Telework, engagement and burnout: heterogeneous effects from a doubly robust study

Background

Whether telework is beneficial or harmful depends on who is working remotely and at what intensity. This study applies doubly robust causal inference to estimate subgroup treatment effects of telework on burnout and work engagement across individual and work-related characteristics.

Methods

Cross-sectional data from 4,088 workers aged 18–65 were drawn from the TELEHEALTH study. Telework exposure was operationalized as four cutpoints relative to no telework: 1–5, 5–10, 10–15, and >15 days/month. Outcomes were burnout (BAT) and work engagement (UWES). Subgroup effects were estimated using augmented inverse probability weighting with five-fold cross-fitting and SuperLearner ensemble models under multiple imputation. Heterogeneity was examined across sex, education, household composition, parental status, job position, work situation, supervisor role, physical limitation, and multiple job-holding. A joint standardized score classified subgroup-by-exposure combinations as beneficial, harmful, or mixed.

Results

Telework effects were heterogeneous across subgroups and exposures. Workers in larger households (≥2 persons) showed the strongest beneficial effects at > 15 days/month (joint score: 0.69, p < 0.01), higher education (0.47, p < 0.01) and no physical limitations (0.35, p < 0.05). Self-employed workers showed the largest harmful effects at high intensities (joint score: 1.01, p < 0.01), with reduced engagement and elevated burnout. Workers living with one other person showed consistently adverse effects (engagement ATE: −0.49 at 10–15d, p < 0.05; burnout ATE: 0.19 at 5–10d, p < 0.05; joint score: −0.67, p < 0.01). Workers with low-to-medium education experienced harmful effects (−0.29, p < 0.05).

Conclusions

These findings challenge the notion that telework is uniformly beneficial or harmful. Effects vary by household structure, employment type, and education. Policies should account for these differences rather than applying uniform arrangements across employees.

Key messages

• Telework at high intensity benefits workers in larger households and the highly educated, but harms self-employed workers and those in small households.

• Effective telework policy requires tailoring to household composition, employment status, and education, which determine whether remote work is beneficial or harmful.