ED NUMBR OF 1987 NIGHTS PER HOSPITL STAY
70 70 NUM87FLG IMPUTATION FLAG FOR NUMNT87X
71 72 HINVA G5 HOSPITAL IS A VA FACILITY
73 74 RSNINHOS G6 MAIN REASON FOR ADMISSION
75 104 RSNHOSOS G6 OTHER REASON FOR ADMISSION
105 105 PREGFLG STAY RELATED TO PREGNANCY
106 109 ICD1 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 1
110 113 ICD2 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 2
114 117 ICD3 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 3
118 121 ICD4 G7/G9/G10/G13/G15 ICD9 CODE-CONDITION 4
122 123 PROCODE1 G8 1ST PROCEDURE CODE
124 125 PROCODE2 G8 2ND PROCEDURE CODE
126 127 PROCODE3 G8 3RD PROCEDURE CODE
128 129 BABYBORN G11 DELIVERY DURING HOSPITAL STAY
130 131 NORMDELV G12 NORMAL DELIVERY
132 139 PREGIDX PREGNANCY ID (ODU+COUNTER)
140 141 BABYNORM G14 BABY NORMAL
142 143 ANYOPER G16 OPERATION DURING HOSP STAY
144 145 PROCODE4 G17 1ST PROCEDURE CODE
146 147 PROCODE5 G17 2ND PROCEDURE CODE
148 149 PROCODE6 G17 3RD PROCEDURE CODE
150 158 EXPFAC G21 ORIGINAL FACILITY CHARGE
159 167 EXPFACX EDITED FACILITY EXPENSE/STAY
168 168 EXPFCFLG IMPUTATION FLAG FOR EDITED FAC EXPENSE
169 176 SOPFSELF PCT PAYMENT FROM SELF OR FAMILY-FAC EXP
177 184 SOPFPRVT PCT PAYMENT FROM PRIVATE INSUR-FAC EXP
185 192 SOPFMCR PCT PAYMENT FROM MEDICARE-FAC EXP
193 200 SOPFMCD PCT PAYMENT FROM MEDICAID-FAC EXP
201 208 SOPFOTFD PCT PAYMENT FROM OTHER FEDERAL-FAC EXP
209 216 SOPFOTST PCT PAYMENT FROM OTHER STATE-FAC EXP
217 224 SOPFWC PCT PAYMENT FROM WORKERS COMP-FAC EXP
225 232 SOPFOTHR PCT PAYMENT FROM OTHER-FAC EXP
233 240 SOPFFFP PCT PAYMENT FREE FROM PROVIDER-FAC EXP
241 241 SOPFFLG IMPUTATN FLG FOR SOURC PAYM INFO-FAC EXP
242 249 EXPDRX EDITED DOCTOR EXPENSE
250 250 EXPDRFLG IMPUTATION FLAG FOR EDITED DOC EXPENSE
251 258 SOPDSELF PCT PAYMENT FROM SELF OR FAMILY-DR EXP
259 266 SOPDPRVT PCT PAYMENT FROM PRIVATE INSUR-DR EXP
267 274 SOPDMCR PCT PAYMENT FROM MEDICARE-DR EXP
275 282 SOPDMCD PCT PAYMENT FROM MEDICAID-DR EXP
283 290 SOPDOTFD PCT PAYMENT FROM OTHER FEDERAL-DR EXP
291 298 SOPDOTST PCT PAYMENT FROM OTHER STATE-DR EXP
299 306 SOPDWC PCT PAYMENT FROM WORKERS COMP-DR EXP
307 314 SOPDOTHR PCT PAYMENT FROM OTHER-DR EXP
315 322 SOPDFFP PCT PAYMENT FREE FROM PROVIDER-DR EXP
323 323 SOPDFLG IMPUTATN FLG FOR SOURC PAYM INFO-DOC EXP
324 332 EXPTOTX EDITED TOTAL EXPENSE/HOSPITAL STAY
333 334 NURSCARE G58 SPECIAL/PRIV DUTY NURSE REQUIRED
335 341 EXPNRS G62 TOTAL NURSE EXPENSE/STAY
342 343 IHSFAC G77 HOSPITAL IS IHS OR TRIBE/ANC
344 355 INCALPER FULL-YEAR WEIGHT
356 358 STRATUMX SAMPLING STRATUM
359 359 SPSU PSEUDO PSU
Codebook
General Information
NMES EVENTS -- HOSPITAL STAYS CODEBOOK
DATE: JUNE 1, 1992
________________________
THIS CODEBOOK PROVIDES UNWEIGHTED AND WEIGHTED FREQUENCIES FOR THE
USE OF HOSPITAL SERVICES FOR CALENDAR YEAR 1987. THE DATA FILE
CONTAINS ONE RECORD PER HOSPITAL STAY FOR EACH PERSON IN THE
HOUSEHOLD SURVEY WHO REPORTED HAVING HAD A HOSPITAL STAY BEGINNING
OR TERMINATING IN CALENDAR YEAR 1987. ALL STAYS INVOLVING
DELIVERIES HAVE MULTIPLE RECORDS, ONE FOR THE MOTHER AND ONE
FOR EACH CHILD. BASIC DEMOGRAPHIC INFORMATION FOR EACH PATIENT,
MAIN REASON FOR ADMISSION AND ADDITIONAL CONDITIONS AS REPORTED
