University of Exeter Medical School 19 January 2017 Prof James Goodwin Chief Scientist Age UK Prof José Iparraguirre Chief Economist Age UK Index Rationale Data sources Process Conceptual framework ID: 550855
Download Presentation The PPT/PDF document "Age UK’s Index of Wellbeing in Later L..." is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Age UK’s Index of Wellbeing in Later LifeUniversity of Exeter Medical School19 January 2017
Prof James Goodwin
Chief Scientist, Age UK
Prof José Iparraguirre
Chief Economist,
Age
UKSlide2
IndexRationaleData sources Process
Conceptual framework
Data
analysis
Factor Analysis
Structural Equation Model
Principal component analysis – Domains
Inequality in Wellbeing in later life
Possible usesSlide3
RationaleSlide4
RationaleUp to now, there has been no way to measure in the round:What
is important in later life
How older people are doing
How much we love later life
Whether this is a great place to grow old(
er
)
“
While researchers have developed important, valid, and reliable instruments to assess different aspects of well-being… we believe that the various models have not yet been integrated into a single and coherent scale covering well-being, overall, and in the most important domains of life
.” (
Prilleltensky
et al, 2015) Slide5
RationaleAn index should:Combine multiple indicators into one single measure, but also allow for dis-aggregation
Include strata such as dimensions or indicators
Help understand inter-relationships between indicators and their direct and indirect effects on wellbeingSlide6
RationaleAge UK aims to measure how older people in the UK are doing. We are using the term wellbeing as the outcome of interest.
We need to be able to understand where and why wellbeing is low to inform our influencing activity. And
to gain
an understanding of the policy and practical levers for improving wellbeing.
Local
Age UKs need data intelligence to target their support services. We hypothesise low wellbeing is a proxy for need. Providing intelligence on where people with low wellbeing live will
help inform local services
. Slide7
2. Data sources Slide8
Data sourcesSlide9
3. ProcessSlide10
What constitutes wellbeing?What does NOT affect your wellbeing?How have the factors changed through your lifetime?How much do the factors differ between people – in general?
Consultation with older people - objectivesSlide11
Consultation with older people
Themes
from workshops
Good physical and mental health
Cognitive ability
Coping with ill health
Coping with stress (in general and stress of ageing)
Mental resilience
Feeling respected
Peace of mind
Religious belief
Being independent
Mobility
Mutual support with other people
Healthcare
Social care
Good family relationships
Good friendships
Not being lonely
Living in own home
Feeling safe
Enough money
Having things to do
Leisure time
Healthy lifestyle
Freedom of expressionSlide12
4. Conceptual frameworkSlide13
Conceptual frameworkSlide14
5.
Data analysisSlide15
Review of existing approaches and creation of conceptual model
Exploration of existing datasets
. Choice
of datasets and variables
Factor analysis to boil down
variables (when necessary
eg
GHQ-12
, Big-5, Cognitive
Ability)
Structural Equation
Modelling
Multigroup
invariance analysis
(by gender, age groups, England/Rest of UK)
Process
Estimation of individual WB
scores
Principal component analysis
Domain selection and namingSlide16
35 factors identified from lit revTwo datasets explored: ELSA & UsocELSA: 50+ in EnglandUSoc: 16+ in UK
USoc
more comprehensive.
35 factors … over 200 variables!
Some variables were immediate (e.g. gender, age). Others resulted from adding different variables (weighted or not –it depends) (e.g. morbidity, benefit income)
And still some needed factor analysis (e.g. GHQ-12, cognitive capacity)
Dataset preparationSlide17
6. Factor AnalysisSlide18
Factor Analysis(P.
Tryfos
, Methods for Business Analysis and Forecasting: Text & Cases, Wiley,
1998)
Factor analysis is a method for investigating whether a number of variables of interest
are
linearly related to a
smaller number of unobservable
factors
Factor analysis consists of a number of statistical techniques the aim of which is to
simplify
complex sets of data.
(P. Kline, An Easy Guide to Factor Analysis, Routledge, 2014)Slide19
Factor Analysis - example
concentration
loss of sleep
playing a useful role
capable of making decisions
constantly under strain
problem overcoming difficulties
enjoy day-to-day activities
ability to face problems
unhappy or depressed
losing confidence
believe worthless
general happiness Slide20
7. Structural Equation ModelSlide21
Once we got the 35 factors, we wrote up a Structural Equation Model: a `comprehensible statistical approach to testing hypotheses about relationships among observed and unobserved variables’(R. Hoyle. Structural Equation Modeling: Concepts, Issues, and Applications, Sage, 1995
. p.1)
WB is also unobservable
It is defined by these 35 factors
(our hypothesis)
Many of these factors are unobserved and inter-related
Structural Equation ModellingSlide22
(Handbook of Structural Equation Modeling, R. Hoyle (ed.). Guilford Publ., 2014, p. 7
Structural Equation ModellingSlide23
SEM
ModelSlide24
SEM Results - Contribution
of individual
indicatorsSlide25
8. Principal component analysis – DomainsSlide26
Principal Components AnalysisThe central idea of principal component analysis is to reduce the
dimensionality of
a data set in which there are a large number of
interrelated variables
, while retaining as much as possible of the variation present
in the
data set. This reduction is achieved by transforming to a new set
of variables, the principal components, which are uncorrelated, and which
are ordered
so that the first
few
retain most of the variation present in
all
of the
original variables.
