The Evidence Base: Research Behind BridgeCalm's Approach

BridgeCalm is built on published clinical research — not marketing claims. This page compiles the studies that inform our product design, from the therapeutic frameworks Jan uses to the outcome tracking in the therapist portal.

Every citation below links to a published, peer-reviewed source. Where a study has limitations, we note them. Where the evidence is mixed, we say so.

Between-session practice and homework compliance

The clinical foundation of BridgeCalm is the evidence that what patients do between therapy sessions meaningfully affects outcomes. These studies inform why the app exists.

Kazantzis, N., et al. (2016). "The Processes of Cognitive Behavioral Therapy: A Review of Meta-Analyses." Cognitive Therapy and Research. PMC5388291

Meta-analysis of CBT process variables found homework compliance to be a significant predictor of outcome (weighted effect size r = .26). Quality of homework engagement predicted outcomes at follow-up (g = 1.07) more strongly than quantity (g = 0.51). This is the basis for BridgeCalm's focus on guided practice quality over repetition volume.

Ryum, T. & Kazantzis, N. (2024). "Homework in Cognitive Behavioral Therapy: A Systematic Review of Adherence Assessment and Its Relationship to Outcome." Clinical Psychology in Europe. PMC11303922

The first causal analysis of homework effects on outcomes, using instrumental variable approaches. Found a medium effect size (d = .53) for homework on treatment outcomes, establishing that homework causes improvement rather than merely correlating with it.

Mausbach, B.T., et al. (2010). "The Relationship Between Homework Compliance and Therapy Outcomes: An Updated Meta-Analysis." Cognitive Therapy and Research. PMC2990801

Meta-analysis of 23 studies (n = 2,183) found a significant positive relationship between homework compliance and treatment outcomes (r = .26). Effect held across diagnostic categories and treatment types.

Tang, W. & Kreindler, D. (2017). "Supporting the Homework Experience in CBT: Essential Features of Mobile Apps." JMIR Mental Health. PMC5481663

Systematic review identifying six essential features for apps that support between-session therapeutic practice: psychoeducation, homework assignment, progress monitoring, guided therapy, motivational messages, and clinician support. BridgeCalm's feature set was designed to address all six.

Schleider, J.L., et al. (2023). "Barriers to Homework Adherence in Therapy: A Systematic Review." Clinical Psychology Review. PMC10844770

Systematic review of 25 studies identifying barriers to homework completion: forgetfulness, time constraints, lack of understanding, emotional avoidance, and low perceived relevance. These findings informed Jan's task delivery system, difficulty reporting, and embedded exercise guides.

Cognitive Behavioral Therapy (CBT)

These studies support the exercises Jan delivers through "The Coach" conversation style.

Hofmann, S.G., et al. (2012). "The Efficacy of Cognitive Behavioral Therapy: A Review of Meta-Analyses." Cognitive Therapy and Research. PMC3584580

Comprehensive review establishing CBT as the most extensively studied form of psychotherapy, with strong evidence across anxiety disorders, depression, eating disorders, insomnia, and chronic pain. Effect sizes ranged from medium to large depending on the condition.

Topooco, N., et al. (2017). "Internet-Delivered CBT for Depression: A Systematic Review and Individual Patient Data Meta-Analysis." Journal of Medical Internet Research. PMC5764800

Established that internet-delivered CBT with therapist support produces large effect sizes (g = .78–.79) for depression treatment, validating the digital delivery model for CBT interventions.

Kaczkurkin, A.N. & Foa, E.B. (2015). "Cognitive-Behavioral Therapy for Anxiety Disorders." Dialogues in Clinical Neuroscience. PMC4610618

Review of CBT techniques for anxiety including cognitive restructuring, behavioral activation, and exposure-based approaches. Provides the clinical framework for Jan's thought record exercises and cognitive reframing prompts.

Dialectical Behavior Therapy (DBT)

These studies support the exercises Jan delivers through "The Anchor" conversation style.

Ramos Vieira, B.C., et al. (2023). "Breathing Practices for Stress and Anxiety Reduction: Systematic Review of Randomized Controlled Trials." Brain Sciences. PMC10741869

Systematic review of 58 studies with 72 breathing interventions. Found that 54 of 72 interventions were effective at reducing stress and anxiety. Supports paced breathing as a core component of Jan's distress tolerance exercises.