IN THE HOUSEHOLD SURVEY, LENGTH OF STAY, AND ALL HOSPITAL EXPENSES
AND SOURCES OF PAYMENT INCLUDING THOSE FOR INPATIENT PHYSICIAN CARE
ARE INCLUDED ON EACH RECORD. FOR VARIABLES CORRESPONDING DIRECTLY
TO QUESTIONNAIRE ITEMS, THE ITEM NUMBER IS PROVIDED IN THE VARIABLE
DESCRIPTOR, WHICH ALSO IDENTIFIES EDITED VARIABLES. FOR VARIABLES
WITH AN ASTERISK IN THE RIGHTMOST COLUMN, EXPLANATORY NOTES ARE
PROVIDED AT THE END OF THE CODEBOOK IN ALPHABETICAL ORDER OF THE
VARIABLE NAME. TO OBTAIN NATIONAL ESTIMATES FOR THE VARIABLES ON
THIS FILE, THE WEIGHT DESCRIBED AT THE END OF THIS CODEBOOK MUST
BE USED. INFORMATION CONSERNING SAMPLE DESIGN AND ESTIMATION IS
PROVIDED IN THE FILE DOCUMENTATION.
Variable Positions 1-55
NAME DESCRIPTION FORMAT TYPE START END NOTE
---- ----------- ------ ---- ----- --- ----
ODUX ORIG DWELG UNIT OF PERSON 5.0 NUM 1 5
VALUE UNWEIGHTED WEIGHTED BY INCALPER
20001-37615 5,432 34,436,582
TOTAL 5,432 34,436,582
PN PERSON NUMBER 3.0 NUM 6 8
VALUE UNWEIGHTED WEIGHTED BY INCALPER
10-266 5,432 34,436,582
TOTAL 5,432 34,436,582
EN EVENT NUMBER 3.0 NUM 9 11 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
11-813 5,432 34,436,582
TOTAL 5,432 34,436,582
PIDX PERSON ID (ODU+PN) 8.0 CHAR 12 19
VALUE UNWEIGHTED WEIGHTED BY INCALPER
VALID PERSON ID 5,432 34,436,582
TOTAL 5,432 34,436,582
EVENTIDX EVENT ID (ODU+PN+EN) 12.0 CHAR 20 31 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
VALID EVENT ID 5,432 34,436,582
TOTAL 5,432 34,436,582
ROUND DATA COLLECTION ROUND 2.0 NUM 32 33 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 59 378,632
1 1,243 7,955,273
2 1,989 12,549,603
3 1,231 7,736,197
4 910 5,816,876
TOTAL 5,432 34,436,582
LASTAGE ED PID AGE AT END LT ELIG RD 3.0 NUM 34 36 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0-17 956 6,775,785
18-44 1,504 11,048,156
45-64 999 7,238,212
65+ 1,973 9,374,429
TOTAL 5,432 34,436,582
SMPSEXR PID SEX 1.0 NUM 37 37
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 MALE 2,133 13,720,333
2 FEMALE 3,299 20,716,249
TOTAL 5,432 34,436,582
RACE6 ED PID RACE 1.0 NUM 38 38 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 AMER INDIAN 41 263,524
3 ASIAN/PACIFIC 45 401,154
4 BLACK 1,136 4,575,130
5 WHITE 4,057 28,359,769
6 OTHER 153 837,004
TOTAL 5,432 34,436,582
RACE3 PID RACE/ETHNICITY 1.0 NUM 39 39 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 HISPANIC 421 2,120,285
2 BLACK NONHISP 1,132 4,556,718
3 OTHER 3,879 27,759,578
TOTAL 5,432 34,436,582
SREGION PID CENSUS REGION 1.0 NUM 40 40
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 NORTHEAST 1,059 6,966,252
2 MIDWEST 1,390 9,006,066
3 SOUTH 2,050 12,442,611
4 WEST 933 6,021,652
TOTAL 5,432 34,436,582
PLACEIDX G1 ID OF FACILITY 8.0 CHAR 41 48
VALUE UNWEIGHTED WEIGHTED BY INCALPER
VALID PLACE ID 5,432 34,436,582
TOTAL 5,432 34,436,582
DATEBMM G2 DATE PID ENTERED HOSP-MNTH 2.0 NUM 49 50
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 15 79,404
-8 DON'T KNOW 8 34,907
-5 NEVER KNOW 2 5,512
1 JANUARY 480 3,069,914
2 FEBRUARY 429 2,633,507
3 MARCH 473 3,005,815
4 APRIL 467 2,950,052
5 MAY 444 2,843,070
6 JUNE 426 2,760,088
7 JULY 484 3,113,523
8 AUGUST 410 2,628,548
9 SEPTEMBER 414 2,689,164
