Transcript Slide 1

Principles & Practice of
Evaluation
Erica Wimbush
Head of Evaluation, NHS Health Scotland
ScotPHO Training, 29th March 2011
Outline







What is it?
Why do we do it?
Who is it for?
When do we do it?
How is population health data used?
What are the different types of evaluation?
Examples of ‘good’ evaluations
Examples
 Monitoring & Evaluation of Scotland’s Alcohol
Strategy (MESAS) – outcome planning
 Evaluation of Keep Well (Wave 1) – pilot phase
 Evaluation of WoSCAP – pre-testing
 Evaluation of the Smoking Ban - implementation
 Review of Scottish Diet Action Plan - review
Evaluation – what is it?
The making of a judgement
about the amount, cost or
value of something
Oxford English Dictionary
“The primary purpose of evaluation
is to improve the human condition
… to help determine if the promised
improvements of social programs
are actually delivered”
(Lipsey 2001)
“The ultimate goal of evaluation
is social betterment to which
evaluation can contribute by assisting
democratic institutions to better
select oversee, improve and
make sense of social programs”
Mel Mark 2007
Ensuring that the interests
of all individuals and groups
in society are served”
(Hopson 2001)
Focus of Evaluation
Planned interventions that are intended to bring
about change
 a policy or policy mechanism
 an agency or organisation
 a service
 a programme or project
 a practice
Why do we do evaluation?
 How effective/successful? To provide sound
evidence of programme effects - what actually
happened vs what was intended
 Better understanding about how programmes
work
 Generate learning from programme
implementation to inform decision-making and
improve practice
 Accountability - Assurance to funders about how
(public) money has been spent
Who is evaluation for?
Evaluation stakeholders
What sort of evaluation is valued?
Policy-makers
Effectiveness; what works?
Funders
Accountability
Planning & Performance
Performance monitoring/ targets
Managers
Developmental/formative evaluations
Researchers
Knowledge-building; research quality and
utility
Service users
Service quality - access, experience,
relevance to needs
Principles for evaluation
 Be focused – on the purpose, what you really need to
know, and what will be useful and used
 Be realistic – about what you can and should evaluate;
what is possible and what it is in your gift to influence
 Be proportionate – about how much evaluation is
appropriate
 Be convincing– to your evaluation audience: what will it
take to convince a reasonable person?
 Be honest – about why you are evaluating, what the
evaluation will be used for, and what you can claim
Types of evaluation
 OUTCOME – assesses effectiveness
 PROCESS – understanding the processes of
implementation and change
 FORMATIVE – feeds directly back into
programme development
 SUMMATIVE – review of evidence and learning
at the end of a period of implementation or
funding
Outcome evaluation designs
Experimental
 True - Random assignment to experimental
and control groups
 Quasi - Controlled design, non-random
assignment; non-equivalent control group
Non-experimental
 Time series analysis (single, comparative)
 Before and after
 Post-intervention (single, comparative)
Non-experimental designs – the
problem of causal attribution
Theory-based evaluation




Realistic Evaluation (UK, Ray Pawson & Nick Tilley)
Theories of Change (US, Aspen Institute)
Programme theory (Australia, Patricia Rogers)
Contribution Analysis (Canada, John Mayne)
 Used in initial planning stage of programme development
 Develops dearer thinking about the logical causal connections
between program goals, activities and outcomes
 Develop a series of statements (assumptions or hypotheses) about
how the program is intended to work that are testable via the
evaluation
Alcohol strategy - Theory of change
Alcohol strategy
Safer environment
Changed attitudes and
social norms
Reduced availability
substitution
Reduced affordability
Improved support from
services
•Drinkers
•Children affected by
parental drinking
Factors affecting delivery
Economic
impact on
alcohol industry
Reduced
alcohol
consumption ;
safer patterns
of drinking
Reduced
alcohol-related
harms
External factors
affecting
outcomes
Key terms: Outputs, Outcomes,
Results, Impacts
OUTPUTS –
Direct Control
INPUTS
ACTIVITIES
RESULTS – Visible
& Direct Influence
OUTPUTS
REACH
OUTCOMES
Expected IMPACTS –
not Visible, Indirect
Influence
OUTCOMES
OUTCOMES
OUTCOMES
Resources
What you do
What is
produced
Who the
outputs reach
(target group)
Immediate
results of
use/uptake
(eg KASA)
Results chain
Changes in
behaviour,
practice or
environments
Changes in
population
health
status
When do we do evaluation?
