INTRODUCTION

Data on causes of death are released by NCHS in a variety of ways including 
published reports, special tabulations to answer data requests, and public-
use data.  Since the inception of the multiple cause-of-death program in 1968, 
a public-use data file has been released for each data year.  Each file 
contains a data record for all deaths processed by NCHS.  Each data record 
contains underlying cause, multiple cause, and demographic data for a death.

With the exception of calendar years 1972, 1981 and 1982, all deaths occurring 
annually in the United States are processed.  In 1972, underlying and multiple 
cause data were coded and processed for only 50 percent of the deaths 
occurring in each State.  In 1981 and 1982, multiple cause data were coded on 
a 50 percent sample basis for deaths occurring in 19 registration areas.  The 
registration areas are the 50 States, New York City and the District of 
Columbia.  The 50 percent sample States are identified in the documentation of 
the 1981 and 1982 files.  For the remaining 33 registration areas, multiple 
cause data were processed on a 100 percent basis.  In 1981 and 1982, 
underlying cause, demographic, and geographic data were processed for every 
death occurring in every State; however the multiple cause-of-death
public-use data contains only those records where the multiple cause field is
also coded.  A public-use data file containing underlying cause, demographic,
and geographic data for every death in the United States is available but contains
no multiple cause data.

This document is intended to provide guidance to the consumer in accessing and
utilizing the multiple cause-of-death public-use data file for 1995.  It
provides the technical data processing information necessary to access the 
datas and the classification structure and coding rules applied to create 
each variable on the file such that the user can readily assess relevance at 
varying levels of detail to his/her own particular research.  Additionally, 
it conveys the characteristics of the multiple cause files sufficient to guide 
the user in analyzing and interpreting multiple cause data.  The user is
alerted to certain pitfalls of interpretation; and the appropriateness of 
each type of multiple cause data to given applications is discussed.  Control
totals are also provided for comparison with user generated counts for 1995
data.

A revised U.S. Standard Certificate of Death was recommended for State use 
beginning on January 1, 1989.  Among the changes were the addition of a new 
item on educational attainment and changes to improve the medical 
certification of cause of death.  In addition, for the first time, the U.S. 
Standard Certificate of Death includes a question on the Hispanic origin of 
the decedent.  Previously a number of States had included an Hispanic-origin 
identifier on their certificates.  A change was also made in the format of 
the item to obtain information on type of place of death from an open-ended 
question to a checkbox.

The Office of Management and Budget revised its designation of metropolitan 
statistical areas based on figures from the 1990 Census. Effective with the 
1990 through 1993 data files, NCHS has been using these new definitions and
codes as indicated in the listing of 320 Metropolitan Statistical Areas
(MSAs), Primary Metropolitan Statistical Areas (PMSAs), and New England County
Metropolitan Areas (NECMAs) included in the documentation for these years.
There are also 20 Consolidated Metropolitan Statistical Areas (CMSAs) which

                                    - 1 -

are made up of PMSAs.  Because other geographic changes based on the 1990
Census became effective with the 1994 data file, the metropolitan statistical
area designations were updated as well.  Effective with the 1994 data file,
there are 311 MSA's, PMSA's, and NECMA's; and 18 CMSA's as indicated in the
listing included with this documentation.

NCHS has adopted a new policy on release of vital statistics unit record data
files.  This new policy was implemented for the 1989 vital event files to 
prevent the inadvertent disclosure of individuals and institutions.  As a 
result, the files for 1989 and later years do not contain the actual day of 
the death or the date of birth of the decedent. The geographic detail is 
also restricted; only counties and cities of 100,000 or more population 
based on the 1980 Census, as well as metropolitan areas of 100,000 or more 
population based on the 1990 Census are identified.












































                                    - 2 -