Subgroup analysis, broadly, aims to measure change within and between groups. Subgroups can be characterized by variables that are easier to define such as gender or those less well defined such as risk status. Capturing change related to treatment impacts within and between groups pervades prevention and intervention science. The challenges in designing research to examine subgroups, the analysis of the data and the interpretation of the results are critically important to answer the question that comes up most regularly in policy-relevant research: What works for whom?
Given the recent attention to evidence-based policy decision making and limited federal funds, answering the question of what works for whom becomes critically important for prevention and intervention science. Subgroup analyses can inform the field on how to maximize treatment and steer resources with informed decision-making. In September 2009 the Administration for Children and Families, along with other federal partners (Office of the Assistant Secretary for Planning and Evaluation, Centers for Disease Control’s Division of Violence Prevention, Institute for Education Sciences, National Institute for Drug Abuse and the National Institute for Mental Health), convened a meeting of experts to discuss the state of the field. Papers and presentations at the meeting centered on innovative methods for conducting subgroup analysis, and a discussion of guidelines for interpretation and reporting of subgroup analyses in prevention and intervention research.
8:40-9:00
Welcome and introductions
Naomi Goldstein, Administration for Children and Families
9:00-10:45
Moderator
Cheryl Boyce, National Institute on Drug Abuse
Overview on subgroups from the perspective of medical and public health research
Rui Wang, Harvard UniversityOverview on subgroups from the perspective of educational and social policy research
Jacob Klerman, Abt Associates
11:00-12:45
Moderator
Liz Ginexi, National Institute on Drug Abuse
Foundational methods: Foundational methods for subgroup effects and key design considerations (statistical power, sampling, statistical vs. clinical significance, multiple comparisons)
Hendricks Brown, University of South FloridaSimultaneous mediation and moderation tests: Examining whether a program is effective for subgroups (including testing for multiple moderators simultaneously; impact of measurement for multiple error testing subgroup effects)
David MacKinnon, Arizona State University
1:45-4:15
Moderator
Amy Madigan, Office of the Assistant Secretary for Planning and Evaluation
Subgroup analysis from a Bayesian perspective: Multi-level modeling approaches and random effects framework to deal with multiple comparisons
Jennifer Hill, New York UniversitySubgroup analysis through meta-analysis (including: meta-analytic techniques (e.g., subgroup definitions differ between studies; coherence in intervention approaches; quality of studies)
Michael Borenstein, BiostatSubgroup analysis from a latent class analysis approach
Stephanie Lanza, Pennsylvania State University
Brendan Kelly, Administration for Children and Families
9:15-10:00
Moderator
Tamara Haegerich, CDC
Causal effect moderation (modification) when treatment or exposure is time-varying
Daniel Almirall, Duke UniversityQuantile treatment effects for testing subgroup impacts
Marianne Bitler, University of California, Irvine, NBER
12:45-1:30
Moderator
Jennifer Brooks, Administration for Children and Families
Post-hoc subgroups (including: Subgroups that emerge as a result of unforeseen changes in design/sampling (i.e., half of the centers have a high percentage of ELL children); populations found post-hoc to be more or less responsive to treatment and modeling who those individuals are)
Chris Price, Abt Associates
Lauren Supplee, Administration for Children and Families
Prevention Science
Volume 14, Issue 2, April 2013
Special Issue: Subgroup Analysis in Prevention and Intervention Science
Abstracts available on http://link.springer.com/journal/11121/14/2
Introduction to the Special Issue: Subgroup Analysis in Prevention and Intervention Research
Lauren H. Supplee, Brendan C. Kelly, David M. MacKinnon, Meryl Yoches Barofsky
Pages 107-110Detecting Moderator Effects Using Subgroup Analyses
Rui Wang, James H. Ware
Pages 111-120Methodological Challenges Examining Subgroup Differences: Examples from Universal School-based Youth Violence Prevention Trials
Albert D. Farrell, David B. Henry, Amie Bettencourt
Pages 121-133Meta-Analysis and Subgroups
Michael Borenstein, Julian P. T. Higgins
Pages 134-143Methods for Synthesizing Findings on Moderation Effects Across Multiple Randomized Trials
C. Hendricks Brown, Zili Sloboda, Fabrizio Faggiano, Brent Teasdale, Ferdinand Keller, Gregor Burkhart, Federica Vigna-Taglianti, George Howe, Katherine Masyn, Wei Wang, Bengt Muthén, Peggy Stephens, Scott Grey, Tatiana Perrino, Prevention Science and Methodology Group
Pages 144-156Latent Class Analysis: An Alternative Perspective on Subgroup Analysis in Prevention and Treatment
Stephanie T. Lanza, Brittany L. Rhoades
Pages 157-168Subgroups Analysis when Treatment and Moderators are Time-varying
Daniel Almirall, Daniel F. McCaffrey, Rajeev Ramchand, Susan A. Murphy
Pages 169-178When is the Story in the Subgroups?
Howard S. Bloom, Charles Michalopoulos
Pages 179-188Exploring Connections Between Moderators and Mediators: Commentary on Subgroup Analyses in Intervention Research
Alexander J. Rothman
Pages 189-192Commentary on Subgroup Analysis in Intervention Research: Opportunities for the Public Health Approach to Violence Prevention
Tamara M. Haegerich, Greta M. Massetti
Pages 193-198Erratum to: Subgroups Analysis when Treatment and Moderators are Time-varying
Daniel Almirall, Daniel F. McCaffrey, Rajeev Ramchand, Susan A. Murphy
Page 199