(I.T. Jolliffe. Principal Components Analysis, Sringer, 2002. p.1) Slide27
Principal Components AnalysisSlide28
Principal component analysis
to group the indicators and results into
broad
areas (domains
).
Nine identified;
grouped into Five Slide29
29
Age UK’s Wellbeing in Later Life Index
1
PERSONAL
2.1 Social participation
2.2 Civic participation
2.3 Cultural participation
2.4
Neighbourliness
2.5 Friends
2.6 Big5 Personality
2
SOCIAL
1.1 Living arrangement
1.2 Marital status
1.3 Children
1.4 Education
1.5
Carer
1.6 Intergenerational
1.7 Cognitive ability
2.1 Social participation
2.2 Civic participation
2.3 Cultural participation
2.4
Neighbourliness
2.5 Friends
2.6 Big5 Personality
2
SOCIAL
3.1 Longstanding illness
3.2 Co-morbidity
3.3 Mental health
3.4 Mental wellbeing
3.5Sports activity
3
HEALTH
4.1 Employment
4.2 Income support
4.3 Pension
4.4
Housing wealth
4.5 Financial wealth
4
RESOURCES
4.6 Home ownership
4.7 material resources
5.1 Health services
5.2 Leisure services
5.3 Public transport
5.4 Shopping
5
LOCAL
5 domains listed across the top; the indicators are listed below within eachSlide30
Results: score for each domain as a percentage of the highest score attained Slide31
9. Inequality in Wellbeing in later lifeSlide32
Comparing the top and bottom
of
the
WB score distributionSlide33
Inequality – bottom 20 percent compared to top 20 percentSlide34
Inequality in overall wellbeing: some early findingsPeople aged 60+ in the UK who are in the bottom fifth of the wellbeing scale are:
More likely to be female and widowed
Half as likely to be married
More than twice as likely to be living alone
Much less likely to take part in cultural, social or civic events
Between three and four times as likely to have a longstanding illness and fourteen times as likely to have three or more diagnosed health conditions
Between four and five times more likely to have no educational qualifications at GCSE or above compared to those in the top fifth for wellbeing
On average, those people with the highest level of wellbeing (the top 20 percent) have:
More than 13 times as much financial wealth and about 14 times the income of those in the bottom twenty percent
They are also seven times more likely to participate in sport and have on average 50% more friends.Slide35
Subgroups
Overall
PERSONAL
SOCIAL
HEALTH
RESOURCES
LOCAL
Total
53.0
59.7
55.0
45.4
49.8
55.0
Sex
Men
54.0
61.5
55.0
46.8
51.6
55.6
Women
52.1
58.2
55.1
44.3
48.3
54.4
Age groups
age 60-64
55.1
67.1
55.1
48.0
49.6
54.4
age 65-69
55.8
65.6
56.4
49.2
51.4
53.7
age 70-79
53.4
59.6
55.9
45.5
50.4
55.8
age 80+
47.3
48.4
51.6
38.1
47.1
55.7
Limiting Long term illness
No
60.0
62.8
57.5
66.5
54.5
56.0
Yes
48.7
57.9
53.5
32.5
46.9
54.3
Group-specific
and
domain-specific wellbeing scoreSlide36
Subgroups
Overall
PERSONAL
SOCIAL
HEALTH
RESOURCES
LOCAL
Tenure
Home owned outright
57.1
62.8
57.0
48.8
60.7
55.2
Outstanding mortgage
54.1
66.4
56.9
46.1
44.4
53.6
Rented
43.7
51.5
49.9
37.8
28.2
54.8
Education
Higher
60.6
71.0
61.2
51.6
58.7
55.7
Not higher
49.9
55.3
52.3
42.8
46.3
54.6
Legal marital status
single, nvr marr/civ p
50.3
53.3
53.7
45.5
45.4
54.8
married
56.9
67.9
56.5
48.6
54.6
55.4
divorced
49.0
54.6
54.0
42.3
40.6
53.5
widowed
48.1
48.1
53.0
40.8
46.2
54.9
Group-specific
and
domain-specific wellbeing scoreSlide37
Subgroups
Overall
PERSONAL
SOCIAL
HEALTH
RESOURCES
LOCAL
Co-morbidity
0 = no morbidity
59.4
64.1
57.2
62.5
54.8
56.5
1
54.9
60.7
56.7
49.7
51.0
55.5
2
51.0
58.3
54.9
39.0
48.9
54.1
3
47.0
56.2
51.8
30.2
45.3
54.1
4
42.7
53.7
47.9
21.5
42.5
52.9
5+ ill-heath conditions
39.9
52.9
47.4
14.5
38.2
50.4
Group-specific
and
domain-specific wellbeing scoreSlide38
Inequality in WB and some indicatorsSlide39
Inequality in WB and some indicatorsSlide40
10. Possible usesSlide41
Possible uses
Record changes in WB of all older people or sub-groups
Identify relative importance of characteristics and variables
Investigate what determines low WB scores in later life
Inform interventions and policy and their evaluations -
Microsimulation
Develop local WB indicesSlide42
Thank you