Evidence
What we know, and what we don't.
Every figure on this website with its source, its core statement and its limits. Laid out openly, including where the findings are inconvenient — that's the difference from a talk that only needs to sound good.
Most figures here come from observational studies. They show links, not cause and effect. Where a finding argues against my own case, it still appears here.
Why the topic is pressing
- +9 %Solid
In the first half of 2026, mental illness became the most common cause of absence for the first time — 9 % more than the year before.
DAK/IGES, Krankenstand 1. Halbjahr 2026 - 28.5 working daysSolid
That's the average length of an absence caused by mental illness. Days lost have risen by 43 % in ten years.
WIdO, Fehlzeiten-Bilanz 2025 und Fehlzeiten-Report 2025 - More than halfSolid
More than half of employees aged 50 and over want to retire early. Among those in poorer health it is 60 %.
DAK-Gesundheitsreport 2026 - 10 % · 77 %Solid
Only 10 % are strongly engaged, 77 % do just what's required. Only 21 % feel fully supported by their manager.
Gallup Engagement Index Deutschland 2025 - 74 %Solid
74 % rate their chances on the job market as good or very good.
Gallup Engagement Index Deutschland 2025
What it comes down to
- Almost twice the riskOR 1.93Solid
Low room to act goes together with almost twice the risk of coronary heart disease. The amount of work, by contrast, was not associated with the risk.
Bosma et al., BMJ 1997 (n = 10.308)LimitsAssociation from an observational study, not proof of cause. What is measured is perceived control, not self-efficacy.
- The middle carries the mostSolid
The middle of the hierarchy shows higher rates of depression and anxiety than either the top or the base.
Prins et al., Sociology of Health & Illness 2015 (n = 21.859)LimitsAssociation from an observational study, not proof of cause.
- 27 % and 54 % lowerRR 0.73 · RR 0.46Solid
Good leadership goes together with a 27 % lower absence rate and 54 % fewer disability pensions.
Kuoppala et al., JOEM 2008 (Metaanalyse)LimitsAssociation, not proof of cause. The meta-analysis finds NO link between good leadership and higher performance — that must never be claimed.
30 minutes, no agenda. We find out whether and how I can help.
Book an intro call →What helps against it
- 136 samplesSolid
Psychological safety is linked to performance, learning and engagement in the team.
Frazier et al., Personnel Psychology 2017 (136 Stichproben)LimitsA link, not a statement of effect.
- £24bn of £51bnWith context only
The largest cost of poor mental health is not absence. It is being present without being able to perform.
Deloitte UK, Mental health and employers 2024LimitsConsultancy study from the UK, not transferable to Germany. Use with context only.
- Small effects that fadeWith context only
Resilience training shows small effects that fade over time. That argues for ongoing support rather than a one-off measure.
Vanhove et al. 2016 (Metaanalyse Resilienztrainings)LimitsUse with context only.
Figures I deliberately don't use
These figures circulate in talks and proposals. They are either misquoted, unverifiable or long refuted. That's why they don't appear here:
- 70 % aller Transformationen scheitern
- Gallup: 10 Billionen US-Dollar
- bis zu 142 Mrd. € Folgekosten
- Happiness Pie 50/40/10
- Amygdala Hijack
- 149.000 fehlende IT-Fachkräfte
- Gute Führung steigert die Leistung
- Senkt Krankentage um X %
Measure, don't assume.
10 questions · about 2 minutes · no sign-up · in German