10 OCTOBER 448 2,838,573
11 NOVEMBER 451 2,794,959
12 DECEMBER 481 2,989,543
TOTAL 5,432 34,436,582
DATEBDD G2 DATE PID ENTERED HOSP-DAY 2.0 NUM 51 52
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 33 172,678
-8 DON'T KNOW 128 639,265
-5 NEVER KNOW 31 146,935
1-31 5,240 33,477,702
TOTAL 5,432 34,436,582
DATEBYY G2 DATE PID ENTERED HOSP-YR 2.0 NUM 53 54
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 12 61,191
-8 DON'T KNOW 2 10,196
-5 NEVER KNOW 1 3,846
85 1 8,465
86 66 408,813
87 5,350 33,944,069
TOTAL 5,432 34,436,582
DATEEMM G3 DATE PID LEFT HOSP-MNTH 2.0 NUM 55 56
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 125 826,365
-8 DON'T KNOW 12 60,658
-5 NEVER KNOW 7 26,462
1 JANUARY 477 3,073,297
2 FEBRUARY 406 2,490,604
3 MARCH 456 2,881,388
4 APRIL 472 2,997,258
5 MAY 449 2,837,850
6 JUNE 417 2,718,414
7 JULY 456 2,956,253
8 AUGUST 414 2,639,498
9 SEPTEMBER 397 2,510,957
10 OCTOBER 442 2,899,864
11 NOVEMBER 419 2,606,863
12 DECEMBER 429 2,621,110
95 STILL IN HOSP 54 289,742
TOTAL 5,432 34,436,582
Variable Positions 57-110
NAME DESCRIPTION FORMAT TYPE START END NOTE
---- ----------- ------ ---- ----- --- ----
DATEEDD G3 DATE PID LEFT HOSP-DAY 2.0 NUM 57 58
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 153 992,915
-8 DON'T KNOW 123 632,386
-5 NEVER KNOW 29 137,372
-1 INAPPLICABLE 54 289,742
1-31 5,073 32,384,167
TOTAL 5,432 34,436,582
DATEEYY G3 DATE PID LEFT HOSP-YEAR 2.0 NUM 59 60
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 58 345,513
-5 NEVER KNOW 2 9,699
-1 INAPPLICABLE 54 289,742
87 5,298 33,671,285
88 20 120,344
TOTAL 5,432 34,436,582
NUMNIGHT G4 # OF NIGHTS PER HOSP STAY 2.0 NUM 61 62
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 160 1,036,417
-8 DON'T KNOW 35 160,347
-1 INAPPLICABLE 5,067 32,367,319
0 18 79,879
1-2 39 236,147
3-5 41 220,318
6-10 42 207,915
11+ 30 128,240
TOTAL 5,432 34,436,582
NUMNGHTX G4 ED # OF NIGHTS PER HOSP STAY 3.0 NUM 63 65 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 497 3,211,233
1-2 1,362 9,327,856
3-5 1,729 11,281,755
6-10 1,003 5,833,612
11+ 841 4,782,125
TOTAL 5,432 34,436,582
NUMNTFLG IMPUTATION FLAG FOR NUMNGHTX 1.0 NUM 66 66
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 FROM HS 2,676 16,619,031
2 FROM MPS 2,599 16,846,690
3 IMPUTED 157 970,861
TOTAL 5,432 34,436,582
NUMNT87X ED # OF 1987 NIGHTS PER HOSP STAY 3.0 NUM 67 69 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 496 3,195,361
1-2 1,392 9,531,367
3-5 1,727 11,253,768
6-10 995 5,788,533
11+ 822 4,667,553
TOTAL 5,432 34,436,582
NUM87FLG IMPUTATION FLAG FOR NUMNT87X 1.0 NUM 70 70
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 FROM HS 2,627 16,379,465
2 FROM MPS 2,650 17,105,529
3 IMPUTED 155 951,588
TOTAL 5,432 34,436,582
HINVA G5 HOSPITAL IS A VA FACILITY 2.0 NUM 71 72
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 60 376,225
-1 INAPPLICABLE 4,700 29,661,930
1 YES 112 646,242
2 NO 560 3,752,185
TOTAL 5,432 34,436,582
RSNINHOS G6 MAIN REASON FOR ADMISSION 2.0 NUM 73 74 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 30 187,518
-1 INAPPLICABLE 31 156,786
1 MEDICAL CONDT 2,464 14,521,352
2 OPER/SURG PRO 1,473 9,416,604
