Do Mental Health Apps Work? What the Research Says

There are thousands of mental health apps promising calmer days, better sleep and fewer anxious thoughts. Some are built on decades of therapy research and tested in clinical trials. Others are a breathing timer with a subscription fee and a marketing claim nobody has checked.

So do they work? The honest answer is: some do, modestly, for some people, if they keep using them. This guide summarizes what published research says by app type, explains why engagement is the real bottleneck and gives you a checklist for judging an app’s clinical claims and privacy practices.

This is general information, not medical advice. If you are in crisis or thinking about harming yourself, call or text 988, the Suicide and Crisis Lifeline, or call 911.

The short answer from the research

The largest recent summary of the evidence is a 2024 meta-analysis in World Psychiatry by Linardon and colleagues, which pooled 176 randomized controlled trials of smartphone apps for depression and anxiety.

Key findings:

• Depression: Apps produced a small but statistically significant improvement in symptoms compared with control groups (an effect size of g = 0.28).
• Generalized anxiety: A similar small effect (g = 0.26).
• What worked better: For depression, effects were larger in apps that used cognitive behavioral therapy (CBT) features or chatbot technology. For anxiety, effects were larger when trials targeted generalized anxiety and used CBT or mood-monitoring apps.
• Weaker evidence: Results for social anxiety, obsessive-compulsive symptoms and PTSD were less certain, because many of those trials were small or at high risk of bias.

What does “small” mean in practice? The authors estimated that roughly 11 to 12 people would need to use an app for 1 additional person to see a meaningful benefit compared with control conditions. That is a real effect for a low-cost, always-available tool. It is also a long way from the dramatic results many app store pages imply.

Key takeaway: The research supports mental health apps as a helpful add-on for mild to moderate depression and anxiety, especially CBT-based ones. It does not support them as a replacement for professional care, and the app you choose matters far more than the category.

What the evidence says by app type

Not all mental health apps do the same thing. Here is how the main types compare, based on the published research.

Prescription apps are a separate category

A few apps go through FDA review as medical devices. In April 2024, the FDA cleared Rejoyn from Otsuka and Click Therapeutics, described by the companies as the first prescription digital therapeutic for treating major depressive disorder symptoms. It is a 6-week program for adults 22 and older who are taking an antidepressant, used as an add-on to clinician-managed care.

Most consumer wellness apps are not FDA-reviewed. FDA policy treats many low-risk apps, such as those that help people with diagnosed conditions practice coping skills, as an area where it does not intend to enforce device requirements. That means an app can talk about anxiety or depression without the FDA ever checking its claims.

What about AI chatbots?

Interest in AI “therapists” has exploded. In November 2025, the FDA’s Digital Health Advisory Committee met specifically to discuss generative AI-enabled digital mental health devices, including premarket evidence and postmarket monitoring. Purpose-built chatbots designed and tested for mental health are different from general-purpose chatbots, which were not designed as treatment. Our guide on the risks of AI chatbots for emotional support covers the safety concerns in detail.

The engagement problem: why most people quit

Clinical trials measure what happens when people use an app as designed. Real life looks different.

A 2019 study in the Journal of Medical Internet Research analyzed real-world usage data for 93 popular Android mental health apps. The median 30-day retention rate was just 3.3%, meaning that for a typical app, only about 3 in 100 people who installed it were still opening it a month later.

Within clinical trials, the 2024 meta-analysis found dropout rates of about 24%, similar to face-to-face therapy. The gap between trial engagement and real-world engagement is one of the biggest reasons app results in daily life can fall short of published findings.

What helps people stick with an app:

• A clear goal, such as completing an 8-week CBT program rather than “feeling better”
• Short, scheduled sessions tied to an existing routine
• Some human connection, such as a therapist, coach or friend who checks in
• Reminders you set, rather than constant promotional notifications

How to judge an app’s clinical claims

App store pages are marketing. Use these questions to see whether claims hold up. They draw on the American Psychiatric Association’s App Evaluation Model, which looks at access, privacy and security, clinical foundation, engagement and data sharing with your care team.

