Wide gaps in mortality risk : comparisons among race - sex groups across time and space dimensions

J 2.

North Carolina State LiDrary

STUDIES

N. a

Doa

A SPECIAL REPORT SERIES BY THE N.C. DEPARTMENT OF HUMAN RESOURCES, DIVISION OF
HEALTH SERVICES, STATE CENTER FOR HEALTH STATISTICS, P.O. BOX 2091, RALEIGH, N.C. 27602

No. 18

ji^^^

o

\9S0

August I9B0

WIDE GAPS IN MORTALITY RISK:
COMPARISONS AMONG RACE-SEX GROUPS
ACROSS TIME AND SPACE DIMENSIONS

Effective May 1, 1980, the State Center for Health Statistics was created.
Accordingly y our special-study monograph series has a new name — SCES STUDIES replacing
the former PHSB STUDIES. In the future, we hope to be more prolific than in the recent
past, and as always, we welcome your suggestions for improvement. Now to the subject
at hand . . .

For many conditions contributing to death, prevention is possible and chances of
survival are good when cases are diagnosed and treated at an early stage. Hence,
through the analysis of risk patterns among population groups, public health programs
can play a major role in reducing morbidity and mortality by concentrating their efforts
on those groups at greatest risk.

As discussed in a prior publication on leading causes of mortality in North
Carolina (1), certain causes of death are associated with wide gaps between the sexes
and between races with males and nonwhites experiencing substantially higher death
rates than their female and white counterparts. In particular, the publication reveals
the following recent patterns in North Carolina:

- Male death rates approach or exceed twice the female rates for 8
major causes: acute myocardial infarction, lung cancer, chronic
obstructive lung diseases, motor vehicle and other accidents,
cirrhosis, suicide and homicide.

- Nonwhite death rates approach or exceed twice the white rates for
6 causes: cancer of the stomach, cervix, prostate; hypertension;
nephr i t i s/nephrosis and homicide.

Differential age structures can account for mortality differences since, obviously,
an older population will experience more deaths and hence higher death rates unless we
adjust for age. Thus, in order to identify those race-sex groups most in need of
particular kinds of service, the present effort examines age-adjusted rates for major
underlying causes of death. Rates for race-sex groups within health service areas
(HSA's), North Carolina and the United States for each year 1973-77 allow for trending
over time (final 1978 U.S. data are not available). In addition, rates for the period
197^-78 are computed for race-sex groups at the county, MSA and state levels. Due to
the high costs involved, comparable data for a prior time period have not been generated.

Consistent with procedures of the National Center for Health Statistics (2), all
rates are adjusted by the direct method using ten-year age intervals and the 19^0 Census
of the total U.S. population as the standard. This allows for comparisons across race-
sex groups, years and geographical areas. All U.S. data are final mortality statistics
published annually by the National Center for Health Statistics as in reference 2.

eo

u%

^

o

^

oo

—

.

f*\

oo

—

OD

■a-

o

OO

V

<J

0^

d

0^

r>%

eg

so

^

SO

SO

eo

f^

N

o

d

OD

w V

eg

eg

e«(

rr\

lA

eg

lA

o

rsi

X

C V

'

•

'

'

'

'

<

•

'

•

+

1

+

T

1

<n

J-

o

w\

eg

•A

^

eo

<

r**

<r%

f«-t

^

^

o^

0 it-

(/I

Z

z

Ln

nO

v\

r*.

o^

#^

r^

so

eg

LA

eg

3

1

eg

eg

1

eg

1

1

1

1

T

eg

1

*

f*\

r^

-a-

ut

r^

lA

ffv

so

^

eg

o

o

SO

OX

so

-3-

2^ *

U

O

o

f^

ro

OO

so

oo

eg

r*>

eg

0^

LA

since
Male

z

r-»

eg

eg

*^

eg

-3-

»*%

-X

m

n-t

1

so

1
eg

1
eg

fv.

+

OO

r-

1
o

+
<

z

+
-*

+

1

1
LA

1
SO

1
O

i*%

**%

u^

u-v

•»

-ar

eg

o

r*.

o

lA

r^

OO

r*

3

n^

eg

eg

eg

n^

eg

eg

V

i

i

1

+

1

1

1

1

I

+

1

CT

C

^

\0

r^

OD

oo

o

rs.

O

lA

LA

so

SO

so

(->

<->

« «,j;

n-v

NO

SO

lA

iA

e^

so

»«^

so

^

a\

OD

LA

LA

CD

Z

f^

eg

eg

e*

eg

1

1

I

1

+

1

1

+

+

t

1

+

vO

.3-

r>.