3 DIAGNOSTC TES 276 1,724,698
4 TO HAVE A BAB 508 3,751,830
5 CHILD BORN 500 3,690,243
91 OTHER 150 987,551
TOTAL 5,432 34,436,582
RSNHOSOS G6 OTHER REASON FOR ADMISSION 30.0 CHAR 75 104 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 5,282 33,449,031
-8 DON'T KNOW 1 10,124
-9 NOT ASCERTAIN 2 12,744
TEXT FIELD 147 964,683
TOTAL 5,432 34,436,582
PREGFLG STAY RELATED TO PREGN 1.0 NUM 105 105 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 YES 1,105 8,101,302
2 NO 4,327 26,335,280
TOTAL 5,432 34,436,582
ICD1 G7/G9/G10/G13/G15 ICD9 CODE-COND1 4.0 CHAR 106 109
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 920 6,673,222
-9 NOT ASCERTAIN 44 248,181
V001-V999 125 823,694
X00X-X999 124 778,025
0001-9999 4,219 25,913,459
TOTAL 5,432 34,436,582
ICD2 G7/G9/G10/G13/G15 ICD9 CODE-COND 2 4.0 CHAR 110 113
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 4,908 31,375,257
-9 NOT ASCERTAIN 44 248,181
V001-V999 4 29,417
X00X-X999 16 69,427
0001-9999 460 2,714,300
TOTAL 5,432 34,436,582
Variable Positions 114-150
NAME DESCRIPTION FORMAT TYPE START END NOTE
---- ----------- ------ ---- ----- --- ----
ICD3 G7/G9/G10/G13/G15 ICD9 CODE-COND3 4 CHAR 114 117
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 5,280 33,593,312
-9 NOT ASCERTAIN 44 248,181
X00X-X999 2 10,373
0001-9999 106 584,716
TOTAL 5,432 34,436,582
ICD4 G7/G9/G10/G13/G15 ICD9 CODE-COND4 4.0 CHAR 118 121
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 5,355 33,987,668
-9 NOT ASCERTAIN 44 248,181
X00X-X999 1 5,489
0001-9999 32 195,244
TOTAL 5,432 34,436,582
PROCODE1 G8 1ST PROCEDURE CODE 2.0 CHAR 122 123
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 3,959 25,019,978
-9 NOT ASCERTAIN 54 336,255
00-99 1,419 9,080,349
TOTAL 5,432 34,436,582
PROCODE2 G8 2ND PROCEDURE CODE 2.0 CHAR 124 125
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 5,180 32,720,157
-9 NOT ASCERTAIN 66 409,505
00-99 186 1,306,919
TOTAL 5,432 34,436,582
PROCODE3 G8 3RD PROCEDURE CODE 2.0 CHAR 126 127
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 5,350 33,906,177
-9 NOT ASCERTAIN 58 360,506
00-99 24 169,899
TOTAL 5,432 34,436,582
BABYBORN G11 DELIVERY DURING HOSPITAL STAY 2.0 NUM 128 129
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 1 9,453
-1 INAPPLICABLE 4,921 30,665,482
1 YES 491 3,644,995
2 NO 17 109,170
3 STILLBIRTH 2 7,481
TOTAL 5,432 34,436,582
NORMDELV G12 NORMAL DELIVERY 2.0 NUM 130 131
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 9 51,201
-1 INAPPLICABLE 4,940 30,785,592
1 YES 380 2,800,009
2 NO 103 799,779
TOTAL 5,432 34,436,582
PREGIDX PREGNANCY ID (ODU+COUNTER) 8.0 CHAR 132 139
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 4,327 26,335,280
VALID PREG ID 1,105 8,101,302
TOTAL 5,432 34,436,582
BABYNORM G14 BABY NORMAL 2.0 NUM 140 141
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 6 41,119
-1 INAPPLICABLE 4,932 30,746,339
1 YES 464 3,432,771
2 NO 30 216,353
TOTAL 5,432 34,436,582
ANYOPER G16 OPERATN DURING HOSP STAY 2.0 NUM 142 143
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 110 758,510
-8 DON'T KNOW 1 5,150
-1 INAPPLICABLE 1,538 9,804,742
1 YES 623 4,302,351
2 NO 3,160 19,565,828
TOTAL 5,432 34,436,582
