james thomas, ginny brunton, alison o’mara-eves eppi-centre, social science research unit,

Community engagement strategies to reduce health inequalities:

a multi-method systematic review of complex interventions.

James Thomas, Ginny Brunton, Alison O’Mara-Eves

EPPI-Centre, Social Science Research Unit, Institute of Education, University of London

• This project was funded by the UK National Institute for Health Research (NIHR) Public Health Research Programme. The views and opinions expressed by authors in this presentation are those of the authors and do not necessarily reflect those of the Public Health Research Programme, NIHR, NHS, or the Department of Health.

• This report is in press with Public Health Research.• Project conducted by a team of researchers at the

Institute of Education, London School of Economics, and University of East London. All authors declare no conflicts of interest.

Funding and conflicts of interest

Project AimsRQ1. What are the range of models and approaches underpinning community

engagement? RQ2. What are the mechanisms and contexts through which communities are engaged? RQ3. Which approaches to community engagement are associated with improved

health outcomes among disadvantaged groups? How do these approaches lead to improved outcomes?

RQ4. Which approaches to community engagement are associated with reductions in inequalities in health? How do these approaches lead to reductions in health inequalities?

RQ5. Which types of intervention work best when communities are engaged? RQ6. Is community engagement associated with better outcomes for some groups when

compared to others? (In particular, does it work better or less well for children and young people?)

RQ7. How do targeted and universal interventions compare in terms of community engagement and their impact on inequalities?

RQ8. What are the resource implications of effective approaches to community engagement?

Aims of this presentation

• What are the theories underpinning community engagement interventions to reduce health inequalities?

• How do these relate to the effectiveness (including cost effectiveness) and implementation of such interventions?

• How do these findings shape new understandings of community engagement?

What is ‘Community Engagement’?

Brief: Community engagement for health inequalities =

‘approaches to involve communities in decisions that affect them’…

’…groups with distinct health needs and/or demonstrable health inequalities’

‘Health inequalities’ = gaps in the quality of the health of different groups of people based on differences in social, economic, and environmental conditions. (Marmot et al. (2010) Fair society, healthy lives: the Marmot review.)

Go ask the ‘experts’…

“I’m here because I’m interested in getting a good definition of community engagement…”

“…Well when you find one let us know”

Connected Communities: Communities, Cultures and Health & Well-Being Research Development Workshop (Cardiff, September 2011)

Community engagement: Panacea…or Pandora’s box?

• Encourages social justice, public accountability and better interventions

• Can “give a voice to the voiceless”: those who are socially excluded and disengaged from services

• Theory behind recommendations for community engagement often not linked to empirical evidence

• Much uncertainty about processes

• Fragmented, questionably poor quality evidence base supporting the effectiveness and cost-effectiveness of community

Community engagement to reduce inequalities in health: a systematic review

Conceptual framework

Statistical synthesis

Synthesis of process evals

Economic analysis

Aim of Conceptual Framework (Synthesis 1)

• To identify the range of models and approaches underpinning community engagement (CE); and

• To identify the mechanisms and contexts through which communities are engaged.

Protocol: Community engagement

The public Populations: • specific health needs• socioeconomic disadvantages

Communities:• of geography• of interests

Reasons for engaging

People invited for;• Ethics and democracy• Better services and health

People engage for;• personal gains: wealth & health• community gains• ideologies

Dimensions of engagement, e.g.• engaged in strategy/ delivery• state/ public initiated• degrees of engagement • individuals/ organised groups

Models of engagement, e.g.• consultations / service development• community development• grants for advocacy and support• controlling local facilities (e.g. sport centre)

Outcomes• Personal development: numbers & inequalities engaged, valued and connected• Community development: social capital• Programme development: communities’ influence on service/ delivery/ access• Health: overall, disadvantaged groups, health inequalities• Economics: time & cost of engagement, services developed, costs saved