O^

so

-3-

-3-

O

<

CD

<r»

LA

-T

SO

Ifl

z

-3-

o

f»>

e>>

eg

O

f^

eg

J-

f>

a.

3

I

1

e4

1

eg

1

eg

1

+

eg

1

1

t

1

"

"

+

+

oo

^

ff\

O

■a-

f^

^

%o

sO

eg

o

-T

O

OO

-3-

U

a\

p^

eg

ff>

u^

<**

O

eg

d

lA

eg

r-

OO

il "

z

*n

eg

N

eg

eg

1

1

1

t

1

+

1

1

+

'

•

1

'

1

1

i-

^

O

r^

oo

O

-3-

r-»

•^

<

lA

r-

lA

OO

eg

^

t/1

z

o

f^

m

r^

o

eg

CO

sO

n-\

sO

n^

-3-

O

=9

1

1

1

eg

1

eg

1

+

1

eg

1

"

T

1

+

+

«->

d

eg

-*

\0

o

-3-

-3-

eg

00
sO

lA
ON

d

LA

-T

SO

eg

|i

Z

J'

<n

oo

e4

eg

C V

*>^

fo

f-.

sX>

o

-;

-a-

O

O

<
Z

o

-»

<r.

(r\

r^

lA

CO

**->

oo

lA

est

eg

-»

lA

d.

LA

f^

eg

3

oo

so

eg

eg

r-^

f^

eg

eg

OO

r*'

eg

<r»

IS\

^

**%

sO

LA

0^

oo

V

^ c r

5^

«->

eg

J"

oc

d

00

r4

so

^

r^

__'

-3-

f^

<T\

^

so

vO

r^

ri^

LT*

SO

J"

z

r*j

m

e-i

a^

CO

so

eo

^

J-

•*\

so

<

z

-3-

O

CO

r^

so

-

w^

lA

r^

-»

o^

so

lA

^

<r\

m

LA

OO

eg

o

-3-

a^

r*.

O

eg

»*%

m

J"

sO

3

=)

O

eg

eg

a

o

a.

~~^

~~^

"

o

OO

^

eg

00

-3-

^

-3"

O

eg

f^

n-i

eg

CD

Ln

LA

o

(_>

°.*>i'

f^

eg

OD

lA

r*

o>

r*«.

sO

eg

e-i

O

r-^

P^

z

i*\

-T

o^

r 100

Whit
Femal

J-

oo

eg

lA

lA

r^

o

eo

<

lA

so

eg

SO

eg

ff\

«1

z

r*

r^

o^

00

o^

C^

r^

eg

O

p*-

eg

V

=J

eg

rr\

-3-

O

a

^

"^

£

«M

l*%

00

-T

lA

eg

o

o

oo

oo

so

<T%

LA

sO

LA

fO

(_>

41

eg

r>.

r^

f*^

a^

d

eg

o

LA

*»^

-»*

SO

f^

" ii «

z

NO

r^

so

lA

eg

r^

r»^

e^

eg

OD

rf\

ix'

lA

O

«o

lA

eg

o

OO

„

<

r^

CO

eg

O

so

eo

to

z

-»

d

r^

d

0^

CO

sO

eg

e-v

d

d

OD

3

OO

cn
eg

lA

sO

<*\

e^

eg

«)

tf)

«

«

M

a

Ol

c

M

w

a

*«

c

^

C

>■ «

•

«

m

•s

— tfl

>

k

1. 9

o

•ft

u

« «

«i

>

O

u

"O t>

u

1.

<

«A

c

lA

3

_l

£

3

«

O

•

U

a

«

C

lA

c

c

H-

«

V

«

0

«

a.

tfl

o

z

u

■D

u>

«

*v

A

"V.

«1

tA

«n

u

«

O

tf)

IA

X

"u

tft

c

lA

tn

N

V

u

1)

3

o

V

o

«

C

u

VI

>

<

V

■D

«

«->

«

u

v

o

■D

W

A'

JC

h.

«

V

3

c

£

U

L.

L.

u

«

0

V

u

J3

o

X

o

«

U

<0

o.

k.

c

«

«»-

u

o.

X

1

«

>

«

C

£.

«

1

"5

<

z

X

Irt

<

u

O

u

u

z

o

«/»

X

Age-adjusted Mortality: United States and North Carolina

Table 1 provides U.S. and N.C. rates for 1977 with corresponding percent changes
since 1973- For total deaths and the five leading causes, U.S. and N.C. race-sex
trends are depicted in Figures 1-6. These graphs use the logarithmic scale in order
to depict the relat i ve (as opposed to absolute) changes in death rates.

