Schools often treat teacher well-being as a benefit to add after the “real” work of improving instruction. A new trial from Uganda tests a more consequential possibility: supporting teachers may be part of how a school delivers mental-health support in the first place.
The cluster-randomized trial, published November 6 in JAMA Network Open, involved 47 urban and rural primary schools and 509 teachers. Schools—not individual teachers—were assigned to one of three groups: professional development focused on evidence-based classroom and family-engagement practices; the same program plus a teacher-wellness component; or a control condition that received no intervention during the first year.
The results are promising, but narrower than a simple “teacher wellness works” headline. The clearest evidence concerns what teachers said they knew and how well they said they managed stress. The study did not measure children’s mental health in this report, and its design cannot tell us whether the wellness lessons themselves caused the apparent advantage.
The extra support changed more than the curriculum
The core program was substantial: three days of training followed by eight one-hour coaching sessions over two to three months. It covered practices such as proactive behavior management, nonharsh discipline, emotion socialization, family engagement and predictable classroom environments. Teacher Training College staff and other non–mental-health professionals delivered it through existing educational structures—a deliberate test of whether the model could function without a large specialist workforce.
The enhanced version added a one-day workshop and three group-support sessions. These focused on recognizing stress and practicing regulation and coping strategies. The registered trial was designed with Ugandan government, university and community partners, building on a previously published implementation protocol.
That addition coincided with greater engagement. Among teachers offered the core program, 64.1% attended all three training days; among those offered the enhanced program, 80.9% attended all four. The enhanced group also attended an average of 6.09 coaching sessions, compared with 4.58 in the core group. Attendance is not a mental-health outcome, but it matters: an intervention cannot change practice if an overloaded workforce cannot stay in the room.
The pattern suggests a useful shift in perspective. Well-being support may not merely be an outcome for staff. It may also help people engage with the training intended to benefit children. That is an interpretation, not a mechanism proven by the trial; the enhanced group received more contact time and content, so several explanations remain possible.
The strongest findings were specific—not universal
Immediately after the intervention, teachers completed measures covering knowledge and use of evidence-based practices, stress and stress management. Compared with controls, the enhanced group showed a medium-sized advantage in knowledge (standardized mean difference 0.57) and stress management (0.50). Those two comparisons remained statistically significant after the researchers adjusted for multiple tests.
Other estimates leaned in the expected direction but were less secure. Reported use of evidence-based practices favored the enhanced program by 0.36 standard deviations, while the stress composite favored it by 0.27 standard deviations. Yet the confidence interval for stress reached zero, and neither comparison remained significant after the multiplicity correction. The ordinary professional-development group showed a smaller knowledge advantage over control, but that result also did not survive the adjustment.
This distinction is easy to lose. The trial supports a short-term change in teachers’ self-reported knowledge and stress-management capacity more confidently than it supports reduced distress or changed classroom behavior. A standardized effect describes separation between group averages; it does not reveal how many individual teachers experienced a meaningful improvement.
What the trial cannot yet tell us about children
Most outcomes were self-reported, which leaves room for social-desirability and expectancy effects. Teachers knew whether they had received training, while the control group received neither an active comparison program nor equivalent attention. Roughly 20% to 27% of post-intervention observations were missing, although the researchers included all randomized teachers through multiple imputation and ran sensitivity analyses.
The most important gap is also the most tempting leap: this article reports no child-level mental-health outcomes. Teachers described more knowledge and stronger stress management; the study did not show that pupils became less anxious, less aggressive or more emotionally secure. Longer-term teacher and child analyses are needed before the program can be called preventive child mental-health care.
Nor can the trial isolate the wellness component. The enhanced group received an extra workshop and support sessions, so content, attention, peer contact and total dose were bundled together. Results from Uganda’s particular educational system may not transfer unchanged to wealthier schools or to places with different workloads and training infrastructure.
Still, the study asks the right systems-level question. In schools where specialist mental-health care is scarce, expanding support does not always begin by importing more specialists. It may begin by strengthening the adults already present—and by recognizing that their capacity to help children depends partly on whether the institution also helps them carry the work.