ImplementationProcess evaluation of community engagement

Process evaluation of community’s intervention

Method of synthesisPrevious systematic reviews

Literature searching

Inclusion/exclusion screening

Reading key located literature for barriers/facilitators of successful CE

Coding

Analysis

Conceptual FrameworkSynthesis

Data extraction/synthesis methods

• Narrative format– Described the models, context and mechanisms of the participants,

interventions and approach to community engagement

• Barriers to, and facilitators of, implementation – Taken from the process evaluations using a formally developed tool – Conducted after the tool had been piloted on a sample of studies

• Findings from meta-analysis and cost-resource analysis

• Iterative ‘drawing together’ of all the above

Results: Included studies

• In total, 943 located potential systematic reviews elicited a total of 7,506 primary study titles and abstracts.

• Searches of other sources provided an additional 1,961 primary study

titles and abstracts.

• Duplicate removal, retrieval and screening of full-text reports resulted in the final inclusion of 361 reports of 319 studies in the map.

• Also purposively selected process-only and background discussion papers that provided key examples of community engagement processes (n=33).

(Health) intervention

Community engagement

Community Engagement in Interventions: Conceptual Framework




Actions ImpactConditionsMotivations

ActionsDefinitions ConditionsMotivationsCommunity

Participation

Community Participation

Definitions

Impact






Participation


Definitions

Communities• Of interests• Of geography

The public

Populations • With specific

needs•

Socioeconomically disadvantaged

Need/Issue• Felt• Expressed•

Comparative• Normative

Impact






Participation


Definitions


The public


needs•




People engage for:• Personal gains:

wealth / health• Community gains• Responsible

citizenship• Greater public

good / ideology

People invited for:• Ethics and

democracy• Better services and

health• Political alliances • Leveraging

resources

For intervention design:• Social learning• Social cognitive• Behavioral

Impact





ActionsDefinitions ConditionsMotivations

Community Engagement in Interventions• Main focus• Secondary focus• Incidental focus

Activity and Extent of Community Engagement• Involved in

intervention: ‐ Design‐ Delivery• Community:‐ Leading‐ Collaborating‐ Consulted‐ Informed



Definitions


The public


needs•







good / ideology




resources


Impact









Mediators of Community Engagement•

Communicative competence

• Empowerment

• Attitudes toward expertise

Context• Sustainability• Context of

the‘outside

world’• Government

policy & targets



Definitions


The public


needs•







good / ideology




resources


Impact











• Empowerment



the‘outside


policy & targets

Process Issues• Collective

decision-making

• Communication

• Training support

• Admin support

• Frequency• Duration• TimingInterventions• Acceptability• Feasibility• Cost



Definitions


The public


needs•







good / ideology




resources


Impact











• Empowerment



the‘outside


policy & targets

Process Issues• Collective

decision-making

• Communication

• Training support

• Admin support

• Frequency• Duration• TimingInterventions• Acceptability• Feasibility• Cost



Definitions


The public


needs•







good / ideology




resources


Impact

Beneficiaries• Direct- Engagees• Indirect- Community- Service providers- Intervention - Government- Researchers

Outcomes• Empowerment• Self-esteem, skills• Social capital• Mutual learning•

Attitudes/knowledge

• Health

Potential harms• Social exclusion• Cost overrun• Attrition• Dissatisfaction

What are the underlying mechanisms/contexts?

Utilitarian perspective• Pragmatic• Health systems focused• Those who initiate engagement

define ‘the community’• Underlying mechanism:

‘engagement’ may lead to better design/delivery

• Understanding what features of engagement improve effectiveness is critical

Social justice perspective• Community empowerment• Democratic right• Power shared/redistributed• Underlying mechanism: if

people are ‘signed up’ to the intervention/programme, participation and health improvements more likely

• Understanding how and why people ‘sign up’ is critical

The issue

The literature included in the review did not fall neatly into either one paradigm or the other…

The public

Patients

Peers

Community development

Consultation

Information

Participation

Empowerment

Service outcomes Social outcomes

Health outcomes

Health improvements

Health inequalities

Community empowerment

Need to bridge utilitarian and social justice rationales for empowerment

Unpacking ‘engagement’

