Summary Measures of Summary Measures of Population HealthPopulation Health

Ian McDowellIan McDowellNovember, 2002November, 2002

1.1. Rationale for summary measuresRationale for summary measures

2.2. Mortality-based measuresMortality-based measures

3.3. Combined disability & mortality Combined disability & mortality methodsmethods

1. Why do we need 1. Why do we need measures of population measures of population health?health?

• We wish to monitor health of citizens …We wish to monitor health of citizens …– To set priorities for health services & policiesTo set priorities for health services & policies– To evaluate social and health policiesTo evaluate social and health policies– To compare health of different regionsTo compare health of different regions– To identify pressing health needsTo identify pressing health needs– To draw attention to inequalities in healthTo draw attention to inequalities in health– Highlight balance between length and quality Highlight balance between length and quality

of lifeof life– Numerical index desirable: a “GNP of Health” Numerical index desirable: a “GNP of Health”

Classifying Population Health Classifying Population Health Measures by their PurposeMeasures by their Purpose

1.1. DescriptiveDescriptive measures: measures:i.i. Current health status (e.g., health surveys)Current health status (e.g., health surveys)ii.ii. EvaluativeEvaluative measures (e.g., to assess measures (e.g., to assess

outcomes of health policies)outcomes of health policies)

2.2. AnalyticAnalytic measures include an implicit measures include an implicit time dimension:time dimension:iii.iii. PredictivePredictive methods (risk assessment; methods (risk assessment;

projections of disease burden) look forward;projections of disease burden) look forward;iv.iv. ExplanatoryExplanatory measures (income inequality or measures (income inequality or

social cohesion) look backwards.social cohesion) look backwards.

Descriptive(measures of

current health status)

Analytic(etiology & determinants)

Predictive(projection &

risk estimation)

Evaluative (process & outcome

measures)

Evidence-basedpolicy

Etiologicepidemiology

HealthServicesResearch

Evidence-basedmedicine

These purposes may correspond to These purposes may correspond to different types of research (shown in different types of research (shown in the ellipses)the ellipses)

Note: the figure is intended to show the typical blend of methods you might use in a particular type of study: HSR would use descriptive and analytic, for example.

Classifying Population Health Classifying Population Health Measures by their FocusMeasures by their Focus1.1. AggregateAggregate measures combine data from individual people, measures combine data from individual people,

summarized at regional or national levels. E.g., rates of smoking summarized at regional or national levels. E.g., rates of smoking or lung cancer.or lung cancer.

2.2. EnvironmentalEnvironmental indicators record physical or social indicators record physical or social characteristics of the place in which people live and cover characteristics of the place in which people live and cover factors external to the individual, such as air or water quality, or factors external to the individual, such as air or water quality, or the number of community associations that exist in a the number of community associations that exist in a neighborhood. These can have analogues at the individual level. neighborhood. These can have analogues at the individual level.

3.3. GlobalGlobal indicators have no obvious analogue at the individual indicators have no obvious analogue at the individual level. Examples include contextual indicators such as the level. Examples include contextual indicators such as the existence of healthy public policy; laws restricting smoking in existence of healthy public policy; laws restricting smoking in public places, or social equity in access to care; social cohesion, public places, or social equity in access to care; social cohesion, etc. etc.

Morgenstern H. Ecologic studies in epidemiology: concepts, principles, and methods. Annu Rev Morgenstern H. Ecologic studies in epidemiology: concepts, principles, and methods. Annu Rev Public Health 1995; 16:61-81.Public Health 1995; 16:61-81.

