An Introduction to Social Networks Methods in Health Services Research
AcademyHealth Annual Meetings, 2007 Doug Wholey
Health Policy & Management, School of Public Health, University of Minnesota
David Krackhardt
The H. John Heinz III School of Public Policy and Management, Carnegie Mellon University
8/11/2008
Social Networks Methods in Health Services Research
1
Presentation Organization
Introduction to social networks methods Three examples of social networks research in health services research
8/11/2008
Social Networks Methods in Health Services Research
2
Presenters
Doug Wholey
Professor, Division of Health Policy & Management, School of Public Health, University of Minnesota Professor of Organizations at the Heinz School of Public Policy and Management and the Tepper School of Business, Carnegie Mellon University Center for Computational Analysis of Social and Organizational Systems (CASOS)
David Krackhardt
2007 Summer Institute, June 25 - July 1, 2007
Social Networks Methods in Health Services Research 3
8/11/2008
Handouts
Social Networks Resources Life in the Pentagon Presentation
8/11/2008
Social Networks Methods in Health Services Research
4
Introduction to Social Networks Methods David Krackhardt
8/11/2008
Social Networks Methods in Health Services Research
5
Examples of Social Network Studies in Health Services Research
Doug Wholey
8/11/2008
Social Networks Methods in Health Services Research
6
Examples
Evaluating the Minnesota Cancer Alliance
Collaborator: Michael Pfeffer, MPH candidate
Evaluating an intervention to coordinate agencies providing services to people in prostitution in North Minneapolis
Collaborator: Kate Downing, MPH candidate Collaborators: David Knoke, Bill Riley, and Amy Wilson
Social Networks Methods in Health Services Research 7
Teams and sub-teams in medical practice
8/11/2008
Issues in Studying Networks: Identifying Network Boundaries
Realist
Network exists as a social object, there is a defined boundary and membership, there is shared knowledge Minnesota Cancer Alliance, medical groups, associations, rural health networks, assertive community treatment teams
Network membership is defined by research interests Agencies serving people in prostitution agencies Public health systems – organizations providing public health related services in a geographic area
Nominalist
8/11/2008
Social Networks Methods in Health Services Research
8
Issues in Studying Networks: Which Relationships?
Types of relations
Collaboration
Levels of collaboration
Who do you share information with about program development? Who do you work with to develop programs? Who are you collaborating with to obtain funding for programs?
Social Networks Methods in Health Services Research 9
8/11/2008
Issues in Studying Networks: Relational, Attribute, & Event Data
Relational data
N (actor) x N (actor) matrix Who interacts with who?
N (actor) x A (attribute) Can use to construct a relational matrix showing similarity between two actors (same gender, difference in age) N (actor) x A (event) matrix Can use to construct a relational matrix to who who attends the same meetings
Social Networks Methods in Health Services Research 10
Attribute data
Event/Affiliation data
8/11/2008
Issues in Studying Networks: Event Data to Relational Data
Attended Meeting 1 2 3 4 0 1 1 1 1 0 1 1 1 0 0 0
George Sam Jane
= A (Event Matrix)
AxA‟ (Common Attendance) =
Use UCINET‟s /Data/Affiliations tool
8/11/2008
George Sam Jane
G 2 2 0
S 2 3 1
J 0 1 1
11
Social Networks Methods in Health Services Research
The Minnesota Cancer Alliance
A coalition of health care organizations founded in 2005 with the goals of
Reducing disparities in cancer screening and treatment Improving access to information about locallyavailable services for cancer patients and their families Increasing colorectal cancer screening Increasing the tobacco excise tax and expanding clean indoor air http://www.cancerplanmn.org/The_Minnesota_C ancer_Alliance.html
Social Networks Methods in Health Services Research 12
8/11/2008
Comprehensive Cancer Control Collaborative
For cancer prevention and treatment in a community,
Develop a shared comprehensive vision Eliminate duplications, integrate efforts, reduce organizational and programmatic fragmentation Plan and implement an evidence based program Identify gaps and prioritize action Reallocate resources Advocate with one voice “Increased coordination of partner activities and enhanced collaboration” (6). Leslie S. Given, Bruce Black, Garry Lowry, Philip Huang, & Jon F. Kerner, 2005, Collaborating to conquer cancer: A comprehensive approach to cancer control, Cancer Causes and Control 16(Suppl): 3-14.
