History
Hispanic ethnicity was added to the death certificate
as a separate checkbox item in 1989.
Mortality statistics for Hispanics were reported in New Jersey
Health Statistics, 1989, however, concerns about the quality of
the data on Hispanic ethnicity led to the discontinuation of publication
of such statistics after 1990.
Reported Hispanic Mortality Data
Hispanics comprised 9.6 percent of New Jersey's population
in 1990. By 2000, the percentage
had risen to 13.3. As the Hispanic
population in New Jersey continues to grow, the need for data on the
mortality experience of this segment of the population increases. Beginning with New Jersey Health
Statistics, 1999, CHS will once again publish Hispanic mortality
data with the understanding that it is incomplete.
The following tables have been created to
bridge the gap between the 1989 and 1999 published data.
As the tables show, reported deaths
for Hispanics do not exhibit stable patterns.
For example, there is a large decline in reported deaths to Hispanics
between 1989 and 1990, and a similar decline between 1991 and 1992. Since 1994, the annual number of reported deaths
has remained roughly constant at about 2,500, yet this number is consistently
lower than what would be expected based on the growing Hispanic population.
As can be seen, the crude death rate among Hispanics is approximately one-third
that for the total population. In the case of age-adjusted death rates, the Hispanic rate is slightly more
than half that of the total population.
Discrepancies of this size are clear indicators of undermeasurement
of Hispanic ethnicity on death certificates. Therefore, while these figures provide some
information about Hispanic mortality, they should be used with caution
and almost certainly do not represent a complete measurement of deaths
to Hispanics.
Background and Analysis
Undermeasurement of
Hispanic ethnicity on death certificates is a nationwide problem. The National
Center for Health Statistics estimates that the undercount is approximately
7 percent nationally1. While
Hispanics are thought to also be undermeasured in the Census, and thus
in the denominator of mortality rates, this is believed to be more than
offset by the undermeasurement on death certificates.
A study done by New Jersey's CHS in 2001 using a Spanish surname
list developed by the U.S. Census
Bureau2 estimated New Jersey's undercount to be approximately
8.4 percent for 1994-1998 data. It
has not been determined if the undercount is uniform across time, age
groups, sex, causes of deaths, countries of origin, and counties of
residence. Therefore, it may be inaccurate to simply adjust
all data upward by 8.4 percent.
The Spanish surname
list has some limitations as a tool to address undermeasurement of Hispanic
ethnicity, since the surnames on the list do not correspond very closely
with the surnames of Hispanics in New Jersey.
For example, many non-Hispanics are included in the universe
of potential Hispanics created by the list.
A 20 percent sample of New Jersey death certificates identified
as Hispanic through use of the surname list (n=439) was examined manually. Based on what was marked on the certificate for ethnicity and also
for race, birthplace, and mother's name, it was determined that only
51.5 percent of the sampled certificates were truly Hispanic.
It is also unfortunately
the case that many Hispanics do not have surnames that are on the list. Only 64.8 percent of New Jersey death records
that were already coded as Hispanic matched this Spanish surname list.
It appears that the list may not be well suited to the countries
of origins of New Jersey's Hispanic population.
While 74.6 percent of those coded as Puerto Rican had names on
the list, that was the case for only 59.3 percent of Cubans, 56.4 percent
of Mexicans, 52.6 percent of Central and South Americans, and 51.7 percent
of other Hispanics.
Further, the surname
analysis only captured exact matches and did not account for misspellings
or hyphenated names. Additionally,
some decedents had one Hispanic parent and one non-Hispanic parent. Yet mother's surname is not on the electronic
death file, thereby making it impossible to make determinations for
persons of mixed ethnicity without searching each paper certificate
manually. Without more information
from the death certificate in the electronic file (specifically, mother's
surname and birthplace for those currently coded as "remainder of world"),
CHS cannot use the list of Spanish surnames to estimate the true mortality
experience of Hispanics in New Jersey on a routine basis.
The complete implementation of the Electronic Death Registry
System in 2005 should alleviate the undercount problem to some degree.
Hispanic Death Data Tables
References
1 Rosenberg
HM, Maurer JD, Sorlie PD, Johnson NJ, et al. Quality of Death Rates
by Race and Hispanic Origin: A Summary of Current Research, 1999.
National Center for Health Statistics. Vital and Health Statistics 2(128).
1999. http://www.cdc.gov/nchs/data/series/sr_02/sr02_128.pdf
2 Word
DL and Perkins RC. Building
a Spanish Surname List for the 1990's- A New Approach to an Old Problem. US Bureau of the Census,
Population Division. Technical Working Paper No. 13.
1996. http://www.census.gov/population/documentation/twpno13.pdf