PROCODE4 G17 1ST PROCEDURE CODE 2.0 CHAR 144 145
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 4,811 30,145,224
-9 NOT ASCERTAIN 6 37,559
00-99 615 4,253,798
TOTAL 5,432 34,436,582
PROCODE5 G17 2ND PROCEDURE CODE 2.0 CHAR 146 147
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 5,340 33,878,340
-9 NOT ASCERTAIN 11 66,638
00-99 81 491,603
TOTAL 5,432 34,436,582
PROCODE6 G17 3RD PROCEDURE CODE 2.0 CHAR 148 149
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-1 INAPPLICABLE 5,411 34,291,246
-9 NOT ASCERTAIN 6 37,559
00-99 15 107,777
TOTAL 5,432 34,436,582
EXPFAC G21 ORIGINAL FACILITY CHARGE 9.2 NUM 150 158
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 155 966,449
-8 DON'T KNOW 757 4,424,760
-5 NEVER KNOW 280 1,662,516
-1 INAPPLICABLE 2,039 12,475,062
1 TO 200 97 610,698
GT 200 TO 500 167 1,182,877
GT 500 TO 1000 247 1,764,626
GT 1000 TO 2000 525 3,746,696
GT 2000 TO 5000 711 4,708,631
GT 5000-10000 256 1,674,100
GT 10000 198 1,220,166
TOTAL 5,432 34,436,582
Variable Positions 159-241
NAME DESCRIPTION FORMAT TYPE START END NOTE
---- ----------- ------ ---- ----- --- ----
EXPFACX EDITED FACILITY EXPENSE/STAY 9.2 NUM 159 167 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 2 14,477
1 TO 200 67 361,192
GT 200 TO 500 360 2,595,633
GT 500 TO 1000 669 4,558,741
GT 1000 TO 2000 1,229 8,190,883
GT 2000 TO 5000 1,759 10,820,171
GT 5000-10000 749 4,517,573
GT 10000 597 3,377,910
TOTAL 5,432 34,436,582
EXPFCFLG IMPUTA FLAG FOR EDITED FAC EXP 1.0 NUM 168 168 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 EXP FROM HS 782 5,445,823
2 EXP FROM MPS 2,620 16,311,896
3 EXP IMPUTED 2,030 12,678,862
TOTAL 5,432 34,436,582
SOPFSELF PCT PAYM FROM SELF OR FAMILY-FAC EXP 8.4 NUM 169 176
0 3,752 22,767,565
GT 0 TO 20 1,071 7,429,188
GT 20 TO 40 229 1,738,415
GT 40 TO 60 57 367,257
GT 60 TO 80 28 198,336
GT 80 TO LT 100 11 82,658
100 284 1,853,162
TOTAL 5,432 34,436,582
SOPFPRVT PCT PAYM FROM PRIVATE INSUR-FAC EXP 8.4 NUM 177 184
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 2,547 14,653,427
GT 0 TO 20 678 3,394,072
GT 20 TO 40 114 645,388
GT 40 TO 60 44 307,008
GT 60 TO 80 386 2,960,489
GT 80 TO LT 100 467 3,721,152
100 1,196 8,755,046
TOTAL 5,432 34,436,582
SOPFMCR PCT PAYM FROM MEDICARE-FAC EXP 8.4 NUM 185 192
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 3,779 26,385,688
GT 0 TO 20 43 191,244
GT 20 TO 40 29 152,999
GT 40 TO 60 74 368,940
GT 60 TO 80 662 3,263,716
GT 80 TO LT 100 615 3,002,212
100 230 1,071,782
TOTAL 5,432 34,436,582
SOPFMCD PCT PAYMENT FROM MEDICAID-FAC EXP 8.4 NUM 193 200
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 4,481 29,303,082
GT 0 TO 20 176 808,065
GT 20 TO 40 46 201,406
GT 40 TO 60 23 100,054
GT 60 TO 80 5 19,759
GT 80 TO LT 100 11 49,799
100 690 3,954,418
TOTAL 5,432 34,436,582
SOPFOTFD PCT PAYMENT FROM OTHER FED-FAC EXP 8.4 NUM 201 208
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,022 31,847,056
GT 0 TO 20 10 66,657
GT 20 TO 40 3 14,463
GT 60 TO 80 16 147,661
GT 80 TO LT 100 21 172,791
100 360 2,187,953
TOTAL 5,432 34,436,582
SOPFOTST PCT PAYM FROM OTHER STATE-FAC EXP 8.4 NUM 209 216
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,267 33,503,806
