📄 social_science

Social Network Factors Associated with the Intention to Use Digital Mental Health Interventions Among AANHPI Emerging Adults: Application and Integration of the Network Episode Model

This study of 1,577 AANHPI emerging adults demonstrates that social network factors, particularly awareness of peers' digital mental health intervention use and current therapy engagement, significantly predict the intention to use such interventions and interact with public stigma, thereby supporting an integrated framework that combines individual and network-level strategies to address mental health disparities.

Original authors: John Bunyi, Shinyi Wu, Angel Hsing-Chi Hwang, Hans Oh

Published 2026-06-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: John Bunyi, Shinyi Wu, Angel Hsing-Chi Hwang, Hans Oh

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: Why Don't People Ask for Help?

Imagine you are feeling down, like a car with a "Check Engine" light on. You know you need a mechanic, but you're scared to go to the shop. Maybe you think people will judge you for having a broken car, or maybe you just don't know which mechanic is good.

This is exactly what happens with Asian American, Native Hawaiian, and Pacific Islander (AANHPI) young adults and mental health. Even though many are struggling, they rarely go to traditional doctors or therapists. Instead, researchers are asking: Would they use a "digital mechanic" (apps, websites, online therapy) if it were available?

This study tries to figure out what makes these young people decide to click "Yes" on a digital mental health app.

The Two Lenses: The "Solo Driver" vs. The "Traffic Jam"

The researchers used two different maps to understand this decision:

The "Solo Driver" Map (Individual Factors): This looks at the person alone. Do they feel sick? Do they know what mental health is? Are they afraid of being judged?
- The Paper's Finding: Yes, these things matter. If you feel a need for help, know what the problem is, or are worried about what others think, you are more likely to want to use an app.
The "Traffic Jam" Map (Social Network Factors): This looks at the people around you. Who are your friends? What are they doing? Do they talk about mental health?
- The Paper's Finding: This is the new part of the study. It turns out that who you know and what they do matters just as much as how you feel inside.

The Key Discoveries (The "Aha!" Moments)

1. The "I Saw Them Do It" Effect (Social Modeling)

Imagine you are hesitant to try a new, weird restaurant. But then, you see your best friend eating there and saying, "It's great!" Suddenly, you feel brave enough to try it.

The study found that if a young person knows a friend or family member who has used a digital mental health app, they are much more likely to want to use one too. It's like seeing a friend successfully navigate a traffic jam; it gives you the confidence to drive your own car.

The Catch: If they don't know anyone who has used it, they are less likely to try. But here is the twist...

2. The "Secret Club" Effect (Stigma and Uncertainty)

This is the most interesting part. The researchers found a strange connection between fear of judgment (stigma) and not knowing if your friends use these apps.

Scenario A: You know your friends use apps. You feel safe using one too.
Scenario B: You are terrified of being judged for having mental health issues, AND you have no idea if your friends use apps (you are "uncertain").
- The Result: In Scenario B, the fear of judgment actually pushes people toward the apps. Why? Because apps are private. If you are scared your friends or family will find out you're struggling, a digital app is like a secret tunnel. It lets you get help without anyone seeing you enter the building.

3. Size Doesn't Matter (Network Size vs. Behavior)

You might think, "If I have a huge group of 50 friends, I'm more likely to get help."

The Paper's Finding: Not necessarily. Having a big network of friends or family didn't make a difference.
What mattered: It wasn't about the number of people; it was about the behavior of the people. Knowing that someone in your circle actually used the tool was the key. It's not about having a big crowd; it's about having a friend who actually bought the ticket.

4. The "Symptom" Myth

Usually, we think: "The more sick I am, the more I need help."

The Paper's Finding: Surprisingly, how "severe" the symptoms were (how bad the depression or anxiety was) did not predict if they wanted to use an app.
The Analogy: It's like having a flat tire. Whether the tire is slightly low or completely flat doesn't matter as much as whether you know you need a spare tire and whether you have a friend who told you where to get one. The "need" and the "social signal" matter more than the actual size of the problem.

The Bottom Line

This study tells us that for AANHPI young adults, deciding to use a mental health app isn't just a solo decision made in a vacuum. It's a social dance.

If they see friends doing it, they do it.
If they are scared of being judged and don't know what their friends are doing, they use the app as a private escape.
Having a big family or many friends doesn't help unless those people are actually talking about or using these tools.

The researchers conclude that to get more people to use these digital tools, we can't just tell them "you are sick, get help." We need to change the social environment—get friends and families talking about it, and show them that using these tools is normal, safe, and something their peers are already doing.

Technical Summary: Social Network Factors Associated with the Intention to Use Digital Mental Health Interventions Among AANHPI Emerging Adults

Problem Statement
Asian American, Native Hawaiian, and Pacific Islander (AANHPI) emerging adults (ages 18–29) experience mental illness rates comparable to or exceeding the general U.S. population yet consistently underutilize formal mental health services. While Digital Mental Health Interventions (DMHIs) are proposed as a solution to barriers such as stigma and lack of culturally responsive providers, the drivers of the intention to use DMHIs within this highly collectivist population remain poorly understood. Previous research, utilizing the Seeking Mental Health Care (SMHC) model, established that individual-level factors (perceived need, mental health literacy, and stigma) predict DMHI intention. However, these models treat help-seeking as a discrete individual decision, neglecting the social context. This study addresses the gap in understanding how social network characteristics influence DMHI intention and whether they moderate established individual-level associations, specifically within the AANHPI demographic.

