Socio-Economic Differences in Mortality by Cause of Deathandrewc/ARCresources/Cairns... · 11...

,

Socio-Economic Differences in Mortality by Causeof Death

Andrew J.G. Cairns

Heriot-Watt University, Edinburgh

Director, Actuarial Research Centre, IFoA

Joint work with D. Blake, K. Dowd, M. Kallestrup-Lamb, C. Rosenskjold, C. Redondo, A.S.Macdonald

Longevity 13, Taipei, September 2017

Andrew J.G. Cairns Socio-Economic Differences in Mortality by Cause of Death 1 / 38

,

The Actuarial Research Centre (ARC)A gateway to global actuarial research

The Actuarial Research Centre (ARC) is the Institute and Faculty ofActuaries’ (IFoA) network of actuarial researchers around the world.The ARC seeks to deliver cutting-edge research programmes that addresssome of the significant, global challenges in actuarial science, through apartnership of the actuarial profession, the academic community andpractitioners.The ’Modelling, Measurement and Management of Longevity andMorbidity Risk’ research programme is being funded by the ARC, theSoA and the CIA.

www.actuaries.org.uk/arc


www.actuaries.org.uk/arc

,

Outline

Danish data:

affluenceeducationcause of death

Statistical significance

US cause of death data by education group


,

Purpose of looking at cause of death data

What are the key drivers of all-cause mortality?

How are the key drivers changing over time?Which causes of death have high levels of inequality:

by education;by affluence?

Beware of

changes in ICD classification of deathsdrift in how deaths are classifiedchanging education levels (grade inflation)

Insight into mortality underpinning life insurance andpensions


,

Danish Data: Cairns et al. Longevity 11, Lyon

Statistics Denmark National Register Database

Key data (amongst others) for each individual:

Date of birth (⇒ age)Date of deathWealthIncomeAffluence=Wealth+15×IncomeEducationCause of Death


,

Education and Affluence Levels

EducationLow education Primary and lower secondary educationMedium education Upper secondary educationHigh education Tertiary education

AffluenceLevel 1 Low affluence decile

......

Level 10 High affluence decile


,

Model-Inferred Underlying Death Rates 2012

60 70 80 90

0.00

20.

010

0.05

00.

200

Group 1Group 2Group 3Group 4Group 5Group 6Group 7Group 8Group 9Group 10

Males Crude m(t,x); 2012

Age

m(t

,x)

(log

scal

e)

60 70 80 90

0.00

20.

010

0.05

00.

200

Group 1Group 2Group 3Group 4Group 5Group 6Group 7Group 8Group 9Group 10

Males CBD−X Fitted m(t,x); 2012Point Estimates

Age

m(t

,x)

(log

scal

e)


,

Education as an Alternative Covariate

Level of Educational Attainment also known to be a goodpredictor

Various US studies

Mackenbach et al. (2003) including Denmark: Std. Mortality RatesBrønnum-Hansen and Baadsgaard (2012) Denmark: LE(x = 30)

As close as possible on a like for like basis:

Crude death rates; age 30+; matching years.

Affluence ⇒Wider spread of SMR’s than M. et al. (2003)Wider spread of LE (30) than BHB (2012)

Issue: “grade inflation” distorts results

More to be done.


,


1985 1990 1995 2000 2005 2010

05

1015

20

Year

Age

Sta

ndar

dise

d M

orta

lity

Rat

e (

per

1000

per

son

year

s)

Affluence Group 1Low EducationHigh EducationAffluence Group 10

Age Standardised Mortality Rates per 1000Ages 45−54; European Standard Population (1976)


,

Cause of Death Data – Health Inequalities

Deaths subdivided into 29 CoD groups

Age groups 31-35, 36-40, ..., 91-95

Year groups 1985-89, 1990-94, 1995-99,2000-2004, 2005-2009

Compare affluence groups

Compare education groups


,

Cause of Death Data – Health Inequalities

1 Infectious diseases incl. tuberculosis 2 Cancer: mouth, gullet, stomach3 Cancer: gut, rectum 4 Cancer: lung, larynx, ..5 Cancer: breast 6 Cancer: uterus, cervix7 Cancer: prostate, testicular 8 Cancer: bones, skin9 Cancer: lymphatic, blood-forming tissue 10 Benign tumours11 Diseases: blood 12 Diabetes13 Mental illness 14 Meningitis + nervous system (Alzh.)15 Blood pressure + rheumatic fever 16 Ischaemic heart diseases17 Other heart diseases 18 Diseases: cerebrovascular19 Diseases: circulatory 20 Diseases: lungs, breathing21 Diseases: digestive 22 Diseases: urine, kidney,...23 Diseases: skin, bone, tissue 24 Senility without mental illness25 Road/other accidents 26 Other causes27 Alcohol → liver disease 28 Suicide29 Accidental Poisonings