Social Networks Methods in Health Services Research 13
8/11/2008
Alliance Evaluation and Committees
Evaluation Who is active where?
People Organizations
How does participation evolve? Who is central in each committee? Who is central across committees? What is the structure of interaction
Structure = repeated patterns of interaction
Committees Community Health Worker Cervical Cancer Screening Colorectal Cancer Screening Access to Information Health Disparities Steering Committee Evaluation Committee Communications Committee Finance Committee Membership Committee
14
8/11/2008
Social Networks Methods in Health Services Research
Methods
Data Source: Meeting Archives
Meeting Attendance Individual - Organizational Used UCINET to translate affiliation data (copmmon meetings attended) to relational data Number of meetings attended by both members of a dyad is strength of attachment to committee
Methods:
8/11/2008
Social Networks Methods in Health Services Research
15
Health Disparities Committee (Three or More Meetings)
8/11/2008
Social Networks Methods in Health Services Research
16
Health Disparities Committee (Four or More Meetings)
8/11/2008
Social Networks Methods in Health Services Research
17
Health Disparities Committee (Four or More Meetings)
8/11/2008
Social Networks Methods in Health Services Research
18
Colorectal Cancer Committee (1 or More Meetings)
8/11/2008
Social Networks Methods in Health Services Research
19
Colorectal Cancer Committee (6 or More Meetings)
8/11/2008
Social Networks Methods in Health Services Research
20
Colorectal Cancer Committee (14 or More Meetings)
8/11/2008
Social Networks Methods in Health Services Research
21
Further Analyses: Which actors are central across committees?
… Disparities Access to Information Cervical Colorectal
8/11/2008
Social Networks Methods in Health Services Research
22
What Can We Learn from Meeting Attendance
Which members within the committee?
Are there core and peripheral members? How stable are relationships?
Which actors coordinate/bridge across committees?
8/11/2008
Social Networks Methods in Health Services Research
23
Next Steps
Initial analyses took advantage of unobtrusive measures: Meeting attendance Why are some members more involved than others in a committee?
What is the content of the relationship? Watching? Collaborating?
Research to develop a better understanding of committees and involvement
Field experience by Michael Pfeffer
Develop a survey to understand involvement and describe networks better
Social Networks Methods in Health Services Research 24
8/11/2008
Improving Service Delivery for People in Prostitution
Research and Intervention led by Folwell Center for Urban Initiatives
North Minneapolis, MN Lauren Martin, Ph.D. Kate Downing, MPH Candidate
8/11/2008
Social Networks Methods in Health Services Research
25
The Problem – North Minneapolis and Prostitution
Poverty Population of Color Mental Health and Substance Abuse A doubly stigmatized, vulnerable population
Project Focus Area, North Minneapolis
8/11/2008 Social Networks Methods in Health Services Research 26
The Problem: Service Delivery
Access to Services
Few prostitution specific services
Politics of Delivery
Ideological Barriers Funding Scarcity
A difficult context to organize
8/11/2008
Social Networks Methods in Health Services Research
27
The Evaluation
Evaluate an intervention to coordinate agencies Network Analysis
Understanding the mechanism Pre-Post Evaluation
Funding search in progress for intervention and evaluation
8/11/2008
Social Networks Methods in Health Services Research
28
The Evaluation
What does Collaboration really mean?
What does it mean to „work with?‟ Asking probing questions:
Do you go to the same meetings? Do you collaborate on programs? Do you go after funding together?
8/11/2008
Social Networks Methods in Health Services Research
29
The Intervention: Northside Women‟s Space
A point of connection:
For service providers to work together For women involved in prostitution to find safety, community and services
The space will offer:
Resource guide Non-judgmental referral services Workshops Safe sex supplies, snacks, etc.