Review
Implement
-ation
Planning
Development
Evaluation in the Programme/Policy Cycle
Review
Impact/
outcome
evaluation
Implement
-ation
Performance
monitoring
Planning
Development
Evaluation of Ban on Smoking in Public Places
Use of population health data
 Use of logic model to map expected outcomes related to
smoke-free legislation
 Routine data on ETS health outcomes (e.g. hospital
admissions for acute coronary syndrome; consultations
for CHD, COPD and asthma)
 Routine data from SHS on smoking behaviour and quit
attempts – changing patterns before-after legislation
 Routine economic data (LFS, GDP, IDBR, ABI) –
changes in the economic performance of the hospitality
sector before-after legislation (employment, turnover,
profitability, openings/closures)
 Primary research – seven linked studies
Journal of Public Health, Volume 28, Number 1, March 2006 , pp. 24-30(7)
http://jpubhealth.oxfordjournals.org/cgi/content/abstract/28/1/24
Logic Model of Expected Outcomes Associated
with Smoke-free legislation
Expected Outcomes
Short-term
0-2 mths
Implementation
of smoke-free
legislation
Enforcement of
smoke-free
legislation
Increasing awareness 1
of health risks of ETS,
change in attitudes
towards ETS exposure
Intermediate
> 2 –12 mths
Long-term
> 12 mths
Reduction in exposure to ETS
2
Reduced ETS exposure
Increasing compliance with
smoke-free legislation
3
Sustained compliance with
smoke-free legislation
Increasing support for 4
legislation and change in
smoking cultures
Sustained cultural change
Reduction in smoking prevalence and tobacco consumption
Reduction in tobacco-related morbidity and mortality
5
6
Reduction in costs to health
service of tobacco-related
illness
Variable economic impact on
hospitality sector
health inequalities
7
Reduction in
health inequalities
8
Long-term
outcomes
Reduced inequal in CHD and cancer
Intermediate
outcomes
Reduced adult smoking rate
Short-term
outcomes
Increased % smokers who have
successfully quit at 1 month
(8% between 2008/9 – 2010/11)
Reach
Adult smokers who want to quit
Outputs
Smoking cessation services
Activities
Actions to deliver effective SC services
(NRT + brief advice) in key settings
Inputs
Budget, staff, training, data infrastructure
Results Chain
Performance
monitoring –
HEAT targets
H6 HEAT target
(2008-2011)
Evaluation in the Programme/Policy Cycle
Review
Impact/
outcome
evaluation
Planning
ImplementDevelopment
ation
Pretesting
Performance
monitoring
systems
Evaluation
of pilot
initiatives
Evaluation of Keep Well (Wave 1)
Evaluation aim
To assess the feasibility and challenges of delivering Keep Well, and
the effectiveness of different approaches to engagement and service
redesign, with a view to incorporating the lessons learned from the
pilots into subsequent waves of implementation
 Understanding of the programme - fit between national and local
pilots
 Rationales for different approaches in terms of feasibility and
doabilty
 Track links between activities-processes-outcomes
 Framework for comparing approaches across pilot areas and
possible unintended effects
 Evaluability assessment
Evaluation of Keep Well (Wave 1)
Use of population health data
Phase 1 (2007-10)
Informed by Theory of Change - understanding the process of implementation
of Keep Well
1: Tracking theories of change at national level and local pilots
2: Tracking the impact of KW on ‘anticipatory care’ in the target population
using secondary data
Phase 2 (2009-10)
Informed by Realistic Evaluation - deeper understanding of certain facets
through use of case studies
Practice level case studies to assess the impact of aspects of Keep Well
(informed by Phase 1 findings)
Patient and practice experiences (2009-10)
Collection of quantitative and qualitative data at practice level, and patient level
including patients recruited via Keep Well practices and community-based
venues.
