Mental health applications number in the tens of thousands. The proportion with published evidence of effectiveness is very small, and there is no signal in an app store that distinguishes them.
What has evidence
Structured programmes based on cognitive behavioural therapy, delivered as a course with exercises and progress tracking. These have been tested in trials and show effects for mild to moderate anxiety and depression – smaller than in-person therapy, and real.
The key feature is structure. A defined programme worked through over weeks, not a collection of tools to browse.
What has less
Mood tracking alone. Recording how you feel is not an intervention, though it can support one.
Meditation apps, which have evidence for stress and much weaker evidence for clinical conditions – and are frequently marketed as though the two are the same.
Chatbot companions, where the research is early, the enthusiasm is high, and the failure modes in a crisis are not well characterised.
The engagement problem
Most people stop using these apps within days. Trial results come from participants who continued, which makes real-world effectiveness considerably lower than trial effectiveness.
This is the largest gap between the research and the reality and it is rarely mentioned.
What to look for
Whether the app cites specific trials of itself, not of the general approach. Whether it says clearly what it is not for. Whether there is a defined programme rather than an assortment of features. And what happens to your data.
The important limit
None of these is a substitute for care in a serious situation. If someone is in crisis, an app is the wrong tool, and any app that suggests otherwise should be deleted.