Social Networks Methods in Health Services Research 30
8/11/2008
The Intervention: Northside Women‟s Space
One central actor managing the space
Folwell Center for Urban Initiatives Kate Downing
Support development and research
Other agencies will staff the space
Adhering to soon available best practices
8/11/2008
Social Networks Methods in Health Services Research
31
Challenges
The questions we can‟t ask… Maintaining neutrality and building bridges Developing trust and keeping it Unknown, invisible population size
A barrier to political will, funding
Identifying agencies to survey
Social Networks Methods in Health Services Research 32
8/11/2008
Expected Outcome: Northside Women‟s Space
More network ties Central node for the Northside
Women‟s Space
Success means:
Care integration Improved quality of care and service Improved outcomes
Greater visibility of the problem Improvement in service delivery Continuing needs assessment Advocacy for change through collaboration Shift from crisis oriented care to primary prevention
8/11/2008
Social Networks Methods in Health Services Research
33
Care Teams in Medical Practice
The Context & Argument
The Mechanism
Health care is becoming more differentiated New technologies and technicians Professional jurisdictions shifts Care is shifting towards chronic rather than acute conditions Chronic conditions require integration of a large variety of services: Social, mental, and medical services Lack of care integration results in inadequate care for individuals with chronic conditions
Care teams Interdisciplinary teams serving a population Task work (service provision) coordinated with team work (systems, coordinating service delivery) Hope Integrated care Greater productivity through reduced waste (lean management, QI) Better patient outcomes Better team member outcomes
8/11/2008
Social Networks Methods in Health Services Research
34
Care Team Examples
Chronic Care Teams http://www.improvingchroniccare.org / Assertive Community Treatment (ACT) Teams (http://www.actassociation.org/actModel/ ) Prepared Practice Teams (http://www.ihi.org/IHI/Topics/ChronicConditi ons/AllConditions/ImprovementStories/Purs uingPerfectionReportfromHealthPartnerson PreparedPracticeTeams.htm) Medical Homes and Patient Homes
Social Networks Methods in Health Services Research 35
8/11/2008
Teams
“A team can be defined as
(a) two or more individuals who (b) socially interact (face-to-face or, increasingly, virtually); (c) possess one or more common goals; (d) are brought together to perform organizationally relevant tasks; (e) exhibit interdependencies with respect to workflow, goals, and outcomes; (f) have different roles and responsibilities; and (g) are together embedded in an encompassing organizational system, with boundaries and linkages to the broader system context and task environment.”
Steve W. J. Kozlowski and Daniel R. Ilgen. 2006. "Enhancing the Effectiveness of Work Groups and Teams." Psychological Science in the Public Interest 7:77-124 (79).
8/11/2008
Social Networks Methods in Health Services Research
36
Issues in Identifying Care Teams
Boundary specification – who is a team member?
Floaters? Visiting consultants? Anyone who has anything do do with care delivery?
Assertive Community Treatment & Individual Treatment Teams Football teams
Teams and sub-teams
Formal and informal teams
Emergent teams Inertial teams
Social Networks Methods in Health Services Research 37
8/11/2008
Identifying Teams and Sub-teams
Structural Individual Team membership perceptions
Groups of interdependent actors
38
Realist
(Social Fact)
ACT teams
Top management teams, Clinics
Nominalist
(Researcher Defined)
8/11/2008
Social Networks Methods in Health Services Research
Using Interdependence to Identify Teams
Organizational theory and complexity theory: “Organizing at the edge of chaos”
Tim Carroll and Richard M. Burton. 2001. "Organizations and Complexity: Searching for the Edge of Chaos." Computational & Mathematical Organization Theory 6:319-337. David Krackhardt, 1994, "Constraints on the Interactive Organization as an Ideal Type." In Charles Heckscher & Anne Donnellan (eds.), The Post-Bureaucratic Organization. Beverly Hills, CA: Sage, p. 211-222. Work with means that you and the other person are interdependent during work – your actions affect each other directly. It does not refer to the situation where you and the other person contribute separately to clinic performance while you work in the same location. Please answer these questions for all others with whom you work (e.g., physicians, rooming nurses, staff, laboratory technicians, and receptionists). Patient care means all direct patient care and patient care related activities (e.g., ordering tests, calling in prescriptions, talking with patients on phone).
Social Networks Methods in Health Services Research 39
Who do you work with to provide patient care?
8/11/2008
Identifying Teams Using Structural Equivalence
(g) are together embedded in an encompassing organizational system, with boundaries and linkages to the broader system context and task environment.