Schumer, M.C., et al. (2019). "Brief Mindfulness Interventions: A Systematic Review and Meta-Analysis." Mindfulness. Springer

Meta-analysis of 85 studies on brief mindfulness practices found significant positive effects on anxiety (g = 0.42), depression (g = 0.36), emotion regulation, and cognitive function — even for practices under 15 minutes. Supports the feasibility of BridgeCalm's micro-practice approach.

Acceptance and Commitment Therapy (ACT)

These studies support the exercises Jan delivers through "The Explorer" conversation style.

Association for Contextual Behavioral Science. "Six Core Processes of ACT." contextualscience.org

Primary scientific organization behind ACT, describing the six interconnected processes — acceptance, cognitive defusion, being present, self-as-context, values, and committed action — that constitute psychological flexibility. Jan's Explorer style guides users through all six.

A-Tjak, J.G.L., et al. (2015). "A Meta-Analysis of the Efficacy of Acceptance and Commitment Therapy." Behaviour Research and Therapy. DOI: 10.1016/j.brat.2014.07.010

Meta-analysis of 39 RCTs finding ACT superior to control conditions for anxiety, depression, addiction, and chronic pain, with medium effect sizes. ACT was not significantly different from established treatments (CBT), supporting it as an evidence-based alternative.

Mood tracking and self-monitoring

These studies inform BridgeCalm's daily check-in system and mood tracking features.

Brattland, H., et al. (2024). "Routine Outcome Monitoring and Clinical Feedback: Recent Advances." Nature Partner Journals. PMC11076375

Systematic review finding that routine outcome monitoring improves treatment outcomes, detects deteriorating cases that clinical judgment alone misses, and facilitates shared decision-making. This is the primary evidence behind BridgeCalm's therapist portal data model.

JMIR Mental Health. (2026). "Mood Monitoring Interventions: Systematic Review and Meta-Analysis." mental.jmir.org

Meta-analysis of 8 RCTs (n = 1,230) found "no robust evidence" that mood monitoring alone reduces depression symptoms. Small effect observed at 12 months but not at 6 months. This study informed BridgeCalm's design decision to pair mood tracking with guided exercises and therapist connection rather than treating it as a standalone intervention. We cite this study because intellectual honesty requires acknowledging where the evidence is limited.

Aguilera, A. (2020). "Digital Micro Interventions for Behavioral and Mental Health Gains." JMIR. PMC7661243

Review of brief digital interventions finding that daily mood check-ins serve as effective entry points for more engaged wellness behaviors. Median app retention after 15 days was only 3.9%, establishing that low-friction design is essential for sustained engagement.

Measurement-based care (MBC)

These studies support the PHQ-9/GAD-7 tracking and outcomes features in BridgeCalm's therapist portal.

Guo, T., et al. (2015). "Implementing Measurement-Based Care for Depression." Psychiatric Services. PMC7813452

24-week RCT comparing MBC to standard care in depression treatment. Found 87% response rate with MBC vs. 63% in standard care, 74% remission rate vs. 29%, and treatment response time halved (5.6 weeks vs. 11.6 weeks). This is the most commonly cited MBC efficacy trial and a core justification for BridgeCalm's outcomes tracking.

Scott, K. & Lewis, C.C. (2015). "Using Measurement-Based Care to Enhance Any Treatment." Cognitive and Behavioral Practice. PMC4910387

Foundational review establishing that MBC improves outcomes across diagnostic groups — depression, anxiety, psychosis, substance use — by providing data-driven feedback. The study positions MBC as a transdiagnostic practice enhancement rather than a standalone treatment.

Frontiers in Health Services. (2025). "Impact of Measurement-Based Care at Scale." frontiersin.org

Implementation study finding 23.5% relative improvement in combined PHQ-9/GAD-7 outcomes. Pre-treatment PHQ-9 scores averaged 16.04; post-treatment dropped to 11.21 (effect size d = 1.14). MBC patients were 22% more likely to discharge due to successful treatment completion.

Psychiatric Services. "Survey of Behavioral Health Providers on MBC Use." psychiatryonline.org

Survey finding fewer than 20% of behavioral health providers use MBC consistently, with only 5% following an evidence-based assessment schedule. Identifies administrative burden as the primary barrier — the problem BridgeCalm's automated assessment workflow is designed to solve.

AI in mental health: safety and limitations

These studies inform BridgeCalm's safety architecture and regulatory positioning.