Age-adjusted Mortality: North Carolina Health Service Areas and Counties

Table 2 shows HSA total adjusted rates for 1978 and percent changes since 1973.
These rates show the greatest differences among HSA's to involve excessive nonwhite
mortality in the Southern Piedmont and excessive white mortality in the Cardinal.
In general, there exists some tendency towards an inverse relationship between white
and nonwhite mortality, e.g., the three easternmost HSA's are experiencing above-
average white mortality and below-average nonwhite mortality.

Examining cause-specific mortality among the HSA's, Table 3 lists an HSA if its
197^"78 cause-race-sex-specific rate exceeded the corresponding N.C. rate by 10^ or
more. Single-year HSA data for the period are also available.

For counties, five-year total age-adjusted rates for race-sex groups are depicted
in Figure 7 where counties are grouped according to quintile. Counties at the upper
end of a range of death rates should carefully consider the present analysis and
request their own cause-specific data in order to ferret out the causes of excessive
mortality in one or more race-sex groups.

Table 2

1978 Age-adjusted Death Rates with Percent Changes
Since 1973. Race-sex Groups

N.C. Health Service Areas

HSA

Race and Sex |

White
Male

White
Female

Nonwhi te
Male

Nonwh i te
Female

1978

Percent
Change

1978

Percent
Change

1978

Percent
Change

1978

Percent
Change

Western

799-5

-10.6

399.5

-13.9

1135.8

-17.3

702.1.

-20.3

Piedmont

8U.'<

-12.1.

ItOli.l

-1I..2

121.6.7

- 9.1

655-9

-20.1.

Southern Piedmont

817.1

-10.2

1.00.7

-12.8

1359.7

+ 1.3

723.2

-16.7

Capital

922.6

- 5.1.

1.32.3

- 9.1.

111.7.3

-13-9

51.9.1

-23-9

Cardinal

Sit?. 6

-12.3

1.57.7

-15.1

1155.1.

-18.9

620.2

-25.2

Eastern

916-9

-11.1.

1.53.7

-13.7

1192.5

-12.4

621.8

-18.3

North Carol ina

850.9

-11.0

1.18.9

-13-1.

1201.9

-12.1

631.0

-20.9

Di scussion

Due to the variability often associated with small numbers, some counties —
especially western counties — may have abnormally high or low rates, especially non-
white rates. This is the case with high nonwhite male rates in Mitchell and Cherol<ee
and the high nonwhite female rate in Alleghany. Otherwise, age-adjusted rates by
county and cause — available for the asking — should help researchers and others to
"zero in" on the particular mortality risks of race-sex groups in local areas.

The graphs below display age-adjusted death rates by race for sex groups (see
key) in the U.S. and N.C. The N.C. data are shown in red; the rate is the
number of deaths per 100,000 population.

RATE

lOOOOr-

WHITES

I

'■•j w

tool—

73

YEAR

WHITES

NONWHITES

FIG. 1 TOTAL MORTALITY

Except for white females, each
N.C. race-sex group exceeds
its U.S. counterpart in age-
adjusted mortality. N.C.
females of both races have
recently experienced good rate
reductions but N.C. males of
both races are lagging behind
females as well as U.S. males.

FIG, 2 HEART DISEASE MORTALITY

North Carolina males are at
greater risk than are U.S.
males, and reductions in male
heart disease mortality have
been less in North Carolina
than in the U.S.

FIG. 3 STROKE MORTALITY

For all four race-sex groups,
stroke mortality is substan-
tially higher in N.C. than in
the U.S. Rate reductions in
the state and nation have been
comparable except that N.C.
nonwhite females are slightly
ahead of their U.S. counter-
parts.

NONWHITES

NONWH I TES

T

The graphs below display age-adjusted death rates by race for sex groups (see
key) in the U.S. and N.C. The N.C. data are shown in red; the rate is the
number of deaths per 100,000 population.

RATE
lOOO-

WHITES

T

->•«-

-*-

lOl—
73

75

YEAR

WHITES

FIG. k CANCER MORTALITY

For all but nonwhite males,
North Carolinians are at less
risk of death from cancer than
are other Americans. However,
except for nonwhite females,
total cancer mortality is
rising faster in N.C. than
nationwide. By site,
increases involve colon/rectum,
especially among nonwhite
males ; pancreas among females ,
especially nonwhites ; lung,
especially females; breast
among white females and
prostate among nonwhites.

