Mode
Text Size
Log in / Sign up

AI-powered CBT reduces anxiety symptoms in 6,284 patients across NHS primary care over 24 weeksYour Wait for Anxiety Therapy Could Predict Your Success Online

AI-generated summary of the cited source, checked by automated accuracy review. How we work

Key Takeaway
Consider AI-CBT as a scalable bridge intervention, noting heterogeneous response and a deprivation-related gap in symptom reduction.

This prospective cohort study examined anxiety symptom trajectories in 6,284 patients aged 18 to 65 referred to an AI-powered CBT platform (CalmLogic) across 187 general practices in four NHS England Integrated Care Systems. The study followed participants for 24 weeks to assess changes in GAD-7 scores and identify predictors of treatment response within routine NHS settings.

Participants demonstrated a significant average linear decline in GAD-7 scores, with an effect size of -0.94 points per month (p < .001). Analysis identified four distinct trajectory classes; membership in the 'Rapid Responder' class was predicted by higher baseline severity, female gender, and greater module completion, representing 28.4% of the cohort.

Practice-level factors also influenced outcomes. Faster improvement trajectories were independently predicted by shorter IAPT wait times (coefficient = -0.31, p = .003). Conversely, patients in the most deprived quintile exhibited slower symptom reduction trajectories (coefficient = 0.22, p = .011). Safety data, including adverse events, discontinuations, and tolerability, were not reported in this study.

The authors note that longitudinal evidence on anxiety symptom trajectories and their predictors in routine NHS settings remains limited. Consequently, the platform's greatest incremental value appears in capacity-constrained areas where IAPT wait times exceed 90 days, suggesting its role as a scalable bridge intervention rather than a definitive cure for all patients.

Anxiety disorders are incredibly common. Millions seek help through the NHS. But demand far outstrips supply.

The average wait for face-to-face cognitive behavioral therapy now exceeds 90 days in many areas. That’s three months of struggling.

To bridge this gap, the NHS has introduced digital therapy platforms. They offer guided CBT exercises online or through an app. It’s a vital stopgap. But do they work equally well for everyone?

Until now, we haven’t had a clear picture.

The Surprising Shift

The old assumption was simple: digital therapy helps some, but not all. This new research moves past that.

It didn’t just ask if people improved. It tracked how they improved over six months. Scientists discovered people don’t follow one path. They follow four.

This changes how we view digital treatment. Success isn't a simple yes or no. It’s a journey with different routes.

Think of anxiety like a loud, persistent alarm system going off too easily. CBT helps you recalibrate that system.

It teaches you to identify and challenge fear-driven thoughts. You learn new ways to respond. An AI-powered platform guides you through these steps with interactive modules.

It’s like having a coach in your pocket. The AI isn’t a therapist. It’s a structured program that delivers proven techniques.

A Snapshot of the Study

Researchers analyzed data from 6,284 adults across 187 GP practices. All had been referred to an AI-CBT platform for anxiety.

They tracked their anxiety scores over 24 weeks. Using advanced modeling, the study could see individual paths. It also checked what factors—from a patient’s age to their clinic’s location—influenced progress.

The Four Paths to Recovery

The data revealed four distinct stories:

1. Rapid Responders (28%): These users felt better quickly. Their anxiety dropped steeply within the first two months and stayed down. 2. Gradual Improvers (34%): This largest group saw a steady, consistent decline in anxiety all the way through the six-month period. 3. Partial Responders (23%): They got some early relief, but then progress stalled. They were left with clinically significant anxiety. 4. Non-Responders (15%): This group saw little to no change. A few even felt slightly worse.

On average, anxiety scores fell significantly. But the "average" hides these very different experiences.

The Capacity Paradox

Here’s a finding that changes the game.

The study found that people from areas with the longest waits for traditional therapy (over 90 days) improved faster on the digital platform.

This is powerful. It means digital therapy isn't just a placeholder. In areas where the system is most strained, it becomes a highly effective lifeline.

Patients there may be more motivated to engage. The digital option feels less like a compromise and more like a necessary solution.

But There's a Catch

A stark inequality emerged, unrelated to how hard people tried.

Patients living in the most deprived communities improved more slowly. This was true even when they completed the same number of therapy modules as people in wealthier areas.

Something else is at play. It could be related to stress, environment, or other barriers that digital tools alone can’t address. This is a deprivation-related treatment gap.