Brown University. (2025). "AI chatbots systematically violate mental health ethics standards." brown.edu

Research finding that general-purpose AI chatbots systematically fail at crisis navigation and ethical boundaries in mental health contexts. Informed BridgeCalm's decision to hard-code crisis routing to 988 and Crisis Text Line rather than attempting AI-mediated crisis management.

American Psychological Association. "Health Advisory: Chatbots and Wellness Apps." apa.org

APA's formal advisory calling for safeguards around AI chatbot interactions, particularly for users in crisis. Supports BridgeCalm's classification as a wellness companion (not a therapist) with mandatory professional crisis routing.

Stanford HAI. "Exploring the Dangers of AI in Mental Health Care." hai.stanford.edu

Found that AI tools constrained by expert medical knowledge perform significantly better than unconstrained models. Supports BridgeCalm's approach of grounding Jan in specific, named therapeutic frameworks rather than open-ended generative responses.

NIST. (2023). "AI Risk Management Framework 1.0." nist.gov

Voluntary federal framework for responsible AI development with four core functions: Govern, Map, Measure, Manage. BridgeCalm's development process follows this structure for AI safety documentation and risk monitoring.

Regulatory landscape

These sources define the legal boundaries within which BridgeCalm operates.

FDA. "Digital Health Center of Excellence." fda.gov

FDA guidance distinguishing general wellness products from Software as a Medical Device (SaMD). BridgeCalm is positioned as a general wellness product that promotes skill-building and self-monitoring without claiming to treat specific disorders.

Illinois IDFPR. (2025). "Legislation Prohibiting AI Therapy in Illinois (HB 1806)." idfpr.illinois.gov

Illinois law prohibiting AI from providing therapy, making diagnoses, or creating treatment plans. Penalties up to $10,000 per violation. BridgeCalm is designed to operate within these boundaries — Jan provides guided practice, not therapy.

California Legislature. (2026). "SB 243 — Companion Chatbot Safeguards." leginfo.legislature.ca.gov

California law requiring AI chatbots to prevent harmful content exposure for minors, notify users they're talking to AI, and include suicide/self-harm crisis protocols. BridgeCalm's crisis detection and routing system was designed with these requirements in mind.

HHS. "HIPAA and Mental Health." hhs.gov

HHS guidance on HIPAA protections specific to mental health information. BridgeCalm maintains HIPAA compliance with end-to-end encryption, available Business Associate Agreements, and no third-party data sharing.

Key study: AI chatbots for depression

Maier, A., et al. (2025). "Therabot AI Chatbot RCT." NEJM AI. ai.nejm.org

Randomized controlled trial (n = 210) at Dartmouth comparing AI chatbot (Therabot) plus self-help to self-help alone. At 8 weeks, 51% of the chatbot group showed clinically meaningful depression improvement (≥50% PHQ-8 reduction) vs. 19.6% in control. The chatbot group also showed significant improvements in anxiety and well-being.

This study informs BridgeCalm's core thesis that AI-guided practice between sessions can meaningfully improve outcomes. Important note: the study tested a conversational AI chatbot providing CBT and other therapeutic techniques — not passive mood tracking alone. BridgeCalm's design reflects this distinction.

What this means for BridgeCalm

Every feature in BridgeCalm maps to a published finding:

The daily check-in exists because self-monitoring is an established gateway to behavioral change — when paired with guided action and therapist connection (Aguilera, 2020; Brattland, 2024).

Jan's conversation styles exist because CBT, DBT, and ACT have extensive evidence bases, and digital delivery of these techniques produces meaningful effect sizes (Hofmann, 2012; Topooco, 2017; A-Tjak, 2015).

The homework system exists because between-session practice is a significant predictor of treatment outcomes, and quality of engagement matters more than quantity (Kazantzis, 2016; Ryum, 2024).

The therapist portal exists because measurement-based care and routine outcome monitoring improve outcomes across diagnostic groups — but fewer than 20% of therapists use them due to administrative burden (Guo, 2015; Psychiatric Services survey).

The safety boundaries exist because the evidence is clear that AI tools require structured constraints, professional crisis routing, and clinical oversight to operate responsibly in mental health contexts (Brown, 2025; APA advisory; Stanford HAI).

Where the evidence is limited — particularly around mood tracking as a standalone intervention — we design accordingly, pairing tracking with action rather than treating it as sufficient on its own.

[Explore the therapist portal →]

This page is updated as new research is published. Last reviewed: March 2026.