Linking the focus of a measure to its Linking the focus of a measure to its applicationapplication

• Aggregate measures are typically used in Aggregate measures are typically used in descriptive studies; focus on the individuals descriptive studies; focus on the individuals within the population, i.e. idiographic. within the population, i.e. idiographic. They measure health They measure health inin the population the population

• Environmental measures can be used in Environmental measures can be used in descriptive, analytic or explanatory studiesdescriptive, analytic or explanatory studies

• Global measures mainly used in analytic Global measures mainly used in analytic studies; focus on generating theory studies; focus on generating theory (nomothetic studies). They measure health (nomothetic studies). They measure health ofof the population the population

Class ofIndicators

Etiological sequence

Individual(+ aggregate)

Global

Humanbiology Lifestyles

Physicalenvironment

Socialenvironment

Generalpolicies

Healthcare policies

Environmental

Levels of Intervention

Individual

Environmental

Policy

Populationdeterminants

Riskfactors

Individualoutcomes

Linking the target of a population Linking the target of a population intervention to the type of intervention to the type of measure measure Interventions can target people, environmental factors, or policy in general

…and to the presumed etiological sequence

These correspond to Morgenstern’s categories of measures used to evaluate the intervention…

History of changing History of changing approaches to measuring approaches to measuring population healthpopulation health

• Originally based on mortality rates. IMR is often Originally based on mortality rates. IMR is often used to describe level of development of a used to describe level of development of a countrycountry

• With declining mortality, people with chronic With declining mortality, people with chronic disease survive; morbidity & disability gain disease survive; morbidity & disability gain importance. Then . . .importance. Then . . .

• Concern with quality of life, not mere survivalConcern with quality of life, not mere survival

• To compare populations at different stages of To compare populations at different stages of economic development, it may be desirable to economic development, it may be desirable to combine mortality and morbidity in a single, combine mortality and morbidity in a single, composite indexcomposite index

2.2. Aggregate Measures:Aggregate Measures:(i) Mortality-Based (i) Mortality-Based

IndicatorsIndicators

Life expectancyLife expectancy

Expected years of life lostExpected years of life lost

Potential years of life lostPotential years of life lost

Expectancies versus GapsExpectancies versus Gaps

• From a typical survival From a typical survival curve, we can either curve, we can either consider the life consider the life expectancy (“E”), or expectancy (“E”), or the gap (“G”) between the gap (“G”) between current life expectancy current life expectancy and some ideal (here, and some ideal (here, the outer rectangle). the outer rectangle).

• Expectancies are Expectancies are generic; gaps can be generic; gaps can be disease-specificdisease-specific

G

0%

20%

40%

60%

80%

100%

0 10 20 30 40 50 60 70 80 90 100

E

Life ExpectancyLife Expectancy

• Summarizes all age-specific mortality Summarizes all age-specific mortality ratesrates

• Estimates hypothetical length of life Estimates hypothetical length of life of a cohort born in a particular yearof a cohort born in a particular year

– This assumes that current mortality This assumes that current mortality rates will continuerates will continue

Gaps: Expected Years of Life Gaps: Expected Years of Life LostLost

• Can use population life expectancy at the Can use population life expectancy at the individual’s age of deathindividual’s age of death

– Problem: different regions may have different life Problem: different regions may have different life expectancies. Cannot identify impact of a disease.expectancies. Cannot identify impact of a disease.

• Or: Standard Expected Years of Life LostOr: Standard Expected Years of Life Lost

– Reference is to an “ideal” life expectancyReference is to an “ideal” life expectancy

• E.g., Japan (82 years for women)E.g., Japan (82 years for women)

• Area between survivorship curve and the chosen normArea between survivorship curve and the chosen norm

Potential Years of Life Lost Potential Years of Life Lost (PYLL)(PYLL)

• PYLL = PYLL = ( “normal age at death” – actual ( “normal age at death” – actual age at death). Doesn’t much matter what age at death). Doesn’t much matter what age is chosen as reference; typically 75age is chosen as reference; typically 75

• Attempts to represent impact of a disease Attempts to represent impact of a disease on the population: death at a young age on the population: death at a young age is a greater loss than death of an elderly is a greater loss than death of an elderly personperson

• Focuses attention on conditions that kill Focuses attention on conditions that kill younger people (accidents; cancers)younger people (accidents; cancers)

• All-causes or cause-specific PYLLAll-causes or cause-specific PYLL

3. Aggregate Measures: 3. Aggregate Measures: (ii) Indicators that (ii) Indicators that

Combine Mortality & Combine Mortality & MorbidityMorbidity

Health expectanciesHealth expectancies

Health gapsHealth gaps

Composite Measures Composite Measures

• Composite measures combine morbidity and Composite measures combine morbidity and mortality into a health index. An index is a mortality into a health index. An index is a numerical summary of several indicators of numerical summary of several indicators of healthhealth

• Aims to represent overall health of a Aims to represent overall health of a populationpopulation

• Mortality data typically derived from life tables; Mortality data typically derived from life tables; morbidity indicators from health surveys, e.g. morbidity indicators from health surveys, e.g.