,

Denmark: Cause of Death Data 2007

Compare education with affluence as covariates:

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationIschaemic heart diseases

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 901e

−05

1e−

031e

−01

Danish Males by AffluenceIschaemic heart diseases

Age Group

Dea

th R

ate

(log

scal

e)

12345678910

Affluence ⇒ wider spread


,


40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationDiabetes

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by AffluenceDiabetes

Age GroupD

eath

Rat

e (lo

g sc

ale)

12345678910

Affluence ⇒ much wider


,


40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationCancer: lung, larynx, ..

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by AffluenceCancer: lung, larynx, ..

Age GroupD

eath

Rat

e (lo

g sc

ale)

12345678910


,


40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationAlcohol −> liver

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by AffluenceAlcohol −> liver

Age GroupD

eath

Rat

e (lo

g sc

ale)

12345678910


,


40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationAccidental Poisonings

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by AffluenceAccidental Poisonings

Age GroupD

eath

Rat

e (lo

g sc

ale)

12345678910

Low affluence ⇒ over 20× at young ages


,


Many causes of death have known risk factors ordriverse.g. smoking, diet, healthy lifestyle etc.⇒ clear socio-economic differences

Biggest differences at ages < 60

Affluence ⇒ stronger predictor than education(sometimes very much stronger)

Other diseases do not have strong differences:


,


40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationCancer: prostate, testicular

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by AffluenceCancer: prostate, testicular

Age GroupD

eath

Rat

e (lo

g sc

ale)

12345678910


,


40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationCancer: lymphatic, blood−forming tissue

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by AffluenceCancer: lymphatic, blood−forming tissue

Age GroupD

eath

Rat

e (lo

g sc

ale)

12345678910

Education ⇒ no effectAffluence ⇒ small effect


,


40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by EducationCancer: gut, rectum

Age Group

Dea

th R

ate

(log

scal

e)

Low EduMediumHigh

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Danish Males by AffluenceCancer: gut, rectum

Age GroupD

eath

Rat

e (lo

g sc

ale)

12345678910


,

Denmark: Cause of Death Data – Health Inequalities

Some causes of death have no obvious link tolifestyle/affluence/educatione.g. Prostate CancerCancerUK:Prostate cancer is not clearly linked to anypreventable risk factors.

But Affluence ⇒ inequalities

Possible explanations (a very non-expert view)

onset is not dependent on lifestyle/affluence/educationBUT less affluent/educated ⇒

??? later diagnosis??? engage less well with treatment process??? lower quality housing


,

CoD Death Rates: Different Shapes & Patterns

40 50 60 70 80 90

1e−

051e

−04

1e−

031e

−02

1e−

01Infectious diseases incl. tuberculosis

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−04

1e−

031e

−02

1e−

01

Meningitis + nervous system (Alzh.)

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−04

1e−

031e

−02

1e−

01

Ischaemic heart diseases

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−04

1e−

031e

−02

1e−

01

Diseases: circulatory

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−04

1e−

031e

−02

1e−

01Diseases: lungs, breathing

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−04

1e−

031e

−02

1e−

01

Diseases: urine, kidney,...

Dea

th R

ate

(log

scal

e)

12345678910


,

CoD Death Rates: Different Shapes & Patterns

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Cancer: gut, rectum

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Cancer: lung, larynx, ..

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Cancer: prostate, testicular

Dea

th R

ate

(log

scal

e)

12345678910

40 50 60 70 80 90

1e−

051e

−03

1e−

01

Cancer: bones, skin

Dea

th R

ate

(log

scal

e)

12345678910


,

Shapes: Conclusions

Typically:Non-cancerous diseases ⇒ approximately exponentialgrowthNeoplasms (cancers) ⇒ subexponential ???polynomial

What does this reveal about different diseasemechanisms?


