Is Mental Health First Aid Evidence-Based? Unpacking the Research and Outcomes

Is mental health first aid evidence-based?

What the research supports, where uncertainty remains, and how employers can evaluate effectiveness claims.

Compare the research findings

Guidance checked 18 October 2026

A group of learners at a Skills Training Group training session

Is mental health first aid evidence-based?

Mental Health First Aid has a research base, but the strength of the evidence depends on the outcome. Studies support improvements in some trainee outcomes, such as knowledge and confidence. Evidence that training improves the mental health of the people receiving help remains much less certain.

It is useful to ask “Evidence for what?” A participant feeling better prepared, actually providing appropriate help and a recipient experiencing better health are three different outcomes. One does not automatically prove the next.

This guide was checked on 18 October 2026. It includes a September 2026 review alongside the broader 2023 Cochrane review. A newer study in a narrower population does not replace a wider review or resolve every question.

What do the major reviews tell us?

2018

Knowledge, stigma and helping behaviour

Morgan, Ross and Reavley’s systematic review and meta-analysis included 18 controlled trials and 5,936 participants. It found improvements in knowledge, confidence and intentions to help, with smaller changes in stigma and the amount of help reported.

The quality of help provided was less clear. The review did not establish convincing benefits for recipients’ mental health. Its search ended in 2017, so it is useful background rather than a complete account of current research.

2023

Mental health and wellbeing outcomes

Cochrane’s review included 21 randomised studies involving 22,604 participants. At six to twelve months, the evidence was too uncertain to draw firm conclusions about effects on mental health and wellbeing compared with no intervention.

Only three studies, with 3,939 participants, contributed to that main comparison—not all 22,604 participants. Differences between the studies and concerns about their methods made the findings less certain. The review found no published data on possible harms from training; that does not show that harms cannot occur. Its searches ran to June 2023 and concerned MHFA-trademarked courses.

2026

A newer review of healthcare students

Yeoh and colleagues’ review, published on 1 September 2026, examined 26 studies involving healthcare professional students. Searches were updated in February 2026. Knowledge and confidence showed the most consistent improvements; findings on stigma were mixed. Peer support, self-care and student wellbeing were examined less often.

Many studies were small and lacked control groups or long-term follow-up. These findings concern healthcare students, not every workplace or course. The review adds recent evidence about learning outcomes without proving that training improves recipients’ health across the wider population.

These reviews overlap in their subject matter and may include some of the same studies. Their study and participant counts should not be added together as though they were three independent trials.

How to read common effectiveness claims

Match the claim to the outcome measured
A claim you may seeWhat you need to check
“Learners know more”Was knowledge tested, compared with another group and checked again later?
“Learners feel confident”Was this self-reported confidence or an independent assessment of what the person could do?
“People receive better support”Did researchers assess the quality of actual help, including the recipient’s experience, or only intentions to help?
“Mental health improves”Whose health was measured, over what period, and how did it compare with a suitable control group?
“The course saves money”Is there a relevant economic evaluation with clear costs and assumptions, rather than a confidence score presented as a financial return?

This is a guide to evaluating a claim, not a new analysis of the studies. A positive course review or satisfaction survey can be useful feedback, but cannot answer every question in the table.

Does the research apply to every course?

No. Check the programme studied, its content, participants and delivery. A result from one named curriculum should not be presented as proof for every course using the words “mental health first aid”. Qualification status and evidence of health benefits answer different questions.

Our Level 3 course leads to a Qualsafe qualification on successful assessment. The reviews above should not be read as direct evaluations of our particular course or as a guarantee of clinical outcomes.

Likewise, research on adult workplace or healthcare-student training should not automatically be used to make claims about children, schools or every type of first-aid training.

What does this mean for an employer choosing training?

Choose a clear purpose and examine how training will fit the organisation. HSE’s guidance places mental health trained first aiders among several possible support arrangements following a needs assessment. It is not an instruction to treat a training course as the whole solution.

  • Define the role. Agree who first aiders support, their availability, boundaries and escalation arrangements.
  • Check the training. Compare the actual syllabus, assessment, trainer support and requirements for keeping skills current.
  • Make help accessible. Keep service information current and explain how staff can find support, including remote workers.
  • Review implementation. Ask whether staff know how to access help and whether first aiders understand their role. Do not ask them to build informal clinical case files to prove impact.
  • Be careful about attribution. A change in absence or staff wellbeing after training does not, by itself, show that training caused it.

Our policy guide helps turn these questions into workplace arrangements. The evidence supports a considered decision about training, with realistic expectations and appropriate support around the people who complete it.