using summary measures of mortality for community planning and policy development

Using Summary Measures of Mortality for Community Planning

and Policy Development

Bruce Cohen, Ph.D.Director, Division of Research, Bureau of Health

Statistics, MDPHNAPHSIS Annual Meeting

June 2008

Context Many public health practitioners feel that mortality

data are not very useful: death is too late of end-point to use for policy, targeting interventions, and evaluation of health care delivery

There are summary, non disease-specific measures that have been developed to enhance the utility of mortality data to identify potential system changes

Two such measures are: premature mortality (PMR) mortality Amenable to Health Care (AM)

As an additional issue, briefly present data on the interaction of race and income—this is an important focus for use of vitals data for community needs assessment and planning

Premature Mortality Rate (PMR)

PMR: Background

Almost 2 out of 3 deaths in Massachusetts occur to people age 75 and older

Although quality of life for our older citizens is important, we wanted to use a measure that focused on the health of younger persons

Why? The rationale is that the vast majority of deaths to persons age 75 and older are due to chronic conditions associated with aging

By examining deaths to persons younger than 75, it is possible to identify many issues that are more amenable to systematic public health approaches to health promotion and disease prevention

PMR: Background

THE PMR is considered an excellent, single measure that reflects the health status of a population, and the need for systematic public health approaches to health promotion and disease prevention.1,2

Sometimes used as an indicator of health care need

1Eyles J, Birch S. A population needs-based approach to health care resource allocation and planning in Ontario: A link between policy goals and practice. Can J Public Health 1993; 84(2): 112-117.2 Carstairs V., Morris R. Deprivation and Health in Scotland. Aberdeen Scotland: Aberdeen University Pres, 1991

PMR: Attractive Properties...

Data used to calculate the PMR are readily available (mortality and age of population);

PMR is easily understandable and intuitive;

PMR provides a mechanism to summarize the burden of multiple adverse conditions creating a broader community perspective.

PMR Definition

The number of deaths to persons age 0-74 divided by the population age 0-74 (per 100,000)

Age adjusted to the 2000 US standard population, age 0-74

PMR: related to many factors

Health care is certainly one of these factors, but not the only factor

PMR may be related to socioeconomic status and its correlates: potential issues such as environmental conditions, housing, education, stress, higher rates of smoking, substance abuse, violence, obesity, and lack of access to care

Other possible reasons for high PMRs: specific sub-populations of younger persons at risk such as: HIV/AIDS; increased motor vehicle deaths in rural areas; heart attack deaths in persons 45-64 in suburbia; violence

Median Household Income and PMR by EOHHS Regions, Massachusetts: 2005

0

50

100

150

200

250

300

350

400

BostonRegion

Western Central Southeast Northeast Metro West

PM

R

$0

$10,000

$20,000

$30,000

$40,000

$50,000

$60,000

$70,000

$80,000

$90,000

$100,000

Med

ian

Ho

use

ho

ld I

nco

me

($)

Age adjusted Rate per 100,000 population

Median Household Income

Source: Income information from the 2000 Census.

Less than High School Education and PMR by EOHHS Regions, Massachusetts: 2005

0

50

100

150

200

250

300

350

400

BostonRegion

Western Central Southeast Northeast Metro West

PM

R

0%

2%

4%

6%

8%

10%

12%

14%

%

Les

s th

an a

Hig

h S

cho

ol

Ed

uca

tio

n

Age adjusted Rate per 100,000 population

Less than High School Education

Source: Education information from the 2000 Census.

Premature Mortality Rates by Race and Hispanic Ethnicity Massachusetts: 2006

300.5279.8

140.3

427.3

298.8

0

200

400

600

White non-Hispanic

Black non-Hispanic

Asian non-Hispanic

Hispanic Total

Dea

ths

per

100

,000

Rates are per 100,000 population. Age-adjusted to the 2000 US standard population persons ages 0-74

* Statistically higher than state rate (p<0.05)** Statistically lower than state rate (p<0.05)

*

**

Rates per 100,000 population under 75 years of age; age-adjusted to the 2000 US standard population.