• Self-rated healthSelf-rated health

• Disability or activity limitationsDisability or activity limitations

• A formal health indexA formal health index

Survivorship Functions for Health Survivorship Functions for Health StatesStates

C

0%

20%

40%

60%

80%

100%

0 10 20 30 40 50 60 70 80 90 100

A

Survivors

Age

This diagram illustrates the composite health of a population.

The lower area ‘A’ shows the proportion of people in good health (however defined); it shows healthy life expectancy. The top curve shows deaths; intermediate curves represent various levels of disability.

Area ‘C’ represents the deficit of this population compared to an arbitrary ideal; this refers back to the notion of health gaps.

Deaths

More details on the combined More details on the combined indicators indicators • From the previous chart: From the previous chart:

– You can read from the bottom, and talk of You can read from the bottom, and talk of “health expectancies.”“health expectancies.”

– Or you can read from the top, and focus on the Or you can read from the top, and focus on the gap between current state and the ideal (these gap between current state and the ideal (these were discussed earlier) were discussed earlier)

– The bands in the middle indicate that the value The bands in the middle indicate that the value of a life lived in less than perfect health is less of a life lived in less than perfect health is less than a healthy life-year. These are “health-than a healthy life-year. These are “health-adjusted life expectancies”adjusted life expectancies”

– The indicators will fall in a descending sequence: The indicators will fall in a descending sequence: overall life expectancy, then health-adjusted life overall life expectancy, then health-adjusted life expectancy, then healthy life expectancy. expectancy, then healthy life expectancy.

2.1 Health expectancies2.1 Health expectancies

• Generic term: any expectation of life Generic term: any expectation of life in various states of health. Includes in various states of health. Includes other, more specific termsother, more specific terms

• Two main classes:Two main classes:– Dichotomous weighting for health states Dichotomous weighting for health states – Health state valuations for a number of Health state valuations for a number of

categoriescategories

Dichotomous expectanciesDichotomous expectancies

• Here full health is rated 1, and any state Here full health is rated 1, and any state of poor health (mild, moderate, or severe of poor health (mild, moderate, or severe disability) is rated 0.disability) is rated 0.

• This leads to Disability-free life This leads to Disability-free life expectancy (DFLE): weight of 1 for “no expectancy (DFLE): weight of 1 for “no disability” and 0 for all other states.disability” and 0 for all other states.

• = Expectation of life with no disability, or = Expectation of life with no disability, or Healthy Life Expectancy (HLE)Healthy Life Expectancy (HLE)

• Very sensitive to threshold of disability Very sensitive to threshold of disability chosenchosen

Illustrating dichotomous weights:Illustrating dichotomous weights:Life Expectancy and Disability-Free Life Life Expectancy and Disability-Free Life Expectancy, Canada, 1986-1991Expectancy, Canada, 1986-1991

0

10

20

30

40

50

60

70

80

90

1986 1991

Years

M F M F

Life expectancyFrom birth

Disability-free Life expectancy

Polytomous states and Polytomous states and valuationsvaluations

• A refinement is to incorporate many levels of A refinement is to incorporate many levels of disability and to count time spent with each level disability and to count time spent with each level of disability. = Polytomous model (three or more of disability. = Polytomous model (three or more health states defined)health states defined)

• Weights are assigned to each state; generally 0 to Weights are assigned to each state; generally 0 to 1.0. These may be added together and compared 1.0. These may be added together and compared across diseasesacross diseases

• This forms the Health-Adjusted Life Expectancy This forms the Health-Adjusted Life Expectancy (HALE) index(HALE) index

• First calculated for Canada by Wilkins. He used First calculated for Canada by Wilkins. He used four levels of severity & arbitrary weights. four levels of severity & arbitrary weights.