GT 0 TO 20 85 440,674
GT 20 TO 40 18 91,272
GT 40 TO 60 6 40,682
GT 60 TO 80 10 64,937
GT 80 TO LT 100 2 12,666
100 44 282,545
TOTAL 5,432 34,436,582
SOPFWC PCT PAYMENT FROM WORKERS COMP-FAC EXP 8.4 NUM 217 224
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,372 33,968,985
GT 80 TO LT 100 2 8,424
100 58 459,172
TOTAL 5,432 34,436,582
SOPFOTHR PCT PAYM FROM OTHER-FAC EXP 8.4 NUM 225 232
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,301 33,630,675
GT 0 TO 20 35 176,433
GT 20 TO 40 3 13,552
GT 40 TO 60 2 15,208
GT 60 TO 80 14 97,695
GT 80 TO LT 100 15 78,542
100 62 424,477
TOTAL 5,432 34,436,582
SOPFFFP PCT PAYM FREE FROM PROVIDER-FAC EXP 8.4 NUM 233 240
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,240 33,140,087
GT 0 TO 20 42 227,234
GT 20 TO 40 20 147,637
GT 40 TO 60 12 82,067
GT 60 TO 80 16 134,796
GT 80 TO LT 100 54 386,294
100 48 318,467
TOTAL 5,432 34,436,582
SOPFFLG IMPUTATN FLG FOR SOURC PAYM INFO-FAC EXP 1.0 NUM 241 241
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 NO EXPENSES 2 14,477
1 ALL SOP FRM HS 2,945 20,013,910
2 ALL SOP IMPUTD 2,485 14,408,195
TOTAL 5,432 34,436,582
Variable Positions 242-315
NAME DESCRIPTION FORMAT TYPE START END NOTE
---- ----------- ------ ---- ----- --- ----
EXPDRX EDITED DOCTOR EXPENSE 8.2 NUM 242 249 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 1,744 11,416,150
1 TO 50 41 212,869
GT 50 TO 100 85 591,575
GT 100 TO 200 368 2,355,094
GT 200 TO 500 952 6,051,818
GT 500 TO 1000 696 4,207,304
GT 1000 TO 5000 1,356 8,423,293
GT 5000 190 1,178,477
TOTAL 5,432 34,436,582
EXPDRFLG IMPUTAT FLAG FOR EDITED DOC EXPENSE 1.0 NUM 250 250*
VALUE UNWEIGHTED WEIGHTED BY INCALPER
2 EXP FROM MPS 3,841 24,371,514
3 EXP IMPUTED 1,591 10,065,067
TOTAL 5,432 34,436,582
SOPDSELF PCT PAYM FROM SELF OR FAMILY-DR EXP 8.4 NUM 251 258
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 3,997 24,891,336
GT 0 TO 20 664 4,496,174
GT 20 TO 40 218 1,516,954
GT 40 TO 60 149 993,846
GT 60 TO 80 49 255,081
GT 80 TO LT 100 16 94,829
100 339 2,188,361
TOTAL 5,432 34,436,582
SOPDPRVT PCT PAYM FROM PRIVATE INSUR-DR EXP 8.4 NUM 259 266
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 3,685 22,656,333
GT 0 TO 20 337 1,750,661
GT 20 TO 40 141 743,321
GT 40 TO 60 108 727,120
GT 60 TO 80 283 2,182,067
GT 80 TO LT 100 193 1,604,861
100 685 4,772,220
TOTAL 5,432 34,436,582
SOPDMCR PCT PAYM FROM MEDICARE-DR EXP 8.4 NUM 267 274
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 4,330 29,012,597
GT 0 TO 20 33 145,199
GT 20 TO 40 60 300,322
GT 40 TO 60 131 690,797
GT 60 TO 80 623 3,097,476
GT 80 TO LT 100 123 596,682
100 132 593,510
TOTAL 5,432 34,436,582
SOPDMCD PCT PAYM FROM MEDICAID-DR EXP 8.4 NUM 275 282
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 4,846 31,269,781
GT 0 TO 20 133 625,710
GT 20 TO 40 42 186,525
GT 40 TO 60 15 69,867
GT 60 TO 80 6 53,814
GT 80 TO LT 100 5 26,590
100 385 2,204,294
TOTAL 5,432 34,436,582
SOPDOTFD PCT PAYM FROM OTHER FEDERAL-DR EXP 8.4 NUM 283 290
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,073 32,298,114
GT 0 TO 20 1 3,161
GT 20 TO 40 2 8,125
GT 40 TO 60 3 27,347
GT 60 TO 80 6 56,488
GT 80 TO LT 100 3 19,598
100 344 2,023,747