1. Did the community identify the health need?2. Level of engagement in design

– Informed– Consulted– Collaborating– Leading

3. Level of engagement in delivery– Informed– Consulted– Collaborating– Leading

1. Empowerment2. Collaboration or consultation in intervention

design3. Lay-delivery

Theories of change identified in the theoretical synthesis

Theory of change for empowerment

Change is facilitated where the health need is identified by the community and they mobilise themselves into action.Example: inner-city childhood immunisation initiative

Community- observed problem

Community- perceived causes

of problem

Community mobilises into

action

Community-designed

intervention programme

Intervention is more

appropriate and greater

community ownership than

before

Outcomes (higher than they would have been

due to empowerment)

Theory of change for collaboration or consultation in intervention design

Observed problem

Health service designs

intervention to tackle the problem

The views of stakeholders are

sought

Intervention is more

appropriate than before

Implement intervention

(which has been altered by

stakeholders)

Outcomes (higher than they would have been

due to stakeholder

input)

The views of stakeholders are sought with the belief that the intervention will be more appropriate to the participants’ needs as a result. Example: healthy eating intervention

Theory of change for lay-delivered interventions

Observed problem

Health service designs

intervention to tackle the problem

Peers deliver the intervention

Delivery more empathetic,

credible, etc. than before

Outcomes (higher than they would have been due to

peer delivery)

Change is believed to be facilitated by the credibility, expertise, or empathy that the community member can bring to the delivery of the intervention.

Example: breastfeeding support

Outcome types

• Health behaviours (n=105)– e.g. breastfeeding, attend cancer screening

• Health consequences (n=38)– e.g. mortality, diagnosis

• Participant self-efficacy (n=20)• Participant social support (n=7)• Also a small number of community outcomes

and ‘engagee’ outcomes – not meta-analysed

• Significant statistical heterogeneity was expected in this review• “When operating across such a wide range of topics, populations

and intervention approaches, however, there is a disjunction between the conceptual heterogeneity implied by asking broad questions and the methods for analysing statistical variance that are in our ‘toolbox’ for answering them”

• Potential confounding variables or interactions amongst variables made it difficult to disentangle unique sources of variance across the studies

• Emphasis on magnitude of the effects and trends across studies

Statistical significance

The results

Results: Effectiveness studies (N = 131)

Countries– 4% (n = 5) UK– 86% (n = 113) USA– 4% (n = 5) Canada– 6% (n = 8) other OECD

Population/Health inequalities– 43% (n = 56) ethnic minorities– 26% (n = 34) low socioeconomic

position– 16% (n = 21) multiple health

inequalities

Age ranges– 60% (n = 79) young

people 11-21yrs– 50% (n = 65) adults 22-

54yrs

Sex– 60% (n = 79) mixed sex– 37% (n = 49)

predominantly female– 2% (n = 3)

predominantly male

Results: Health topic

Substa

nce abuse

Cardiovascu

lar dise

ase

Breastfeeding

Obesity preve

ntion / weig

ht reducti

on

Smoking ce

ssation

Public health

/ Health

promotion/ p

revention

Antenatal (pren

atal) c

are

Cance

r prev

ention

Diabetes p

revention/ m

anagem

ent

Physica

l acti

vity

Healthy eati

ng/ nutri

tion

Parenting

Immunisa

tion

Injury prevention

Smoking/to

bacco preve

ntion

Child illn

ess and ill

health

Disabiliti

es & ch

ronic i

llness

Child ab

use preven

tion

Hyperten

sion

Infant morta

lity0

2

4

6

8

10

12

14

16

1818

1413 13

12

87

6 6 65 5

4 43

2 21 1 1

Health Topics (N=131 studies)

Results: Overall mean effect

*** p < .001Statistical significance indicates the effect size estimate is significantly different from zero Note. 95% CI = 95% confidence intervaln = number of effect sizesτ2 = between studies variance

Heterogeneity

Outcome Pooled effect size estimate

95% C.I. n τ2 Q statistic I2

Health behaviours .33*** .26, .40 105 .093 604.62*** 82.80

Health consequences .16** .06, .27 38 .076 196.36*** 81.16

Participant self-efficacy .41** .16, .65 20 .278 480.44*** 96.05

Participant social support

.44*** .23, .65 7 .067 42.67*** 85.94

In general, interventions are effective!