CHNA PMR1. Community Health Network of Berkshire = 330.82. Upper Valley Health Web-Franklin County = 334.23. Partnership for Health in Hampshire County = 285.74. The Community Health Connection = 370.75. Community Health Network of Southern Worcester County = 354.96. Community Partners for Health = 277.17. Community Health Network of Greater Metro West = 2508. Community Wellness Coalition = 3399. Fitchburg/Gardner Community Health Network = 324.510. Greater Lowell Community Health Network = 333.411. Greater Lawrence Community Health Network = 285.612. Greater Haverhill Community Health Network = 328.113. Community Health Network North = 248.8

14. North Shore Community Health Network = 303.515. Greater Woburn/Concord/Littleton Community Health Network = 217.816. North Suburban Health Alliance = 294.517. Greater Cambridge/Somerville Community Health Network = 255.718. West Suburban Health Network = 208.219. Alliance for Community Health = 365.420. Blue Hills Community Health Alliance = 297.321. Hampshire County Partnership = 367.522. Greater Brockton Community Health Network = 349.223. South Shore Community Partners in Prevention = 31324. Greater Attleboro-Taunton Health & Education Response = 347.125. Partners for a Healthier Community = 378.426. Greater New Bedford Health & Human Services Coalition = 35127. Cape Cod & Islands Community Health Network = 295.1

45

8

24

26

22

21

25

19

1

2 9

3

27

23

20

12

1011

1416

7

6

15

13

18

17

Rates are per 100,000 persons under 75 years of age, age-adjusted to the 2000 US standard population and are calculated using MDPH population estimates for 2005, which are the most updated estimates available at the sub-state level by age groups.

PMR by CHNA

Significantly higher than state rate

Not different from state rate

Significantly lower than state rate

Premature Mortality Rates (PMR) by Community Health Network Area Massachusetts: 2006

Premature Mortality Rates (PMR) to Chronic Diseasesby Community Health Network Area (CHNA), Massachusetts: 2006


CHNA PMR1. Community Health Network of Berkshire = 220.22. Upper Valley Health Web-Franklin County = 2183. Partnership for Health in Hampshire County = 184.44. The Community Health Connection = 225.95. Community Health Network of Southern Worcester County = 242.16. Community Partners for Health = 187.67. Community Health Network of Greater Metro West = 170.78. Community Wellness Coalition = 216.69. Fitchburg/Gardner Community Health Network = 213.410. Greater Lowell Community Health Network = 220.111. Greater Lawrence Community Health Network = 192.512. Greater Haverhill Community Health Network = 226.613. Community Health Network North = 174.5

14. North Shore Community Health Network = 188.815. Greater Woburn/Concord/Littleton Community Health Network = 151.816. North Suburban Health Alliance = 197.417. Greater Cambridge/Somerville Community Health Network = 16618. West Suburban Health Network = 136.919. Alliance for Community Health = 223.520. Blue Hills Community Health Alliance = 198.921. Hampshire County Partnership = 240.522. Greater Brockton Community Health Network = 228.923. South Shore Community Partners in Prevention = 216.924. Greater Attleboro-Taunton Health & Education Response = 235.325. Partners for a Healthier Community = 240.226. Greater New Bedford Health & Human Services Coalition = 219.627. Cape Cod & Islands Community Health Network = 192.2

45

24

22

21

25

19

1

2 9

3

6

8

27

23

26

20

12

1011

1416

7

15

13

18

17


PMR to Chronic Diseases by CHNA




Massachusetts PMR to Chronic Diseases= 203.2

Premature Mortality Rates (PMR) to Non-Chronic Diseasesby Community Health Network Area (CHNA), Massachusetts: 2006