Groups based on structurally equivalence: “Two nodes are said to be exactly structurally equivalent if they have the same relationships to all other nodes.”
At least two actors Dense interdependence Occupational heterogeneity Informal consulting (c) possess one or more common goals; (d) are brought together to perform organizationally relevant tasks;
(a) two or more individuals who
(e) exhibit interdependencies with respect to workflow, goals, and outcomes;
(f) have different roles and responsibilities;
(b) socially interact (face-to-face or, increasingly, virtually)
Inferred
8/11/2008
Social Networks Methods in Health Services Research
40
A Work With Sociomatrix: Pediatrics Clinic
MD08 MD08 0 MD21 0 MD25 0 MD54 0 MD64 0 MD82 0 N00 3 N08 1 N21 2 N25 2 N54 3 N71 0 N79 3 N82 2
MD21
MD25 MD54
0
0 0
0
2 0
2
0 0
0
0 0
0
0 0
2
3 0
3
3 1
2
1 3
2
2 0
3
3 0
3
3 0
0
0 0
3
3 0
2
2 1
MD64
MD82
0
0
0
2
0
3
0
0
0
0
0
0
3
2
2
2
0
1
2
2
2
2
0
0
2
2
2
2
N00
N08 N21
3
1 2
3
2 2
3
1 2
1
3 0
3
2 0
2
2 1
0
1 1
1
0 2
1
2 0
1
2 1
3
2 2
2
2 1
2
2 1
2
1 1
N25
N54 N71 N79 N82
2
3 0 3 2
8/11/2008
3
3 0 3 2
3
3 0 3 2
0
0 0 0 1
2
2 0 2 2
2
2 0 2 2
1
3 2 2 2
2
2 2 2 1
1
2 1 1 1
0
2 1 2 1
2
0 3 2 2
1
3 0 3 2
2
2 3 0 2
41
1
2 2 2 0
Social Networks Methods in Health Services Research
Graphing the Full Sociomatrix: Pediatrics Clinic
8/11/2008
Social Networks Methods in Health Services Research
42
Structurally Equivalent Actors: Pediatrics Clinic
Occupational Grouping
2N 6 MDs, 1 N
4N
1N
8/11/2008
Social Networks Methods in Health Services Research
43
Graphing the Full Sociomatrix: Family Practice Clinic
8/11/2008
Social Networks Methods in Health Services Research
44
Structurally Equivalent Actors: Family Practice Clinic
2 MD, 1 ST
Teams
2 MD, 2 N 2 MD, 5 N
4 ST
2 MD, 1 N,
2 ST, 1 T N, 3 ST, T 2 ST
8/11/2008 Social Networks Methods in Health Services Research 45
Predictive Validity for Interdependence Method for Identifying Teams
Informal consulting
is more likely to occur within teams than between teams Is greater in more cross-functional and interdependent teams Pediatrics – pool staffing Family practice – practice teams
Teams consistent with clinic staffing rules
Consistent with organizational and complexity theory
Organizing at the edge of chaos
Social Networks Methods in Health Services Research 46
8/11/2008
Usefulness of Methodology
Where are sub-teams in
Chronic care teams? Primary care practices? Assertive community treatment teams?
How are sub-teams connected?
Is it too many? Process loss and perturbations? Is it too few? So loosely coupled system is not integrated.
Social Networks Methods in Health Services Research 47
8/11/2008
Social Networks Research Challenges (And Benefits)
Observing relations rather than actors Defining network boundaries Defining relations to measure
Friendship? Respect? Work-with? Helping? Level of detail
Obtaining high response rates
Social Networks Methods in Health Services Research 48
8/11/2008
Managerial Benefits of a Social Network Approach
Managing relations is a key function of management
Relations can be managed with relatively simple interventions (staffing) Accurate understanding of network structure, which supports effective management
Social Networks Methods in Health Services Research 49
Measuring networks supports
8/11/2008
Research Benefits of a Network Approach
Evaluate interventions designed to integrate care Understand the black box between formal and informal structure
8/11/2008
Paths (or lack of paths) for information flow Disconnects Too many connections
Social Networks Methods in Health Services Research 50