TOTAL 5,432 34,436,582
SOPDOTST PCT PAYM FROM OTHER STATE-DR EXP 8.4 NUM 291 298
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,272 33,555,939
GT 0 TO 20 57 307,058
GT 20 TO 40 35 160,652
GT 40 TO 60 5 31,319
GT 60 TO 80 1 4,621
100 62 376,992
TOTAL 5,432 34,436,582
SOPDWC PCT PAYM FROM WORKERS COMP-DR EXP 8.4 NUM 299 306
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,400 34,193,640
GT 60 TO 80 1 3,907
100 31 239,035
TOTAL 5,432 34,436,582
SOPDOTHR PCT PAYM FROM OTHER-DR EXP 8.4 NUM 307 314
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,315 33,605,904
GT 0 TO 20 14 63,478
GT 20 TO 40 3 18,748
GT 40 TO 60 9 57,164
GT 60 TO 80 23 149,671
GT 80 TO LT 100 21 166,915
100 47 374,701
TOTAL 5,432 34,436,582
SOPDFFP PCT PAYM FREE FROM PROVIDER-DR EXP 8.4 NUM 315 322
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 5,322 33,631,581
GT 0 TO 20 44 330,016
GT 20 TO 40 14 80,343
GT 40 TO 60 8 61,189
GT 60 TO 80 13 104,239
GT 80 TO LT 100 12 107,976
100 19 121,238
TOTAL 5,432 34,436,582
Variable Positions 323-344
NAME DESCRIPTION FORMAT TYPE START END NOTE
---- ----------- ------ ---- ----- --- ----
SOPDFLG IMPUTA FLG FOR SRC PAYM INFO-DOC EXP 1 NUM 323 323 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 NO EXPENSES 1,744 11,416,150
1 ALL SOP FRM HS 1,889 12,573,786
2 ALL SOP IMPUTD 1,799 10,446,646
TOTAL 5,432 34,436,582
EXPTOTX EDITED TOTAL EXPENSE/HOSP STAY 9.2 NUM 324 332 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
0 2 14,477
1 TO 200 50 269,721
GT 200 TO 500 254 1,818,538
GT 500 TO 1000 561 3,862,486
GT 1000 TO 2000 1,015 6,802,907
GT 2000 TO 5000 1,795 11,222,400
GT 5000-10000 977 5,977,290
GT 10000 778 4,468,762
TOTAL 5,432 34,436,582
NURSCARE G58 SPECIAL/PRIV DUTY NRS REQUIRED 2.0 NUM 333 334
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 244 1,619,710
-8 DON'T KNOW 10 73,945
-1 INAPPLICABLE 54 296,694
1 YES 135 798,809
2 NO 4,989 31,647,424
TOTAL 5,432 34,436,582
EXPNRS G62 TOTAL NURSE EXPENSE/STAY 7.2 NUM 335 341
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 108 644,209
-1 INAPPLICABLE 5,297 33,637,773
GT 0 27 154,600
TOTAL 5,432 34,436,582
IHSFAC G77 HOSP IS IHS OR TRIBE/ANC 2.0 NUM 342 343
VALUE UNWEIGHTED WEIGHTED BY INCALPER
-9 NOT ASCERTAIN 24 146,093
-1 INAPPLICABLE 5,385 34,144,761
2 TRIBE/ANC 4 25,953
3 NOT IHS/TRIBE 19 119,775
TOTAL 5,432 34,436,582
INCALPER FULL-YEAR WEIGHT 12.6 NUM 344 355 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
781.59-24900.84 5,432 N/A
TOTAL 5,432 N/A
STRATUMX SAMPLING STRATUM 3.0 NUM 356 358
VALUE UNWEIGHTED WEIGHTED BY INCALPER
301-401 5,432 34,436,582
TOTAL 5,432 34,436,582
SPSU PSEUDO PSU 1.0 NUM 359 359 *
VALUE UNWEIGHTED WEIGHTED BY INCALPER
1 2,758 17,235,016
2 2,674 17,201,566
TOTAL 5,432 34,436,582
CODEBOOK NOTES
An asterisk in the rightmost column of the codebook indicates
that an explanatory note provides greater detail on information
necessary for the use of this variable. These notes are listed
below in alphabetical order by variable name.
VARIABLE NOTE
EN EN is an event number where each event is a unique
EVENTIDX hospital stay. EVENTIDX is the unique record
identifier on the file for these events.