CHNA PMR1. Community Health Network of Berkshire = 110.62. Upper Valley Health Web-Franklin County = 116.23. Partnership for Health in Hampshire County = 101.24. The Community Health Connection = 144.85. Community Health Network of Southern Worcester County = 112.86. Community Partners for Health = 89.57. Community Health Network of Greater Metro West = 79.38. Community Wellness Coalition = 122.59. Fitchburg/Gardner Community Health Network = 111.110. Greater Lowell Community Health Network = 113.311. Greater Lawrence Community Health Network = 93.112. Greater Haverhill Community Health Network = 101.613. Community Health Network North = 74.3

14. North Shore Community Health Network = 114.615. Greater Woburn/Concord/Littleton Community Health Network = 66.016. North Suburban Health Alliance = 97.117. Greater Cambridge/Somerville Community Health Network = 89.618. West Suburban Health Network = 71.419. Alliance for Community Health = 141.920. Blue Hills Community Health Alliance = 98.421. Hampshire County Partnership = 127.022. Greater Brockton Community Health Network = 120.323. South Shore Community Partners in Prevention = 96.124. Greater Attleboro-Taunton Health & Education Response = 111.825. Partners for a Healthier Community = 138.226. Greater New Bedford Health & Human Services Coalition = 131.427. Cape Cod & Islands Community Health Network = 102.9

4

2625

19

1

2 9

3

56

8

27

2423

20

12

10

22

21

11

1416

7

15

13

18

17


PMR to Non-Chronic Diseases by CHNA




Massachusetts PMR to non Chronic Diseases= 107.9

Premature Mortality Rate by Race/EthnicityChronic Diseases1, Massachusetts: 2006

197.8 195.0

97.9*

146.8*

245.0*

0

50

100

150

200

250

300

White non-Hispanic

Black non-Hispanic

Hispanic Asian non-Hispanic

ALL

Dea

ths

per

100

,000

Age-adjusted to the 2000 US standard population under 75 years of age.

(*) Statistically different from State (p ≤.05)

1 Includes Cancer, heart disease, stroke, CLRD, nephritis, chronic liver disease, diabetes, Parkinson, and other chronic diseases

Premature Mortality Rate by Race/EthnicityNon Chronic Conditions/Diseases,

Massachusetts: 2006

101.0 105.5

42.4*

133.0*

182.3*

0

20

40

60

80

100

120

140

160

180

200

White non-Hispanic

Black non-Hispanic

Hispanic Asian non-Hispanic

ALL

Dea

ths

per

100

,000

Age-adjusted to the 2000 US standard population under 75 years of age.

(*) Statistically different from State (p ≤.05)

PMR: Limitations

PMR does not identify specific reasons why some areas may be high or low

summary measures may sometimes obscure important subgroup differences

mortality might not be a good measure of important public health issues (e.g. arthritis, poor housing, etc.)

PMR: summary

The PMR is a useful tool…

to begin discussions that allow policy makers, community advocates, public health professionals, and cities and towns to consider more effective and cost efficient approaches to improving the quality of life and health of the public;

to focus on the inter-connected roots of early death and direct us towards considering the overall health of our communities.

Mortality Amenable to Health Care

Amenable Mortality: Background

Definition: deaths from certain causes that should not occur in the presence of timely and effective health care.1,2

Originally developed in US in 1970’s; adopted and updated by many researchers especially in Europe.2

This concept has been revitalized and used to assess the quality of health care systems

Potentially useful tool to assess performance of health care systems and track changes over time.1

1Nolte E and McKee CM. Measuring The Health of Nations: Updating An Earlier Analysis. Health Affairs 2008; Vol 27, Number 1: 58-71; Jan/Feb 2008.

