Someone downloads a top-rated app hoping for therapy and later wonders where their entries go. The mental health app features that matter most are published evidence, a transparent privacy policy and clear limits on what the app replaces.
Use this checklist alongside the wider digital health guide. It separates what apps promise from what research and privacy terms actually support.
Evidence: popularity is not proof
A psychiatrist quoted in a 2024 article from Hartford HealthCare’s rehab network says most of the 10,000-20,000 mental health apps are untested, lack evidence that they work and are privacy risks. A 2017 study, described in KFF’s policy analysis, found that none of the most popular apps rated highly by consumers for anxiety symptoms were evidence-based treatments. A high rating therefore says little about whether an app helps.
In practice, effectiveness is judged by research. Hartford’s selection advice is to choose apps with research-based evidence, ideally published in a peer-reviewed journal. Treat that as the first filter. An app that points to a study you can look up is a stronger candidate than one that cites no research at all, however polished its design.
| Common assumption | What the evidence supports | Planning use |
|---|---|---|
| A high rating means it works | A 2017 study found none of the most popular consumer-rated anxiety apps were evidence-based treatments. | Look for peer-reviewed research on the specific app, not its star rating. |
| Any app that mentions therapy is a treatment | Some apps carry disclaimers that use does not replace care from a licensed provider. | Read the disclaimer to learn what the app itself claims to be. |
| A listed app is a maintained app | Updates within the past 180 days are a sign the app is well maintained. | Check when the app was last updated before relying on it. |
| A privacy policy is boilerplate | A transparent policy ideally promises not to collect, sell or transmit sensitive data. | Read it before you enter mood or symptom entries. |
The criteria come from Hartford HealthCare’s app-selection guidance, and the study and disclaimer findings from KFF.
Evidence is only half the decision, because an app that works but mishandles personal data is still a poor choice. Privacy is where the next set of misunderstandings sits.
Privacy: what a policy, HIPAA and updates tell you
Privacy varies from app to app, so a general assumption of safety is not enough. The policy wording, the maintenance record and the legal framework each answer a different part of the question.
KFF describes HIPAA’s reach more narrowly. Before the public health emergency, providers were required to ensure that audio-video technology and third-party apps used in telehealth sessions were HIPAA-compliant, and in 2020 the US Department of Health and Human Services exercised enforcement discretion over that technology during the pandemic. Ask any app that cites HIPAA how it applies to your data. Providers and product teams can find the operational side in Health Data Compliance: A Practical 2026 Playbook For Providers And Tech Teams – Bioconnected.
Several features indicate how seriously an app treats your information:
- Policy wording. Hartford recommends a transparent privacy policy that ideally promises not to collect, sell or transmit sensitive data. Vague language about sharing with partners is a reason to look elsewhere.
- Data not collected at all. The CBT-I app created by the U.S. Department of Veterans Affairs is described by Hartford as an app that does not collect your data, which is the strongest form of privacy a policy can offer.
- Stated safeguards. Hartford notes that WYSA emphasizes confidentiality and protects user data with a number of different safeguards. Look for apps that describe their protections rather than simply asserting trust.
- Security maintenance. A resource from the University of Illinois Chicago, citing Psyberguide, says the eCBT app has not been recently revised and users should use caution regarding security updates. An app with good content but stale updates carries a security question.
Privacy and evidence decide whether an app is worth trusting. The therapeutic features decide whether it is worth using.
Therapeutic features worth having
The content should follow a recognised therapy format, and the app should be clear about whether a licensed person is involved. Self-guided apps and live-session apps solve different problems, and mixing them up is a frequent source of disappointment.
Self-guided skills
Self-guided apps teach techniques you practise alone. The eCBT app provides skills and education based on cognitive behavioral therapy techniques, with versions for mood, anxiety and trauma, according to the UIC resource. WYSA coaches users through concerns using CBT and dialectical behavior therapy (DBT) techniques, and Hartford lists a free version with a premium plan at $11.99/month or $74.99/annually.
Narrower tools can also be useful. William James College describes Breathe2Relax, created by the National Center for Telehealth and Technology, as teaching diaphragmatic breathing. It describes Calm as offering people with stress and anxiety guided meditations, sleep stories, breathing programs and relaxing music, at $14.99/month or $69.99/year.
Self-guided apps suit skill practice and relaxation. They are not a substitute for a clinician when symptoms are severe, so check what the app itself says it replaces.
Live sessions with licensed providers
Some apps are really access points to teletherapy. KFF describes telehealth apps that offer therapy sessions with a licensed provider via live video or audio. Hartford describes NOCD as offering virtual sessions with a licensed therapist who specializes in Exposure and Response Prevention (ERP) therapy, with a subscription around $99/week depending on location and availability. That weekly price sits well above the self-guided subscriptions above, so it suits someone who needs condition-specific therapy rather than general stress relief.
Tracking and clinician reporting
Tracking is a feature that helps when a clinician is involved. Hartford describes a medication app that sends personalized reminders and tracks medication history, mood, symptoms and side effects for the provider. The value lies in the report your provider receives, not in the logging alone. Tracking records are no proof that the app improves outcomes, which is why the evidence check comes first.
Matching an app to the need
Choose by the problem rather than the brand. Practising breathing or meditation points to a self-guided tool such as Breathe2Relax or Calm. Learning CBT-based skills points to apps built on that therapy, such as eCBT or WYSA. A diagnosis that calls for ERP points to a live licensed therapist rather than a content library. Medication management points to a tracker that reports to your provider.
Treat conversational AI features as a separate question from any of these. The Hidden Dangers Of AI In Healthcare: What Clinicians And Patients Must Know In 2026 – Bioconnected sets out the risks clinicians and patients should weigh before relying on them.
Frequently asked questions
| Question | Answer |
|---|---|
| How is app effectiveness measured? | By published research, ideally in a peer-reviewed journal, according to Hartford HealthCare. Research on the WhatsMyM3 screening tool found it a valid, efficient and feasible one-page screen in primary care, as noted by UIC. That is the kind of validation to look for, though it concerns screening rather than treatment. |
| Can an app be prescribed? | Computerized behavioral therapy devices are subject to FDA approval and may need a prescription, according to KFF. They form a distinct category from general wellness apps. |
| Can an app use my entries for research? | Some do. The UIC resource says data from the Mood Mission app is used for a research study and that users are asked for information before unlocking missions. Read the consent wording before you enter anything. |
| Do I have to pay for a good app? | Not necessarily. William James College says many listed apps are free and the rest cost a small fee, and Breathe2Relax is free. Hartford says the VA’s CBT-I app is free. Live therapy apps cost more. |