Pre-testing – Bowel Cancer Campaign (WoSCAP)
Use of Qualitative research
5 concepts pre-tested in 6 focus groups
 ‘niggling worries’ addressed real
barriers to action –inertia and fear
 Proved compelling and intriguing due
to dialogue with man confronting his
fears
 It had a direct call to action –‘go to
your doctor if your bowel habits
change or you have blood in your
motions’
 It was felt to be an empathetic way of
tackling the fear that surrounds the
subject area
Evaluation in the Programme/Policy Cycle
Outcome
planning
Health Impact
Assessment
Review
Impact/
outcome
evaluation
Implementation
Performance
monitoring
Planning
Development
Pretesting
Evaluation
of pilot
initiatives
Alcohol strategy – MERGA/MESAS
Role of evaluation in strategic planning
 Planning stage – scope out the scale and nature
of the problem and potential solutions.
 Role of MERGA and MESAS
 Developing a shared understanding of problem
 Mapping the interlinked outcomes and potential
pathways; identifying range of plausible measures
needed and target groups (logic modelling)
 Developing monitoring and evaluation plans
Outcome/Results Planning
Intermediate outcomes
Model
2:
Model
3:
Children in need
receive timely and
appropriate support
Long term outcomes
A culture in which low alcohol consumption is valued and accepted as the norm
Less absenteeism +
presenteeism in
educational
establishments
Safer drinking + wider
environments
Fewer children
affected by parental
drinking
Reduced
acceptability of
hazardous drinking
and drunkenness
Model
4:
Study 3
Model
5:
Study 1
Model
6:
Model
7:
Study 2
Less absenteeism +
presenteeism the
workplace.
Study 4
Increased knowledge
and changed
attitudes to alcohol +
drinking
Less alcohol related
incapacity
Reduction in
Individual and
population
consumption
Reduced availability
of alcohol
Study 5
Reduced affordability
of alcohol
Individuals in need
receive timely,
sensitive &
appropriate support
Increased
educational
attainment
Less alcohol related
violence/abuse,
offences and ASB
Safer &
happier
families and
communities
National
outcomes
Reduce
significant
inequalities
Reduced
health, social
care, justice
costs
Study 6
Safer drinking
patterns
Study 7
Increased
workplace
productivity
Reduced alcohol
related injuries,
physical and
psychological morbidity
+ mortality
Fewer children
affected by maternal
drinking during
pregnancy
Healthier
individuals
and
populations
Reducing alcohol
related harm: strategic
logic model
Evaluation in the Programme/Policy Cycle
Outcome
planning
Review
processes
Review
Impact/
outcome
evaluation
Implementation
Planning
Health
Impact
Assessment
Development
Pretesting
Performance
monitoring
systems
Evaluation
of pilot
initiatives
Review of Scottish Diet Action Plan
Use of population health data
Aim - Progress with implementation and impacts 1996-2005
Community level impacts
Twenty-07 cohort study
Qualitative research
Population level impacts
Dietary targets
Trends in eating out,
breastfeeding, food retailing
Review Panel
SDAP Implementation
Programme evaluations
Evidence from
implementation bodies
and food industry
International comparisons
Expert commentary on
food and health policies and
implementation in 13
countries
Evaluation in the Programme/Policy Cycle
Outcome
planning
Review
processes
Review
REAL
EFFECTS
Impact/
outcome
evaluation
Implementation
Planning
Development
Health
Impact
Assessment
LIKELY
EFFECTS
Pretesting
Performance
monitoring
systems
Evaluation
of pilot
initiatives
What makes a ‘good’ evaluation?
Influences
decision making
Contributes
to the
evidence
base