2Nolte E and McKee CM . Does Health Care Save Lives? Avoidable Mortality Revisited. The Nullfield Trust. 2004. London, England


Causes amenable to secondary prevention through early detection and treatment: this includes causes where screening and treatment are effective; for example breast, cervical, and skin cancer

Causes amenable to improved treatment and medical care: this group includes infectious diseases; causes that respond to antibiotic treatments and immunizations as well as causes that require direct medical and/or surgical intervention such as appendicitis and hypertension or causes that rely on efficient medical care delivery (accurate and timely diagnosis, transport, and treatment.)

(Adapted from Does Health Care Save Lives? p.30)


Operationalized as a set of 33 cause of death codes for persons under age 751

Subset of PMR

1 Online data supplement to Nolte and McKee, Measuring the Health Of Nations. Health Affairs. Vol. 27,

no. 1. (http://content.healthaffairs.org/cgi/content/full/27/1/58/DC1)

List of Causes of Death Considered Amenable to Health care

Intestinal infections Tuberculosis Other infectious (Diphtheria, Tetanus,

Poliomyelitis) Whooping cough Septicemia Measles Malignant neoplasm of colon and rectum Malignant neoplasm of skin, Malignant neoplasm of breast, Malignant neoplasm of cervix uteri Malignant neoplasm of cervix uteri and

body of the uterus Malignant neoplasm of testis

Hodgkin’s disease Leukemia Diseases of the thyroid Diabetes mellitus Epilepsy Chronic rheumatic heart disease Hypertensive disease Ischemic heart disease Cerebrovascular disease All respiratory diseases (excl.

pneumonia/influenza) Influenza

List of Causes of Death Considered Amenable to Health care (continued)

Pneumonia Peptic ulcer Appendicitis Abdominal hernia Cholelithiasis & cholecystitis Nephritis and nephrosis Benign prostatic hyperplasia Maternal deaths Congenital cardiovascular anomalies Perinatal deaths, all causes excluding

stillbirths Misadventures to patients during surgical and

medical care

List of Causes of Death Considered Amenable to Health care (continued)

Reasons Considered Amenable

Percent Amenable Deaths Massachusetts: 2006

All Deaths

90%

Amenable

Deaths10%

Amenable

Deaths28%

All Deaths Ages 0-

7472%

All Deaths Deaths Persons Ages 0-74

Mortality Rates for Causes Amenable to Health Care by Race and Ethnicity

Massachusetts: 2000 and 2006

105.4

142.9

40.6

103.6

64.5

84.081.4

48.1

128.4

82.5

0

40

80

120

160

White non-Hispanic

Black non-Hispanic

Asian non-Hispanic

Hispanic Total

Dea

ths

per

100

,000

2000 2006


**

** Statistically lower than 2000 rate (p<0.05)

**

133.7

86.0

103.6

84.0

67.6

101.8

0

40

80

120

160

Male Female Total

Dea

ths

per

100

,000

2000 2006


** Statistically lower than 2000 rate (p<0.05)

**

****

Mortality Rates for Causes Amenable to Health Care by Gender

Massachusetts: 2000 and 2006

Premature Mortality Rates & Amenable Mortality

by Race and Ethnicity, Massachusetts: 2006

29%34%30%

28%28%

427.3

298.8300.5

140.3

279.8

0

100

200

300

400

500

600

Total White non-Hispanic

Black non-Hispanic

Asian non-Hispanic

Hispanic

Dea

ths

per

100

,000


Uses of Amenable Mortality

Amenable mortality is a useful tool…

to begin discussions that allow policy makers, community advocates, and public health professionals, to consider more effective and cost efficient approaches to improving the quality of life and health of the public;

to move us away from considering only individual diseases, and directs us towards considering the overall health and access issues.

The Interaction between race and poverty: examples from natality analyses

No direct measure of income on the birth certificate

Education is useful, but teens haven’t completed schooling and foreign born have different educational experiences

Is it race? (surrogate for unequal treatment, cultural differences, linguistic isolation, etc.) OR

Is it poverty? (lack of financial access to purchase medical care, other necessities, surrogate for other detrimental exposures such as higher pollution, crime, stress, etc.)