This study, published on the health sciences archive medRxiv, provides crucial real-world evidence. It shows digital therapy is a potent tool for system-wide capacity problems. But experts warn its one-size-fits-all delivery may widen health inequalities if not carefully managed.

The value is clear in overstretched areas. The challenge is making it work for everyone.

If you are on a waitlist for therapy, a recommended digital CBT platform is a valid and potentially very helpful option. This data strongly supports trying it.

Engage with it consistently. The study linked completing more modules with a better chance of being a "Rapid Responder."

However, be patient with your own progress. Your path might be gradual. If you feel you're in the "Partial" or "Non-Responder" groups, it’s a critical sign to circle back with your GP. You may need a different approach.

Understanding the Limits

This study observed people using a service in the real world. It doesn’t prove the app caused the improvement, though the design strongly suggests it. The platform studied was one specific AI-CBT system. Results might vary with other apps.

Most importantly, it highlights a problem—the deprivation gap—but doesn’t yet solve it.

The next steps are critical. NHS services must use these insights to identify patients on slower trajectories early and offer extra support. Developers need to adapt platforms for greater effectiveness in deprived communities.

This isn't about replacing therapists. It's about using smart tools to help more people faster, while relentlessly working to ensure no one is left behind because of their postcode. The future of mental health care is a blended one—and it must be a fair one.

Study Details

Study typeCohort
Sample sizen = 6,284
EvidenceLevel 3
PublishedMar 2026
View Original Abstract ↓
BackgroundThe NHS Improving Access to Psychological Therapies (IAPT) programme, now rebranded as NHS Talking Therapies, faces persistent capacity constraints with average wait times exceeding 90 days for cognitive behavioral therapy (CBT) in many Clinical Commissioning Group areas. AI-powered CBT platforms have been introduced as a digital adjunct within stepped care, yet longitudinal evidence on anxiety symptom trajectories and their predictors in routine NHS settings remains limited. ObjectiveTo model individual anxiety symptom trajectories among patients referred to an AI-powered CBT platform within NHS primary care, identify distinct trajectory classes, and examine patient-level and practice-level predictors of differential treatment response using multilevel growth curve modeling. MethodsA prospective cohort study was conducted using linked clinical and administrative data from 6,284 patients (aged 18-65) referred to the CalmLogic AI-CBT platform across 187 general practices in four NHS England Integrated Care Systems (ICSs) between April 2023 and September 2025. Patients completed GAD-7 assessments at baseline, 4 weeks, 8 weeks, 12 weeks, and 24 weeks. Three-level growth curve models (assessments nested within patients nested within practices) with random intercepts and random slopes were fitted. Growth mixture modeling (GMM) was subsequently applied to identify latent trajectory classes. Predictors were examined at Level 2 (patient demographics, baseline severity, comorbidities, digital literacy, engagement intensity) and Level 3 (practice deprivation index, list size, urban/rural classification, and IAPT wait time). ResultsThe unconditional growth model revealed a significant average linear decline in GAD-7 scores of -0.94 points per month (p < .001), with substantial between-patient variation in both intercepts (variance = 14.82, p < .001) and slopes (variance = 0.38, p < .001). Significant between-practice variation accounted for 8.7% of intercept variance (ICC = 0.087). Growth mixture modeling identified four distinct trajectory classes: Rapid Responders (28.4%, steep early decline stabilising by week 8); Gradual Improvers (34.1%, steady linear decline through 24 weeks); Partial Responders (22.8%, modest early improvement followed by a plateau at clinically significant levels); and Non-Responders (14.7%, minimal change or slight deterioration). Higher baseline severity, female gender, and greater module completion predicted membership in the Rapid Responder class. Practice-level IAPT wait times exceeding 90 days independently predicted faster improvement trajectories (coefficient = -0.31, p = .003), suggesting that AI-CBT has its greatest incremental value in capacity-constrained areas. Patients in the most deprived quintile showed slower trajectories (coefficient = 0.22, p = .011) despite equivalent engagement levels, indicating a deprivation-related treatment response gap. ConclusionsAI-powered CBT platforms integrated within NHS primary care produce significant anxiety symptom reduction on average, but treatment response is heterogeneous, with four distinct trajectory classes identified. The finding that longer IAPT wait times predict better AI-CBT outcomes supports the platforms positioning as a scalable bridge intervention for capacity-constrained services. The deprivation-related response gap warrants targeted support strategies for patients in the most disadvantaged communities.
Free Newsletter

Clinical research that matters. Delivered to your inbox.

Join thousands of clinicians and researchers. No spam, unsubscribe anytime.