FIG. 5

MOTOR VEHICLE ACCIDENT MORTALITY

N.C. exceeds the nation with
the nonwhite male rate
exceeding the U.S. rate by
59%. On a positive note,
however, all race-sex groups
in N.C. have experienced
higher declines than their
U.S. counterparts. The N.C.
nonwhite female rate is down
by more than half.

RATE
1000|

NONWHITES

I

^

10
73

7M

7S

YEAR

NONWHITES

FIG. 6

NON-MOTOR-VEHICLE ACCIDENT MORTALITY

N.C.'s recent experience is
disturbing. Not only do all
race-sex groups, especially
nonwhites, surpass their U.S.
counterpcirts in death rates,
but recent improvements have
been substantially less in
N.C, except for white
females.

NONWH I TES

T

Table 3

Health Service Areas Experiencing 197''-78 Age-adjusted
Death Rates 101 or More Above Corresponding Statewide Rates

Underlying
Cause

Race-sex

Groups

White Male

Whi te Female

Nonwhi te Male

Nonwhite Female

Heart Disease

Cardinal, Eastern

Cardinal

S. Piedmont

S. Piedmont

Hypertens ion

Cardinal , Eastern

Cardinal, Eastern,
Capital

Cardinal , Eastern

Cardinal . S . Pi edmont

Strolce

Cardinal, Eastern

Cardinal

Eastern

Cardinal . Eastern

Arter iosclerosis

S. Piedmont, Cardinal

Cardinal

5. Piedmont, Cardinal

S. Piedmont. Cardinal,
Western

Cancer

Eastern

Piedmont, S. Piedmont

Stomach

Western

Western, Cardinal
Eastern

Western

Colon/Rectum

Piedmont, S. Piedmont

Piedmont, S. Piedmont

Pancreas

Capital

Western

Western, S. Piedmont

Western, S. Piedmont,
Piedmont

Trachea. Bronchus
and Lung

Cardinal, Eastern

Eastern

Western, S. Piedmont,
Capi tal

Female Breast

Capital

Piedmont

Cervix Uteri

Cardinal

Eastern

Ovary, Fallopian Tube
and Broad Ligament

Piedmont

Western, Eastern

Prostate

Piedmont

Leukemia

Cap! tat

Capi tal , Cardinal

Capital , Piedmont

Cardinal , Piedmont ,
S. Piedmont

Diabetes Hel 1 i tus

Cardinal

S. Piedmont

S. Piedmont, Western,
Piedmont

Cardinal

Inf luenza/Pneumon ia

Eastern

Eastern

S. Piedmont

Western

Chronic Obstructive
Lung

Cirrhosis of the Liver

Cardinal

Capital, Cardinal,
Eastern

Capital

Capital , Cardinal ,
S. Piedmont, Eastern

Western, Piedmont,

S. Piedmont
S. Piedmont, Piedmont

Western, Capital

Piedmont, S. Piedmont,
Western

Nephr i t i s/Nephros i s

S. Piedmont. Cardinal

S. Plednont, Cardinal,
Capital

S. Piedmont

S. Piedmont, Cardinal,
Piedmont

Motor Vehicle Accidents

Cardinal , Eastern

Cardinal, Eastern

Cardinal

Cardinal, Eastern

Other Accidents

Western, Cardinal,
Eastern

Western

Suicide

Capital

Capital

Capital, Piedmont,
Eastern

Western, S. Piedmont

Homicide

Western, Cardinal

Cardinal, Eastern

Western, S. Piedmont

Western, S. Piedmont,
Capital

All Causes

Cardinal, Eastern

CO
Q.

e>

X

UJ 1^

o a>

o o>
I^ll. —

o
<

I

UJ

to

UJ

UJ
U-

UJ

I-

<

— « 2 •

ui o 9) » r«
. (D s m lo

■oaao

UJ

UJ

I-

CO

X

asG3aa

■SOGQ

In this report, we have deliberately not shown rates for total populations (all
races and sexes combined) since race and sex then become confounding factors.

This study of mortality risk has shown that, after adjustment for age, wide
gaps exist between N.C. and the U.S., and within N.C., among HSA's, between the races
and between the sexes. In fact, comparisons of adjusted and unadjusted rates reveal
that adjustment serves to alter mortality patterns by substantial margins and in
different directions in some cases. For example, whites of the Cardinal HSA and non-
whites of the Southern Piedmont have below-average unadjusted rates but above-average
adjusted rates. Hence, age-adjustment is crucial to the analysis of mortality risk.