,

Denmark Males: Statistical Significance

Which CoD’s are significantly affected bysocio-economic status?

H0: Affluence groups all have the same CoDdeath rate mi(c , t, x) = mj(c , t, x) ∀i 6= jversus

H1: Affluence groups do not all have the sameCoD death rates


,

Denmark Males: Statistical Significance

For each cause of death (29), and age group (13)

Rank the death rates for the 10 groups i = 1, . . . , 10

For each year group, tR(i , t) = rank of m(i , t) out of m(1, t), . . . ,m(10, t)Rank 1: highest death rateRank 10: lowest death rate

Data (i ,R(i , t))

Test statistic, S = cor(i ,R(i , t)

)Under H0 the ranks are a random permutation of 1, . . . , 10

Under H0, S is approximately N(0, σ2) where σ = 0.149.

One-sided test: Reject H0 if S > σΦ−1(α)

Large S ⇒ low affluence ∼ high CoD mortality


,

Cause of Death Inequalities: p-values

p−values for Danish Male Affluence Group Rankings

CoD Group

Age

Ban

d

2

4

6

8

10

12

5 10 15 20 25

0

0.0010.0050.010.05

1

Very low or zero mortality: CoD 5, 6, 24 & low agesHigh age convergence


,

US Education Data

Males and Females (2)

Single ages 55-75 (21)

Single years 1989-2015 (27)

Causes of death (29)

Low, medium & high education level (3)


,

US Education Data: Growing Inequality

US Males All Cause MortalityRatio of Low to High Education Mortality

Year 1989−2015

Age

555657585960616263646566676869707172737475

1990 1994 1998 2002 2006 2010 2014

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5


,

US Education Data

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−

01 Low EduMediumHigh

Year 2000 Cancer: prostate, testicular

Age

CoD

Dea

th R

ate

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2015 Cancer: prostate, testicular

Age

CoD

Dea

th R

ate

Recall: Denmark ⇒ very narrow gap


,

US Education Data

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2000 Ischaemic heart diseases

Age

CoD

Dea

th R

ate

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2015 Ischaemic heart diseases

Age

CoD

Dea

th R

ate

Widening gap


,

US Education Data

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2000 Meningitis + nervous system (Alzh.)

Age

CoD

Dea

th R

ate

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2015 Meningitis + nervous system (Alzh.)

Age

CoD

Dea

th R

ate

Widening gap


,

US Education Data

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2000 Accidental Poisonings

Age

CoD

Dea

th R

ate

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2015 Accidental Poisonings

Age

CoD

Dea

th R

ate

Case & Deaton (2015) ⇒ Accidental poisoning ↗


,

US Education Data

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2000 Alcohol −> liver

Age

CoD

Dea

th R

ate

55 60 65 70 75

1e−

051e

−04

1e−

031e

−02

1e−


Year 2015 Alcohol −> liver

Age

CoD

Dea

th R

ate

Widening gap


,

US Males: Low versus High Education

Do Low and High education groups have the sameCoD rate?

Four × 5-year age groups

29 causes of death

Signs Test (count low edu. > high edu. mort.)

29× 4 = 116 individual tests

115/116 ⇒ reject H0 equality

Accept H0 (p = 0.08) for only one pairing(Meningitis + nervous system (Alzh.), 70-74)

Most p-values < 10−6


,

4. Summary

Affluence better than education for all CoD ifyou have the data

Impact of affluence/education varies with CoD

Different growth patterns cancers versus otherdiseases

Work in progress!

E: [email protected] W: www.macs.hw.ac.uk/∼andrewc


,

Thank You!

Questions?

E: [email protected] W: www.macs.hw.ac.uk/∼andrewc


,


Dig a bit deeper:Affluence + Education: average ASMR’s over 5 years

Mortality Improvement Rates (%)Period 1987−2009; Age Band 45−54By Affluence and Education Group

Affluence Group

Edu

catio

n Le

vel

LOW

MED

HIGH

1 2 3 4 5 6 7 8 9 10

−1012345


Socio-Economic Differences in Mortality by Cause of Deathandrewc/ARCresources/Cairns... · 11...

Documents

Transcript of Socio-Economic Differences in Mortality by Cause of Deathandrewc/ARCresources/Cairns... · 11...