EXPDRX Constructed expense variables. The constructed
EXPFACX doctor expense variable (EXPDRX) is equal to the
EXPTOTX sum of expenses for all physicians billing
separately for inpatient services during the
hospital stay. The constructed facility expense
variable (EXPFACX) is the sum of all expenses for
hospital care, including room and board,
laboratory charges, x-rays, and services provided
by physicians not billing independently from the
hospital. The total hospital expense variable
(EXPTOTX) is the sum of facility expense (EXPFACX)
and physician expense (EXPDRX). See section 2.3.1
for details.
EXPDRFLG A flag indicating whether the physician expense
for that record was imputed or provided from the
Medical Provider Survey.
EXPFCFLG A flag indicating whether the facility expense for
that record was imputed, provided by the household
respondent or taken from the Medical Provider
Survey.
INCALPER Weight adjusted for nonresponse and post-
stratified to U.S. Census data. Estimates of
utilization and expenditure for hospital stays
require the use of weighted data. For details on
this weight, see the Technical and Programming
section on this file and the hard copy
documentation accompanying this tape.
LASTAGE An edited variable which identifies person age, in
years, as of the end of the last round in 1987 for
which the person was eligible. Less than 0.1
percent of the cases were edited.
NUMNGHTX NUMNGHTX is an edited variable indicating the
NUMNTFLG number of nights in the hospital for this stay.
See section 2.1 for details. NUMNTFLG
is a flag indicating whether the number
of nights for the record was imputed,
provided by the household respondent, or
taken from the MPS.
NUMNT87X NUMNT87X is an edited variable indicating
NUM87FLG the number of hospital nights in 1987
for this stay. Stays starting in 1986 or ending
in 1988 were truncated to show only the nights in
the hospital in 1987. NUM87FLG is a flag
indicating whether the number of nights in 1987
for the record was imputed, provided by the
household respondent or taken from the MPS.
PLACEIDX Responses to question G1 were converted to
numerical identifiers for the facility where the
hospital stay occurred. This variable may be used
as a link to similar variables on NMES Pubic Use
Tapes 9, 14.2, and 14.5 (usual source of care,
home health visit, and ambulatory medical visit
data). The identifier consists of the ODUX
variable plus a three digit number, which was
assigned to each facility reported within a
dwelling unit. Since this number was assigned
within a dwelling unit, this identifier is not
unique across dwelling units. In order to link
the files all identification variables provided
for medical providers, physicians and place of
care (PLACEIDX, DOCRIDX, PROVIDX and IHSANIDX; the
last three are not provided on the current tape
but on NMES Tapes 9, 14.2 and 14.5) must first be
renamed to a common variable name. Consistancy
across files for provider type has not been
verified, e.g., a medical provider identified on
the current file may link to a place of care on
another file rather than a medical provider. It
should be noted that if a respondent saw a
provider in more than one location then it was
possible to have a different ID for each location.
PREGFLG A constructed variable which indicates whether the
stay was associated with a pregnancy. Pregnancy
related events were identified using the following
criteria: i) any condition code in the ranges
6300-6769, V220-V222, V230-V239, V240-242 and
V300-392; ii) a procedure code in the range 72-75;
iii) category 4 or 5 given for the visit in
RSNINHOS; iv) coverage by a flat fee which also
covers events having any of characteristics i) -
iii). It is set to 1 for both mothers and
newborns. To identify all newborn stays in which
the birth occurred use RSNINHOS=5.
PREGIDX An identification variable to link hospital
records for delivering mothers and newborn babies.
RACE3 A person-level variable constructed to facilitate
the poststratification of the NMES person-level
sampling weights by race and ethnicity,
considering three mutually exclusive
classifications: Hispanic, Black--non-Hispanic,
and white or other.
RACE6 An edited variable indicating race. Less than 0.2
percent of cases were edited.
ROUND Indicator of the round of data collection.
Reference period dates specific to each person's
rounds of data collection are provided on NMES
Public Use Tape 13.
RSNINHOS Categories 4 and 5 can be used to identify most
pregnancy related stays. To identify all
pregnancy related stays, see PREGFLG. To identify
all newborn stays in which a birth occurred
category 5 should be used.
SOPDFLG Flags indicating whether all the source of payment
SOPFFLG percentages for the physician (SOPDFLG) and
facility (SOPFFLG) expense for a stay were
provided by the household respondent or imputed.
SPSU To obtain variance estimates of sample statistics
STRATUMX by means of standard statistical programs that use
the Taylor series linearization method of variance
estimation, variables must be used that denote the
strata and primary sampling unit (PSU) within a
given stratum. The variables STRATUMX and SPSU
are these variables, respectively. For details on
sampling weights and variance estimation, see the
Technical and Programming section on this file and
the hard copy documentation accompanying this
