
Hybrid Experimental Designs (HEDs) are transforming the way researchers optimize the integration and adaptation of human-delivered and digital intervention components. Since human-delivered components and digital components can be adapted on different timescales, slow (for example monthly, for human-delivered components) and fast (for example daily, for digital components), HEDs involve sequential (repeated) randomizations on multiple timescales.
Three studies led by d3center affiliates show these designs in action. Sylvie Naar, PhD, is using the HED to develop an intervention for HIV and alcohol use among young adults. Rebecca Sripada, PhD, and Stefanie T. LoSavio, PhD, use the HED to develop an intervention for PTSD among Veterans. Simon Goldberg, PhD, is using the HED to develop a meditation intervention for promoting mental health. Together, these projects showcase the versatility and impact of HEDs in addressing critical health and behavioral challenges.
Optimizing HIV and Alcohol Interventions for Young Adults
Project SUSTAIN, led by Sylvie Naar, PhD, and funded by an NIAAA P01 grant, examines a multimodal adaptive intervention (MADI) to improve HIV and alcohol management among Young People Living with HIV (YPLWH). The study employs a HED with randomizations on two timescales:
Slow timescale randomizations: After participants receive an initial web-based self-management intervention (HC-WB) targeting medication adherence and alcohol use, they undergo randomization at weeks 2, 4, and 6 to step up to tele-health coaching (HC-TB) delivered by trained staff or continue HC-WB.
Fast timescale randomizations: Simultaneously, participants undergo daily randomization to receive a youth-developed mobile device prompt or no prompt over 8 weeks.
The study aims to evaluate both the optimal timing of HC-TB implementation and the impact of different prompt types on daily medication adherence, stress, and alcohol use outcomes.
Changing the Treatment Course for Veterans with PTSD
Rebecca Sripada, PhD and Stefanie T. LoSavio, PhD are leading a HED with 302 Veterans with PTSD, funded by the Department of Defense to improve the effectiveness and efficiency of PTSD treatment.
Slow timescale randomizations: Participants are first randomized to one of two first-line treatments—Prolonged Exposure or Cognitive Processing Therapy.
- After 6 sessions, Non-responders are re-randomized to one of two RESCUE strategies—either continuing with a modular version of the same treatment for sessions 7-14 or switching to the alternate treatment.
- After 6 sessions, Responders are re-randomized to one of two MAINTENANCE strategies—either continuing treatment or stepping down for the remaining 8 sessions.
Fast timescale randomization: Participants are randomized every day to either a daily digital prompt to enhance engagement with homework assignments, or no digital prompt.
Goals include evaluating whether adjusting treatments for non-responders improves outcomes and whether prompts enhance between-session homework completion, supporting long-term progress.
Improving Engagement and Mental Health Outcomes in a Meditation Intervention
Simon Goldberg, PhD, is exploring how to best integrate human and digital support to improve engagement and mental health outcomes in a meditation-based intervention among 41 participants. Participants were individuals who had recently downloaded the Healthy Minds Program app who were reporting elevated symptoms of depression, anxiety, or both.
Slow timescale randomization: Participants are first randomized into four groups using a 2×2 factorial design with two factors:
- Initial coaching session vs. initial informational session (control group)
- Access to email support vs. no access
Fast timescale randomization: Over 4 weeks, participants are randomized daily to a mobile-based prompt to encourage informal practice, a mobile-based prompt to encourage formal practice, or no prompt.
The study evaluates feasibility, acceptability, and engagement metrics, including app usability and engagement with formal and informal practice. Goldberg shares: “We were excited to deploy a hybrid experimental design to efficiently examine multiple research questions in the same sample. We hope that this work will ultimately clarify what elements of human and digital support are most impactful, and cost-effective, for improving outcomes in digital meditation interventions.”
LET’S STAY IN TOUCH
Join the d3center Mailing List
Keep up to date with the latest news, events, software releases, learning modules, and resources from the d3center.