Inequities in Cancer Care

Bill MackillopQueen’s Cancer Research Institute

What is meant by “inequities” in the context of cancer care?

Differences in access to care or quality of care, among groups of people who have equal needs, and who have equal entitlement to care

Why do inequities matter?

• Equity is valuable in its own right• Equity in cancer care is necessary to

achieve optimal outcomes in the population at large– Inequities represent opportunities for improving

overall outcomes

How may inequities in care lead to inequities in outcomes?

• Differing access to primary care and diagnostic services– unequal stage at presentation– unequal general health

• Differing access to cancer treatment– unequal delays in starting treatment– unequal options for treatment

• Differing quality of cancer treatment– unequal knowledge and skill among providers

Inequities among whom?

• different countries• different regions of the same country• different subgroups of residents of the same

community• e.g. different age groups, ses groups, racial groups,

disease groups

International Variations in Age-Standardized 5-Year Relative Survival Rates (%) for Selected Cancers (1985-1989)

36

21

5

5673479

Europe (Eurocare)

14

23

6

3870406

China (Shanghai)

-365352603981Prostate

27395143534260Colon

3571212713Lung

-

13

4

50

China (Qidong)

365245Pancreas

21

44

84

Japan (Osaka)

17

8

43

India (Madras)

17

22

71

Singapore

413144Leukemia

251321Stomach

806782Breast

SwitzerlandUKUS (SEER)

Cancer

Chia et al, Int. J. Cancer 2000; Sankaranarayanan et al, International Agency for Research on Cancer, 1998

Potential causes of differences in outcomes among rich countries

• Differences in degree of income inequality• Differences in % GDP spent on health care • Differences in health systems

– Variations in universality, accessibility and/or comprehensiveness of health insurance

Fem ales

US

France

Germ any

Netherland

Switzerland

Italy

Finland

Denm ark

England

Spain

0

2

4

6

8

10

12

0 2 4 6 8 10 12Rank of Incom e Inequality

Rank of Relative Survival Rate

R=0.03

Relationship between Income Inequality and Relative Survival for all cancer in females, (1989)

g

Data from Evans et al, Public Health 2000; Lynch et al, Lancet 2001

11.0

9.28.4 8.3

7.7 7.2 7.26.4 6.0 6.0

0

4

8

12

U SF r

ance

G erm

any

N eth

erlan

d sSw

itzer

land

F in l

and

Italy

D enm

ark U K

Spain

C o u n try

%GDP

Variations in %GDP Spent on Health Care among Wealthy Countries, 1989-90

Relationship between % of GDP spent on health care, and Relative Survival for all cancers in females (1989)

Evans et al, Public Health 2000

Fem alesSpain

England

Denm ark

Finland

Italy

Switzerland

Netherland

Germ any

France

US/SEER

0

2

4

6

8

10

12

0 2 4 6 8 10 12

Rank of % GDP Spent on Health

Rank of Relative Survival R

ate

1

1

R=0.82

p<0.01

Potential for bias in international comparisons of cancer outcomes

• Differences in analytic methods.• Differences among registries

– Definitions: rules for assigning site, date of Dx, etc – Data quality: proportion of DCO cases, etc – Completeness of follow-up

• Differences among populations– demographic characteristics, comorbidity, stage

distribution (lead time bias)

Sant et al, Eur. J. Ca 2001; Gatta et al, Cancer, 2000, Quinn et al, Eur J Ca. 1998.

0

10

20

30

40

50

60

70

80

90

Large bowelBreast

Prostate

Stomach

Cervix

Ovary

NHLNorthern Europe Western EuropeDenmark and UK Eastern Europe

Rel

ativ

e Su

rviv

al (%

)Variations in age-standardized, 5-year relative survival

for selected cancers in Europe (1987-1989)

Sant et al, Eur. J. Ca. 2001; Gatta et al, Cancer, 2000

Conclusions from international comparisons of cancer survival

• Survival is worse in poor countries than in rich countries

• Survival is generally better in countries that devote a higher proportion of their GDP to health care.

• International differences in survival are generally greater in diseases where early diagnosis affects survival, and smaller in diseases with a more uniformly bad prognosis

The UK’s “NHS Cancer Plan”

• Background: “cancer patients in England often have poorer survival prospects than in other European countries”

• Goal: “by 2010 our 5 year survival rates will compare with the best in Europe”

• Rx: “an extra 570 million pounds a year for cancer services by 2003/4” (US $0.9bn),

What do we know about international variations in other

outcomes?