• Recent work uses utility weights. E.g. from Health Recent work uses utility weights. E.g. from Health Utilities Index, Quality of Well-Being Scale, Utilities Index, Quality of Well-Being Scale, EUROQoL, etc.EUROQoL, etc.

Health Expectancy by Income Level Health Expectancy by Income Level and Sex, Canada, 1978 (Wilkins)and Sex, Canada, 1978 (Wilkins)

0

10

20

30

40

50

60

70

80

1 2 3 4 5 1 2 3 4 5

Males Females

YearsSeverely disabled

Restricted

Minor limitations

Healthy

Low HighIncome Quintiles

Relationship between Life Expectancy, Relationship between Life Expectancy, Health Expectancy and Health-Adjusted Health Expectancy and Health-Adjusted Life ExpectancyLife Expectancy

Health-AdjustedLife Expectancy

LifeExpectancy

HealthyLifeExpectancy By down-weighting the

various levels of disability,the HALE falls between LE and HLE

Some HALE Results for Some HALE Results for CanadaCanada• Wolfson & Wilkins at Statistics Canada used data Wolfson & Wilkins at Statistics Canada used data

from the National Population Health Survey to from the National Population Health Survey to calculate HALEs, using the “Health Utilities Index” to calculate HALEs, using the “Health Utilities Index” to weight different levels of imperfect healthweight different levels of imperfect health

• The difference between LE and HALE is 11% for men, The difference between LE and HALE is 11% for men, and 15% for women, because women live longer and and 15% for women, because women live longer and suffer more chronic disease at older agessuffer more chronic disease at older ages

• They recalculated HALEs, deleting certain types of They recalculated HALEs, deleting certain types of disability, and found that sensory problems (eyesight, disability, and found that sensory problems (eyesight, hearing) were the major contributor in Canada to lost hearing) were the major contributor in Canada to lost years. Vision problem have a very minor effect on years. Vision problem have a very minor effect on health status, but are very common… Pain was the health status, but are very common… Pain was the second largest causesecond largest cause

• They also showed that less educated people both live They also showed that less educated people both live shorter lives, and also experience more disabilityshorter lives, and also experience more disability

• Source: Wolfson MC. Health Reports 1986;8(1):41-46Source: Wolfson MC. Health Reports 1986;8(1):41-46

Health Expectancies and Health Expectancies and Health GapsHealth Gaps

SLE

LE

HALE

HLE

LE SEYLL SURVIVAL

HALE HALYPOLYTOMOUS

HLE ?DICHOTOMOUS

Birth

LE = Life Expectancy; SLE = Standard LE; HALE = Health-Adjusted LE; HLE = Healthy LE; SEYLL = Standard Expected Years of Life Lost

Gaps

Expectancies

Age

Classifying Health GapsClassifying Health Gaps

• Gaps: Compare population health to Gaps: Compare population health to some target. = Difference between time some target. = Difference between time lived in health states less than ideal lived in health states less than ideal health, and the specified targethealth, and the specified target

• The implied norm or target is arbitrary The implied norm or target is arbitrary and must be explicit, but as long as and must be explicit, but as long as same for all populations being same for all populations being compared, does not mattercompared, does not matter

Examples of Health Gap Examples of Health Gap MeasuresMeasures• Gap measures use a weighting for Gap measures use a weighting for

intermediate health states. This is intermediate health states. This is necessary to combine time lost due to ill necessary to combine time lost due to ill health with time lost due to premature health with time lost due to premature mortalitymortality

• Quality Adjusted Life Years (QALYs)Quality Adjusted Life Years (QALYs)

– Common outcome measurement in clinical Common outcome measurement in clinical trials, program evaluationtrials, program evaluation

– Record extra years of life provided by therapy Record extra years of life provided by therapy and quality of that lifeand quality of that life

– Typically use utility scale running from 0 to 1Typically use utility scale running from 0 to 1

• DALYS (disability-adjusted life years)DALYS (disability-adjusted life years)

4. Environmental and 4. Environmental and Global MeasuresGlobal Measures

Environment,Environment,

Income inequalities,Income inequalities,

Health inequalities.Health inequalities.

