Related Experiment Videos
Unit-Level Burnout Dimensions and Patient Experience: A Multilevel Study of Aggregated Contextual Strain in a
Edona Haxhija1,2, Drita Kruja3
1European University of Tirana Tirana Albania uet.edu.al.
Purpose:
To examine whether aggregated unit-level nurse burnout is associated with patient-reported experience across hospital units using a multilevel framework, conceptualizing burnout as aggregated contextual strain rather than homogeneous unit climate.
Methods:
A cross-sectional multilevel study was conducted across 11 regional public hospitals in Albania (October 2024-July 2025). Nurse burnout was assessed using the Maslach Burnout Inventory among 2000 nurses, and patient experience was measured with HCAHPS among 679 patients. Burnout scores were aggregated at the clinical-unit level. Multilevel linear modeling with patients nested within units examined associations between emotional exhaustion, depersonalization, personal accomplishment, and patient experience. Cluster-robust standard errors and leave-one-unit-out sensitivity analyses were conducted to assess estimate stability.
Results:
Patient experience varied significantly between units (ICC = 0.229). In the multilevel model, higher unit-level emotional exhaustion and depersonalization were both associated with lower patient experience. Under cluster-robust estimation (t-distribution, df = 9), depersonalization remained statistically significant (robust p < 0.001), and emotional exhaustion remained significant (robust p = 0.028), whereas personal accomplishment was not significant (robust p = 0.187). Leave-one-cluster-out and exact wild-cluster-bootstrap sensitivity analyses confirmed depersonalization as the most robust association across all specifications (bootstrap p = 0.003), whereas emotional exhaustion, while directionally consistent, was only borderline under the most conservative correction (exact wild-bootstrap p = 0.052), underscoring the statistical fragility inherent to a 10-cluster design.
Conclusion:
Aggregated unit-level depersonalization, and to a lesser and more fragile extent emotional exhaustion, may be relevant to patient-centered care, but the strength and consistency of these unit-level estimates are constrained by the small number of analytical clusters and by heterogeneity in how those clusters were formed across hospitals. Given the modest number of clinical units, findings should be interpreted as exploratory contextual associations. These findings support further investigation of workload support and team-level resources as potential targets for improving nurse well-being and patient-centered care.