Variation amongst studies needs to be explained

• Conducted moderator and regression analyses• Most of the analyses conducted on health

behaviour outcomes only because of small number of data points

• Not unexpected: none of the variables tested were statistically significant predictors of effect.

• Emphasis on trends across the data

Attempts to explain variation

Moderator of effect on health behaviours: Theory of change

• Most interventions were compared to a comparison condition that differed from the intervention in more ways than just community engagement

• For health behaviour outcomes, there were seven studies for which the only difference between the treatment conditions was the presence or absence of community engagement

• Analysis did not detect a significant difference between the studies with a direct comparison (effect size = .34) or indirect comparison (effect size = .33)

Direct comparisons

Moderator of effect on health behaviours: Marmot Review themes

Outcomes Marmot Review theme Mean ES 95% CI nHealth behaviours a

Modifiable health risks .24*** .11, .37 34

Best start in life .38*** .19, .56 24

Prevention of ill-health and injury .38*** .28, .48 47

Health consequences b

Modifiable health risks .23** .06, .40 17

Best start in life .05 -.29, .39 7

Prevention of ill-health and injury .12 -.06, .30 14

Other moderators tested• Single component interventions tended to be more effective

at improving health behaviours than multiple component interventions

• Universal interventions tended to have higher effect size estimates for health behaviour outcomes than targeted interventions.

Features of the interventions• Interventions conducted in non-community settings tended to be

more effective than those in community settings for health behaviour outcomes.

• Interventions that employed skill development or training strategies, or which offered contingent incentives, tended to be more effective than those employing educational strategies for health behaviour outcomes.

• Interventions involving peers, community members, or education professionals tended to be more effective than those involving health professionals for health behaviour outcomes.

• Shorter interventions tended to be more effective than longer interventions for health behaviour outcomes; this is probably confounded by levels of exposure or intensity of contact with the intervention deliverer.

Conclusions

• Overall, public health interventions using community engagement strategies for disadvantaged groups are effective in terms of health behaviours, health consequences, participant self-efficacy, and participant perceived social support.

• These findings appear to be not due to systematic methodological biases.

Conclusions

• However, unexplained variation exists amongst the effect sizes

• “…the evidence suggests that community engagement in public health is more likely to require a ‘fit for purpose’ rather than ‘one size fits all’ approach.”

Conclusions• Strengths

– Theories of change helped us to articulate proposed causal mechanisms

– Effects were evident despite substantial heterogeneity • Limitations

– Broad scope didn’t enable us to identify the ‘active ingredients’ of community engagement (i.e., which components work?)

– Lack of direct comparisons mean we don’t know how much of the effect is unique to community engagement

• More work to be done to understand more about which components contributed to effectiveness– Different methods of analysis may be required– Theories of change need further development

Acknowledgements

Co-authors:

David McDaid, Sandy Oliver, Josephine Kavanagh, Farah Jamal, Tihana Matosevic, Angela Harden

Thanks also to authors of and participants in the reviewed studies

EPPI-CentreSocial Science Research UnitInstitute of EducationUniversity of London18 Woburn SquareLondon WC1H 0NR

Tel +44 (0)20 7612 6397Fax +44 (0)20 7612 6400Email [email protected] eppi.ioe.ac.uk/

The protocol of the review is available to download at http://www.phr.nihr.ac.uk/

Thank you!

James Thomas [email protected] Brunton [email protected]

Alison O’Mara-Eves [email protected]

mailto:[email protected]



james thomas, ginny brunton, alison o’mara-eves eppi-centre, social science research unit,

Documents

poorer health

department of health

improved health outcomes

terms of community engagement

modifiable health determinants

aka community participation

institute of education

disadvantaged groups