Some Examples of Some Examples of Environmental IndicatorsEnvironmental Indicators

• Environmental indicators of health Environmental indicators of health status (water, air quality, etc.) status (water, air quality, etc.)

• Indicators of social interactions: Indicators of social interactions: changing patterns of crime; changing patterns of crime; volunteerismvolunteerism

• Scope of, and access to, social & Scope of, and access to, social & mental health institutionsmental health institutions

• Urban environmental quality; housing, Urban environmental quality; housing, hospitals, etc. hospitals, etc.

Some Examples of Global Some Examples of Global MeasuresMeasures

• Social solidarity; sense of identity; artistic Social solidarity; sense of identity; artistic output; public interest in health issues, output; public interest in health issues, etc.etc.

• Indicators of societal support: the “safety Indicators of societal support: the “safety net”net”

• Quality of social institutions for health Quality of social institutions for health (health protection laws, etc.) (health protection laws, etc.)

• Social cohesion, neighbourhood quality, Social cohesion, neighbourhood quality, social capitalsocial capital

Gini Coefficient: Measure of Gini Coefficient: Measure of Income InequalityIncome Inequality

• L(s) lies below line L(s) lies below line of equality when of equality when income inequality income inequality favours the richfavours the rich

• Gini coefficient is Gini coefficient is twice the area twice the area between the curve between the curve and the line of and the line of equality equality

% of income

% of population

L(s)

0 100

100

Measures of Health Inequalities Measures of Health Inequalities (I)(I)

• Index of Dissimilarity: Index of Dissimilarity: Absolute Absolute number or percentage of all cases that must number or percentage of all cases that must be redistributed to obtain the same mortality be redistributed to obtain the same mortality rate for all SES groups. rate for all SES groups.

• Index of Dissimilarity in Length of Index of Dissimilarity in Length of Life: Life: The absolute number or proportion of The absolute number or proportion of person-years of life that should be person-years of life that should be redistributed among SES strata to achieve redistributed among SES strata to achieve equal length of life in all.equal length of life in all.

Measures of Health Inequalities Measures of Health Inequalities (II)(II)

• Relative Index of Inequality: Relative Index of Inequality: Ratio of Ratio of morbidity or mortality rates between those at bottom morbidity or mortality rates between those at bottom of SES range to those at top. This is estimated using of SES range to those at top. This is estimated using regression and corrects for other factors. regression and corrects for other factors.

• Slope Index of Inequality: Slope Index of Inequality: Expresses health Expresses health inequality between top and bottom of social inequality between top and bottom of social hierarchy in terms of rate differences rather than rate hierarchy in terms of rate differences rather than rate ratiosratios

Measures of Impact of Measures of Impact of Interventions to Reduce Interventions to Reduce InequalitiesInequalities

• Population attributable risk: Population attributable risk: The The reduction in mortality that would occur if reduction in mortality that would occur if everyone experienced the rates in the everyone experienced the rates in the highest socioeconomic group highest socioeconomic group

•Population attributable life lost Population attributable life lost index: index: The absolute or proportional The absolute or proportional increase in life expectancy if everyone increase in life expectancy if everyone experienced the life expectancy of the experienced the life expectancy of the highest SES grouphighest SES group

Standardized Index of Standardized Index of Health InequalityHealth Inequality

• L(s) lies above line of L(s) lies above line of equality when ill-health equality when ill-health is concentrated among is concentrated among poor. poor.

• L*(s) is indirectly L*(s) is indirectly standardized curve standardized curve indicating unavoidable indicating unavoidable inequality (e.g., due to inequality (e.g., due to age-sex distribution)age-sex distribution)

• Inequality favours rich Inequality favours rich if L(s) lies above L*(s) if L(s) lies above L*(s)

Cum % of ill-health

Cum. % of population ordered by income

0100

100L(s)

L*(s) Line ofequality