Canonical text (the exact bytes fingerprinted)

TRUVACE RECORD VERSION
record: TRV-2026-1997
version: 1
kind: certified
reason: Certified into the record
timestamp: 2026-10-20T10:41:52.252974Z
status: published
lens: g_space
sector: labor
headline: Explainable AI for resource-aware intervention prioritization of digital burnout in healthcare workplaces
dek: Digital burnout is an emerging workforce risk in digitally intensive healthcare systems, yet existing studies remain largely descriptive and provide limited guidance for prioritizing modifiable risk factors. This study develops an explainable artificial intelligence framework for model-based prioritization of digital burnout drivers among healthcare professionals. Cross-sectional survey data were collected from Malaysian healthcare professionals (N = 400). Digital burnout was modelled as a continuous outcome usi…
gain_title: A gradient boosting model with SHAP explanations predicted digital burnout in 400 healthcare professionals and pointed to digital minimalism practices and technostress as the main levers for prioritizing interventions.
problem_title: (none)
trace_subject: (none)
gain_reading: A gradient boosting model with SHAP explanations predicted digital burnout in 400 healthcare professionals and pointed to digital minimalism practices and technostress as the main levers for prioritizing interventions.
gain_evidence: SHAP analysis identified digital minimalism practices and technostress as the dominant opposing predictors of digital burnout.
problem_reading: (none)
problem_evidence: (none)
quick_read: Researchers surveyed 400 Malaysian healthcare professionals and trained linear regression, random forest, and gradient boosting models to predict digital burnout from job demands, resources, technostress, and digital minimalism practices. Gradient boosting performed most stably, and SHAP analysis showed which predictors drove individual predictions.

The work matters because earlier digital burnout research was mostly descriptive, offering little help in choosing which risk factors to address first. The framework moves toward resource-aware decision support, but its scenario results are simulated and non-causal, so real-world benefit is not established until longitudinal or experimental testing is done.
limitation: The counterfactual burnout reductions are simulated, non-causal estimates from cross-sectional data and need longitudinal or experimental validation before being read as real intervention effects.
tag: Evidence-backed gain
key_points: The sample was cross-sectional survey data from 400 Malaysian healthcare professionals. | Gradient boosting reached a mean R2 of 0.653 in five-fold cross-validation and 0.660 across 100 repeated 80/20 splits. | Digital minimalism practices and technostress were the dominant opposing predictors of burnout. | Single-factor scenario rankings varied by individual, risk threshold, and perturbation size. | Combining digital minimalism enhancement with technostress reduction gave the largest model-predicted burnout reduction in every condition tested.
rundown: The study modelled digital burnout as a continuous outcome. Predictors included digital job demands, organizational and personal resources, technostress, and digital minimalism practices. Three model types were compared using train-test validation, cross-validation, and repeated resampling.

SHAP values were used to describe predictor contributions and the heterogeneity of predictions across individuals. Constrained counterfactual scenario analysis then estimated how predicted burnout would change under alternative adjustments to modifiable factors.

Because single-factor priorities differed across individuals and thresholds, the authors frame the approach as individualized, resource-aware prioritization. The combined scenario of enhancing digital minimalism and reducing technostress was the most effective in the model.
sources:
- peer_reviewed | Health Information Science and Systems | https://doi.org/10.1007/s13755-026-00485-4 | 2026-10-15
prev: 0000000000000000000000000000000000000000000000000000000000000000

sha256

1771dff35fb75eeac108716ee39ed983b754efd400d502568e138dff7dd4626f

previous

0000000000000000000000000000000000000000000000000000000000000000