Upload
others
View
6
Download
0
Embed Size (px)
Citation preview
L'influence des Liens Sociaux sur la Santé
Thomas W. Valente, PhD
November 28, 2012Montréal, Quebec
Professor, Institute for Prevention ResearchPreventive Medicine, Keck School of Medicine
University of Southern [email protected]
1
Acknowledgements
• Thanks:– Yun Jen
• National Institutes of Health (NIH)– NIDA/OBSSR
• Center for Prevention Implementation Methodology
• University of Miami
– NIAAA RC1AA019239NIAAA RC1AA019239• Social Networks and Networking ...
– NCI: R0110866440• Social Networks and Health
• Kar-Hai Chu, PhD, for animations
16es Journées annuelles de santé publique 1
What is a network?
• A set of relations connecting unitsF i d hi ( i h l li )– Friendships (e.g., in schools or onlin)
– Trading relations among countries
– Exchanges between firms
– Collaboration & cooperation among organizations.
Sample Social Network
Bob Sue
Ted
Sam
Liz
AmyDistance Bob to Sam=2Distance Sam to Bob=4
16es Journées annuelles de santé publique 2
Why Study networks?
• Spread Disease
• Spread Behaviors
• Spread new ideas
• Important Planning and Policy
• Understanding Markets and Diffusion
And so onAnd so on …
Relationships of 10th graders
16es Journées annuelles de santé publique 3
Discussion Network for SF MDs
After School Program Directors
1
14
5
410
12
11
15
13
16
55
57
59
KK1
Y1
KK- Kids
YMCA
14
3
1920
23
64
72
6
10
789
2
17
24
26
27
30
3241 51
48
9133
36
40
5354 4544
47
50
5856
65
77KK2
28
3029
3739
40
31
4635
38
42
43 95
52
34
88
49
60 6162
6380
89
118
Y2
Y3
16es Journées annuelles de santé publique 4
Network Properties (1)
• Homophily (like sorts with like)
Network Properties (2)
• Transitive (friends of friends become friends))
Bob
John Ken John
Bob
KenJohn Ken John Ken
16es Journées annuelles de santé publique 5
Network Properties (3)
• Scale Free: Centralized (e.g., some people are popular)p p )
Network Properties (4)
• Small World: Local clusters connected by bridgesy g
16es Journées annuelles de santé publique 6
Diffusion/contagion
• Today we’ll focus on the importance of networks for understanding how to usenetworks for understanding how to use networks to understand and change behavior.
Network Diffusion
16es Journées annuelles de santé publique 7
Network Exposure/Peer Influence
=non-user =user
A B
Person A has only 1 user in his/her network whereas person has 4 and so much more likely to adopt.
Evidence for network effects
• Adolescents with smoking friends are more likely to smoke;more likely to smoke;
• Physicians use the same therapies their colleagues use;
• Women in developing countries use the same contraception as their friends; and
• Smokers quit when their network quits.
• …
16es Journées annuelles de santé publique 8
Networks Can Also be Organizations or Countries
• Inter‐firm and market competition;
• Inter‐organizational collaboration; and
• Inter‐state or inter‐country diffusion.