Methodology
The study employed a cross-sectional survey design using data from the Personalized Normative Feedback (PNF) Project, collected via Qualtrics panels between July and September 2024. The analytic sample consisted of 1,577 self-identified AANHPI emerging adults (ages 18–29) residing in the United States, recruited to ensure representation across seven major ethnic subgroups (Chinese, Filipino, Indian, Japanese, Korean, Native Hawaiian/Pacific Islander, Vietnamese) plus "Self-Described" and multi-ethnic/racial categories.

The analysis integrated two theoretical frameworks:

Seeking Mental Health Care (SMHC) Model: Measuring individual-level predictors including perceived need for support, mental health literacy (MHL), perceived public stigma, and self-stigma.
Network Episode Model (NEM): Measuring social network characteristics, including awareness of friends/family DMHI use, structural network size (Lubben Social Network Scale-6 for family and friends), perceived social support availability (Medical Outcomes Study Social Support Survey), and current formal therapy engagement.

Statistical analysis utilized nested multivariable logistic regression in Stata 19. The modeling proceeded in three stages:

Model 1: Replicated the SMHC-only specification.
Model 2: Added NEM variables to assess their independent contribution.
Models 3a–3c: Tested theoretically motivated interaction terms (Perceived Need × Family Network Size; MHL × Friends/Family DMHI Use; Public Stigma × Friends/Family DMHI Use). A Bonferroni-corrected significance threshold ( $\alpha = .017$ ) was applied to the interaction tests. Covariates included sociodemographics, symptom severity (categorized by PHQ-9 and GAD-7 scores), and vendor source.

Key Results

Model Fit and Network Variables: Adding NEM variables to the SMHC model (Model 2) significantly improved model fit ( $\Delta \chi^2 = 39.96, p < .001$ ).
Significant Network Predictors: Awareness that friends or family use DMHIs was strongly associated with DMHI intention (AOR = 1.79, $p < .001$ ). Current engagement in formal therapy was also a strong predictor (AOR = 1.84, $p < .001$ ).
Non-Significant Network Predictors: Structural network size (family and friend subscales) and perceived social support availability did not reach statistical significance in predicting DMHI intention.
Individual-Level Robustness: All SMHC predictors (perceived need, MHL, public stigma, self-stigma) remained significant after the inclusion of network variables. Symptom severity remained non-significant across all models.
Interaction Effects: Only one interaction term reached significance after Bonferroni correction: Friends/Family DMHI Use × Public Stigma. Specifically, among participants who were uncertain ("not known") whether their network used DMHIs, higher public stigma was associated with substantially greater DMHI intention (AOR = 3.12, $p = .003$ ). This effect was not observed for those who knew their network used DMHIs.
Ethnic Heterogeneity: Significant differences persisted across ethnic subgroups, with Filipino, Indian, Japanese, and Self-Described AANHPI participants showing higher odds of DMHI intention compared to Chinese participants.

Key Contributions
The paper's primary contribution is the empirical integration of the SMHC and NEM frameworks to demonstrate that DMHI intention is not solely an individual cognitive process but is significantly shaped by social network dynamics.

Social Modeling: The study confirms that "experiential homophily" (knowing others who use DMHIs) acts as a powerful social modeling mechanism, increasing the likelihood of intention to use digital tools.
Refinement of Network Constructs: The findings suggest that the content and behavior of a network (e.g., knowing someone uses a tool) are more predictive of DMHI intention than the structure of the network (e.g., network size) or the general availability of social support.
Stigma Dynamics: The significant interaction reveals a nuanced relationship where public stigma suppresses network conversations about digital help-seeking. For those unaware of peer usage, high public stigma paradoxically increases intention, likely driving individuals toward private, anonymous digital channels to avoid social exposure.
Symptom Severity: Consistent with prior work, symptom severity alone does not predict DMHI intention; rather, the social and attitudinal processing of symptoms (perceived need, literacy, stigma) is the critical antecedent.

Significance and Claims
The authors claim that this study provides a more comprehensive framework for understanding digital help-seeking by moving beyond individual-level factors to include social negotiation and network influence. The findings underscore that interventions aiming to increase DMHI engagement in AANHPI communities must extend beyond individual education to include network-level strategies. Specifically, the authors suggest that:

Peer and family education programs could leverage social modeling to normalize DMHI use.
Interventions for individuals lacking peer exposure to DMHIs should address the specific dynamic where public stigma drives reliance on private digital tools.
Mental health literacy programs should focus on recognizing the need for support rather than solely targeting symptom severity, as the latter is not a direct predictor of help-seeking intention in this context.

The paper concludes that while the cross-sectional design limits causal inference, the integration of individual and network models offers a robust theoretical basis for future research and the development of culturally tailored, network-informed digital mental health interventions for AANHPI emerging adults.