Infant Mortality Rate by Race and Education Mothers Ages 25+, Massachusetts – 2000-2006

6.1

3.7 3.6

7.2

4.9

2.5

11.9*

9.0*

16.0*

0

2

4

6

8

10

12

14

16

18

High School or Less Some College College Graduate or More

Infa

nt D

eath

s p

er 1

,000

Liv

e B

irth

s

White, Non-Hispanic

Black, Non-Hispanic

Hispanic

* Significantly Different from White Non-Hispanic

Infant Mortality Rate by Percent in Poverty and Race-Hispanic Ethnicity

Bla

ck N

H

Bla

ck N

H

Bla

ck N

H

Bla

ck N

H

His

pan

ics

His

pan

ics

His

pan

ics

His

pan

ics

Asi

an N

H

Asi

an N

H

Asi

an N

H

Asi

an N

H

Wh

ite

NH

Wh

ite

NH

Wh

ite

NH

Wh

ite

NH

0

5

10

15

0- 4.9% 5- 9.9% 10-19.9% 20-100 %

Percent in Poverty

IMR

(p

er 1

000

live

bir

ths)

Infant Mortality Rate by Race-Hispanic Ethnicity and Percent in Poverty

0-

4.9

%

0-

4.9

%

0-

4.9

%

5-

9.9

%

5-

9.9

%

5-

9.9

%

10-1

9.9

%

10-1

9.9

%

10-1

9.9

%

20-1

00 %

20-1

00 %

20-1

00 %

0-

4.9

%

5-

9.9

%

10-1

9.9

%

20-1

00 %

0

5

10

15

White NH Black NH Hispanics Asian NH

Race-Hispanic Ethinicty

IMR

(p

er

1000 l

ive b

irth

s)

LBW by Percent in Poverty and Race-Hispanic Ethnicity

Bla

ck N

H

Bla

ck N

H

Bla

ck N

H

Bla

ck N

H

His

pan

ics

His

pan

ics

His

pan

ics

His

pan

ics

Asi

an N

H

Asi

an N

H

Asi

an N

H

Asi

an N

H

Wh

ite

NH

Wh

ite

NH

Wh

ite

NH

Wh

ite

NH

0

5

10

15

0- 4.9% 5- 9.9% 10-19.9% 20-100 %

Percent in Poverty

LB

W (

%)

Smoking During Pregnancy by Percent in Poverty and Race-Hispanic Ethnicity

Bla

ck N

H

Bla

ck N

H

Bla

ck N

H

Bla

ck N

H

His

pan

ics

His

pan

ics

His

pan

ics

His

pan

ics

Asi

an N

H

Asi

an N

H

Asi

an N

H

Asi

an N

H

Wh

ite

NH

Wh

ite

NH

Wh

ite

NH

Wh

ite

NH

0

5

10

15

20

25

0- 4.9% 5- 9.9% 10-19.9% 20-100 %

Percent in Poverty

Sm

oki

ng

Du

rin

g P

reg

nan

cy(%

)

Smoking During Pregnancy by Race-Hispanic Ethnicity and Percent in Poverty

0-

4.9

%

0-

4.9

%

0-

4.9

%

5-

9.9

%

5-

9.9

%

5-

9.9

%

10-1

9.9

%

10-1

9.9

%

10-1

9.9

%

20-1

00 %

20-1

00 %

20-1

00 %

0-

4.9

%

5-

9.9

%

10-1

9.9

%

20-1

00 %

0

5

10

15

20

25

White NH Black NH Hispanics Asian NH

Race-Hispanic Ethinicty

Sm

okin

g D

uri

ng

Pre

gn

an

cy(%

)

Concluding thoughts

We should be as creative as possible making our statistics and analyses vital for public health policy development and community uses

There are emerging frameworks that allow for use of vital statistics in these ways—we should be standardizing and promoting these applications

using summary measures of mortality for community planning and policy development

Documents