According to unpublished results of the Fall 1979 North Carolina Citizen Survey,
females and nonwhites utilize health departments more than males and whites. Thus,
present findings for nonwhite females should be of particular interest to public
health administrators. These findings include the following:

- Between 1973 and 1977, N.C. nonwhite females experienced
substantial reductions in age-adjusted mortality, leading
other N.C. race-sex groups and their U.S. counterparts in
reducing total mor*-Tlity by 20%. By cause, the state's
nonwhite females experienced higher reductions than others
in heart disease, stroke, cancer, chronic obstructive lung
disease, cirrhosis and motor vehicle accidents.

- At the same time, N.C.'s 1978 nonwhite female age-adjusted
rate remained 50% above the white female rate with excesses
in most major causes. Recent trends reveal substantial
increases in lung cancer and suicide with smaller increases
in pancreatic cancer and nephr i t i s/nephros i s. And with
higher rates, N.C.'s nonwhite females are lagging behind
their U.S. counterparts in red :ing death from hypertension,
arteriosclerosis, diabetes and non-motor-vehicle accidents.

Except for a static nonwhite male rate in the Southern Piedmont, Table 2 reveals
that all HSA's have shared in the recent mortality declines of all four race-sex
groups, especially nonwhite females, and these declines are reducing the gaps between
whites and nonwhites of both sexes. In contrast, greater downturns in female mortality
have widened the gaps between N.C. males and females of both races. These trends
are observed in the race and sex ratios of Table k where it is also shown that race
differences are greater for females, sex differences are greater for whites and sex
differences are greater than race differences. Based on the rates of Table 2, this
is generally true in all HSA's. The 1978 race and sex ratios for HSA's also reveal that

- race differences are greater, in the three westernmost HSA's;

- sex r 'fferences are highest in the Capital HSA.

The decline in mortality from stroke and heart disease in North Carolina reflects
a nationwide trend. The reasons for this decline are unclear; a number of primary
and secondary prevention factors have been cited including improved coronary care
techniques, changes in diet and increased exercise. While there is still debate on
the role factors such as these may have played, there is little disagreement that the
improved detection and treatment of hypertension and reduced cigarette smoking have
played important roles. In spite of these declines, cardiovascular diseases remain
the leading causes of death in both North Carolina and the nation. Obviously, there
remains a great deal of preventive medicine work to be done.

Table k

Race and Sex Ratios:
Age-adjusted Mortality Rates

North Carolina, 1973 and 1978
Ratio of Nonwhite Ratio of Male to

to Wh

te Rates

Fema

e Rates

Year

Males

Females

Whites

Nonwhi tes

1973

l.'<3

1.65

1.98

1.71

1978

1.1.1

1.51

2.03

1.90

Some other comments are that (l) unadjusted rates are certainly valid and prefer-
able indicators for allocating health manpower, facilities, supplies, etc., but in
assessing mortality "risk," adjustment for confounding factors such as age is the only
way to go; (2) all mortality rates are subject to spatial differences or temporal
changes in certification practice and/or accuracy of diagnoses and (3) as always, the
accuracy of rates examined here is also contingent upon the accuracy of population
bases.

In summary, it has been demonstrated empirically that both race and sex,
especially sex, are differentiating factors in mortality risk, even more so in North
Carolina than nationwide, and North Carolina's sex differential is widening. The age-
adjusted data also underscore the need for expanded initiatives in the area of accidents,
both motor vehicle and other types.

REFERENCES

(1) North Carolina Department of Human Resources, Division of Health Services,

Administrative Services Section, Public Health Statistics Branch. Leading
Causes of Mortality, North Carolina Vital Statistics 1978, Volume 2.
Raleigh, October 1979-

(2) U.S. Department of Health, Education, and Welfare, Public Health Service. Monthly

Vital Statistics Report: Advance Report Final Mortality Statistics, 1977,
from the National Center for Health Statistics. Vol. 28, No. 1. Hyattsvi lie,
Maryland, May 11, 1979-

Rates for this study were produced by adjustment programs developed
in the State Center for Health Statistics. Available to other users,
these programs use the direct method to adjust for all or any combi-
nation of age, race and sex.

STATE LIBRARY OF NORTH CAROLINA

\m

Hill

nil

ill 1

liii

11111

1111

III

3 3091 00753 7376

State Center for Health Statistics
Division of Health Services
Department of Human Resources
P.O. Box 2091
Raleigh, North Carolina 27602

Bulk Rate
U.S. Postage
PAID
RALEIGH, N.C. 27602
PERMIT NO. l}^'*

Dociments Dspartment

Librarian

NT.C. State Library