3 Treaties Diffused 2001‐2010
1
FCTC Corrupt Pollute
0,4
0,5
0,6
0,7
0,8
0,9
0
0,1
0,2
0,3
16es Journées annuelles de santé publique 9
Diffusion at Policy LevelDiffusion at Policy LevelCountry Ratifications of FCTCCountry Ratifications of FCTC
100
nt Italy
40
60
80
lati
ve
Ad
op
tio
n P
erc
e
Lebanon
Iceland
19
0
20
20032003
20042004
20042005
20052005
20062006
20062007
20072007
20082008
2008
Time
Cu
mu
l
Norway
Co-membership on GlobalinkEarliest (15.5%) Non-Ratifiers (17.1%)
16es Journées annuelles de santé publique 10
FCTC Diffusion
To accelerate diffusion:Use Networks to Identify Opinion Leaders
• OL interventions are the most typical network interventionnetwork intervention
• Easy to measure
• Intuitively appealing
• Proven effectiveness
• Over 20 studies using network data to• Over 20 studies using network data to identify OLs and hundreds of others using other OL identification techniques
16es Journées annuelles de santé publique 11
Central People are Potential Opinion LeadersCentral People are Potential Opinion Leaders
Central Members
23
Simulation w/ 3 Initial Adopter Conditions
40
60
80
100
ent
Ad
op
ters
Opinion Leaders
Random
Marginals
0
20
1 2 3 4 5 6 7 8 9 10
Time
Per
ce
Marginals
16es Journées annuelles de santé publique 12
Role of Leaders / Influentials in the Adoption Process
6. ConfirmationOngoing experience confirms decision to
adopt May move to be an advocateadopt. May move to be an advocate
3. Persuasion / EvaluationTalks to others to gather information and
Mentally considers use
4. Decision to TryConvinced to trial to see if
it will be useful
5. AdoptionCompares personal experience with
expectations and considers future use
Regional
Local OLs
Peers
Six Steps in the
Adoption Process
1. AwarenessFirst learns of new product,
but unaware of benefits
2. InterestSearches for information about the new
product. Talks to trusted friends & experts
Mentally considers use
National Opinion Leaders
Regional
Opinion Leaders
Network Effects
Steps to Opinion Leader InterventionsSteps to Opinion Leader Interventions
1) Collect/obtain network data
2) Identify opinion leaders
3) Recruit them as champions
4) Convert them (if need be)
5) Assist them in their behavior5) Assist them in their behavior change promotions
26
16es Journées annuelles de santé publique 13
Network Interventions
• Network interventions are any change program designed to improve p g g porganizational performance, promotion campaigns, and/or social mobilization that uses network data to:– Select change agents– Define groupsg p– Affect network structure – Assist behavior change program
implementation
Principle 1: Program Goals Matter
• In some cases want to increase cohesion in others increase fragmentationin others increase fragmentation
• Increase/decrease centralization
• E.g., slowing spread of STDs requires different strategy than accelerating adoption of office automation
• Network Interventions Are not Agnostic to Content.
28
16es Journées annuelles de santé publique 14
Principle 2: Theory
• The type of change desired will be guided by theory (Behavior v Attitude)guided by theory (Behavior v Attitude)
• Understanding motivations for and barriers against behavior change is critical.
• A well-articulated theory of the behavior is often critical for successful interventions.
29
Principle 3: Learn As Well As Induce
• The interventionist should use network methodology to learn from the communitymethodology to learn from the community as much as try to influence it.
• Programs which meet the needs of their audiences are better received than those designed asymmetrically.
30
16es Journées annuelles de santé publique 15
Network Interventions
Strategy Tactic Operationalization
Identification LeadersBridgesKey PlayersPeripheralsLow Thresholds
Degree, Closeness, etc.Mediators, BridgesPositive, Negative
Proportions, Counts
Segmentation GroupsPositions
Components, CliquesStructural Equivalence, Hierarchies
Induction WOMSnowballMatching
Random ExcitationRDS, OutreachLeaders 1st, Groups 1st
Alteration (Manipulation)
Deleting/Adding NodesDeleting/Adding LinksRewiring
VitalityOn Cohesion, OthersOn Network, On Behavior
Graphical Displays of Intervention Choices
16es Journées annuelles de santé publique 16
Graphical Displays of Intervention Choices
Selecting a NI
• Availability and type of data
T f k– Types of networks
– Existing network structure
• Behavioral characteristics
– Existing prevalence
– Perceived characteristics such as cultural compatibility; cost; trialability; etc.
34
16es Journées annuelles de santé publique 17
Linking Theory to Intervention Strategy
• There are several theoretical mechanisms that drive contagion and behavior changedrive contagion and behavior change
• Evidence for a particular mechanisms suggests choice of intervention strategy or tactic
Influence Mechanisms Aligned with Interv. Choices
Mechanism Tactic
PowerConflict
LeadersBridges
CohesionIsolationThresholds
gKey PlayersPeripheralsLow Thresholds
Group IdentificationStructural Equivalence
GroupsPositions
Information diffusion WOMInformation diffusionHard to reach populationsClosureHomophily
WOMSnowballOutreachMatching
AttributesStructureStructure!!
Deleting/Adding NodesDeleting/Adding LinksRewiring
16es Journées annuelles de santé publique 18
Conclusions Conclusions
• Network theories and methods can be used t d t d ti ti f d to understand motivations for and processes of human behavior change.
• Network interventions have the promise of improving outcomes (health, organizational performance, etc.); and
• Test theoretically interesting network theories.
37
More reading and information:www-hsc.usc.edu/~tvalente/
16es Journées annuelles de santé publique 19