1. Is there published research on this specific app? Look for randomized controlled trials in peer-reviewed journals, not just “based on science” or “developed with psychologists.” Research on CBT in general does not prove a particular app works.

2. Who was studied, and compared with what? A trial that compares an app with doing nothing tends to show bigger effects than one that compares it with an active alternative. The 2024 meta-analysis found larger effects against inactive controls, which the authors suggest may partly reflect a “digital placebo” effect.

3. Does the app use an established approach? CBT, behavioral activation, acceptance and commitment therapy and mindfulness have research behind them. Vague claims about “rewiring your brain” are a warning sign.

4. Is it FDA-authorized, and for what? If an app claims to treat a condition, check whether it actually has FDA clearance and what it was cleared for.

5. What does it do in a crisis? A responsible app clearly directs users to 988 or emergency services and does not try to handle a crisis on its own.

Privacy: what happens to your mental health data

Mental health data is some of the most sensitive information you can share. Many consumer apps are not covered by HIPAA, which generally applies only when an app is provided by a health care provider, health plan or their business associate.

The Federal Trade Commission has acted against health apps that broke privacy promises. In 2023, the FTC finalized an order requiring the online counseling service BetterHelp to pay $7.8 million and banning it from sharing consumers’ health data for advertising, after alleging it shared users’ information, including health questionnaire answers, with platforms such as Facebook and Snapchat. The FTC’s updated Health Breach Notification Rule, effective July 29, 2024, also covers many health apps outside HIPAA. Some states have added their own consumer health data laws; our US state data privacy laws guide explains the landscape. This is not legal advice.

Before you sign up:

• Read the privacy policy for whether data is sold or shared for advertising.
• Check whether you can use the app without linking social media accounts.
• Look for a way to export and delete your data.
• Turn off ad tracking permissions on your phone.

Where to start: choosing and using an app

Watch the cost. Many apps use free trials that convert to annual subscriptions. Set a reminder before the trial ends, and see our guide on how to find and cancel subscriptions if you forget.

If you want human support with digital convenience, online therapy connects you with licensed professionals, and some clinicians use apps between sessions as “homework.” For physical health signals like sleep and activity that can affect mood, see our comparison of smartwatches and fitness trackers.

When an app is not enough

Apps are a reasonable first step for mild symptoms, everyday stress or building skills. Reach out to a professional instead of, or in addition to, an app when:

• Symptoms last more than 2 weeks and affect work, school or relationships
• You have thoughts of death, suicide or self-harm
• You are using alcohol or drugs to cope
• You have symptoms such as hallucinations, extreme mood swings or panic attacks that feel unmanageable
• The app makes you feel worse, more anxious or more isolated

If you are thinking about suicide or are worried about someone else, call or text 988 any time, or call 911 in an emergency. Asking an AI tool about symptoms is no substitute; our article on whether to ask an AI chatbot about your symptoms explains why.

Conclusion

Mental health apps work, but modestly, and only when they are well designed and actually used. CBT-based apps have the best evidence for mild to moderate depression and anxiety, mindfulness apps can help with stress, and FDA-cleared prescription apps are a separate category used under clinical care. Pick an app with published research and a clear privacy policy, give it a fair trial and measure your progress. If it is not enough, that is a signal to reach out for professional support, not a personal failure.

Sources

• Linardon et al., World Psychiatry (2024): meta-analysis of 176 randomized controlled trials — effect sizes by app type, dropout and control-group findings
• Baumel et al., Journal of Medical Internet Research (2019): Objective user engagement with mental health apps — real-world 30-day retention
• FDA: Examples of software functions for which the FDA will exercise enforcement discretion — low-risk mental health app policy
• FDA: November 6, 2025 Digital Health Advisory Committee meeting — generative AI mental health devices
• American Psychiatric Association: The App Evaluation Model — framework for judging apps
• FTC: Final order against BetterHelp — health data sharing enforcement