Research & Insights for Workplace Wellbeing

This study does not test a new wellbeing programme. Instead, it tackles a harder question: how should we properly evaluate organisational-level interventions aimed at reducing sickness absence and turnover?

Using a large Swedish public-sector healthcare intervention, the authors combine mixed-effects modelling, time-series analysis, process evaluation, and reference group comparisons to separate real intervention effects from seasonal trends, time trends, and wider organisational change.

They found that accounting for seasonality and time trends significantly alters results. When using this rigorous approach, the intervention was associated with an overall reduction in total sickness absence and, with time lags applied, reductions in employee turnover. However, effects varied substantially between intervention groups, highlighting the importance of context and implementation fidelity.

Because evaluation quality determines whether we can trust results.

• Sickness absence and turnover fluctuate naturally over time and season

• Without correcting for trends, effects can be misinterpreted

• Organisational interventions rarely produce uniform results

• Context and implementation fidelity strongly influence outcomes

• Delayed effects (e.g., turnover) must be considered in analysis

• Comparing against reference groups helps rule out false positives

• Small percentage shifts in long-term outcomes can still be practically meaningful

For HR and workplace wellbeing leaders, this paper reinforces an important message: strong design and rigorous evaluation matter just as much as the intervention itself.