• Outcomes of palliative care– Symptom control – Quality of life

• Other outcomes of curative therapy– Organ conservation– Complication rates

Treatment of advanced laryngeal cancer Ca (T3,T4, Supraglottis)

Ontario SEER

Total laryngectomy 13.5% 53.8%Conservative treatment 75.0% 36.9%

Radiotherapy 72.5% 26.1%

Conservative surgery 2.5% 10.8%

No treatment identified11.5% 9.3%

Groome et al, 2001

Treatment

Inequities in cancer outcomes within countries

• Differences in outcomes among different subgroups of residents of the same region

• Differences in outcomes among different regions

30%

40%

50%

60%

70%

80%

90%

100%

0 1 2 3 4 5

Tim e from Diagnosis (Years)

Survival

Q 5

Q3

Q1

Head & Neck Cancer Survival in Ontario by SES(1987-92 Cohort)

P<0.001

Mackillop et al , J Clin Oncol 1997

Cause-specific Survival in Ontario and the U.S.(SEER), by Income Group (1987-92)

Boyd et al, J Clin Oncol, 1999

1Highest Quintile

2 3 4 5Lowest Quintile

Median Household Income

All sites except prostate

45

50

55

605-

year

Sur

viva

l (%

)

OntarioU.S.

Inter-regional Variations in Cancer Survival in Ontario

Zhang-Salomons, 2002

RRs from Cox regressionAdjusted RRs by removing random variations

Pancreas Colon Cervix H & N HD Testis

0.4

0.5

0.75

1

1.25

1.5

2

Primary Site

Relative Risk of Death

Potential causes of inequities in outcomes within countries

• Unequal access to care• Unequal quality of care

Potential causes of inequities in outcomes within countries

• Unequal access to care– A case study, RT in Ontario

• Unequal quality of care

UseofRadiotherapyW ithin1YearofDiagnosisofCancer

Ottawa

Kingston

Toronto

Ham iltonLondon

W indsor

Sudbury

SouthernOntarioW indsor

Sudbury

ThunderBay

#

#

#

#

#

#

#

#

Northern Ontario

(CasesdiagnosedinyearendingJune30th,1997)

QuintileN/A17.5% -24.3%25.1% -26.8%27.2% -29.1%29.2% -30.7%30.8% -35.2%

# Cancercentres

©

Counties where Cancer Centres with short waiting times are locatedCounties where Cancer Centres with long waiting times are locatedCounties where no Cancer Centreis located

0 %

10 %

20 %

30 %

40 %

50 %

0 10 20 30 40 50

Counties in Order of XRT Rate

Propo

rtio

n of

Cas

es

Proportion of cancer cases treated with radiotherapy within one year of diagnosis in Ontario

(cases diagnosed in year ending June 30th 1997)

©

RT Use in O lder Patients

0%

20%

40%

60%

80%

100%

45-54 55-64 65-74 >75

Age G roups

RT Rate or Functional State *

Adjuvant RT forresected BreastCa

Palliative RTwithin yr ofdeath fromProstate Ca

Radical RT forOropharynx Ca

Estim atedFunctionalStatus

*Relative to 45-54 years Tyldesley et al, 2000

Selected Regions

Ham ilton 1.20 (1.14,1.26)

O ttaw a 1.00

Toronto 0.88 (0.84,0.92)

Diagnosing Hospital

Affiliated with a cancer centre 1.35 (1.30,1.40)

Not affiliated w ith a cancer centre 1.00

Median household incom e, Can$

High (>50,000) 1.17 (1.11,1.24)

M edium (20,000 - 50,000) 1.09 (1.04,1.15)

Low (< 20,000) 1.00

Factors Affecting the Use of Palliative RT

Controlling for disease site, gender & age

Huang et al, DCCE 2001

Potential causes of inequities in outcomes within countries

• Unequal access to care• Unequal quality of care

– Variations in appropriateness of treatment decisions

• differences in individual knowledge among doctors• differences in decision processes

– Variations in skill in implementing treatments• The “volume effect”

Quintiles35.9% -45.9%46.6% -50.2%52.5% -56.8%58.0% -62.0%62.3% -66.7%data notavailable

#S

#S

#S

#S

#S

#S

#S

#S

Sudbury

ThunderBayOttaw a

Kingston

Toronto

Ham iltonLondon

W indsor

200 0 200 400 Miles

Northern Ontario

100 0 100 200 Miles

Southern Ontario

RateofLum pectom yUseAm ong BreastCancerPatientsin Ontario,1995-1997

Elements of a strategy for diminishing internal inequities

• A societal commitment to equity in health care is necessary, but not sufficient

• Other requirements are:– Set standards– Provide necessary resources– Monitor structure, access, quality and outcomes– Create accountability

• informed, caring, and capable governance

The role of the central registries in diminishing inequities in outcome

• Creating awareness of the problem!• Quantifying it

– Complete, accurate case identification, universal definitions, common practices, well trained personnel

• Explaining its causes– additional information may be needed

• Monitoring progress