Positive communication for decreasing burnout in intensive‑care‑unit staff: a cluster‑randomized trial

Azoulay @ElieAzoulay5. Intensive Care Med, 2025; 51:2031-2041. doi.org/10.1007/s00134-025-08134-2

Clinical Question

• In ICU healthcare professionals, does a multicomponent positive communication intervention, compared with usual care, reduce the prevalence of occupational burnout?

Background

• Burnout prevalence in healthcare professionals in ICU is high
• Recent international reports have prioritised targeting burnout
• Observational studies suggest that better communication among staff reduces burnout
• Previous research has often focussed on individual interventions requiring considerable personal effort and time, such as mindfulness, coaching or yoga
• Long term positive effects have rarely been demonstrated

Design

• International cluster randomised controlled trial
• Randomised at the unit level in a 1:1 ratio
• Stratified on ICU size, country and mortality
• Baseline data collection on demographics of units, participating healthcare professionals, and burnout scores with Maslach Burnout Inventory (MBI)
• MBI has 3 subscales, for emotional exhaustion, deperonsalisation, and personal accomplishment
• Respondents with high emotional exhaustion (27+) and/or high depersonalisation (10+) scores were classified as having burnout
• Units randomised to intervention received training then implemented the intervention over a four week period
• Intention-to-treat analysis
• Outcomes assessed at individual HCP level with adjustment for clustering
• No blinding
• Primary outcome of burnout assessed with MBI score following intervention
• HCP responses were anonymous and therefore not paired over time
• Secondary outcomes included MBI sub-scores, visual analogue scales assessing work satisfaction, patient-centeredness of care, family-centeredness of care, ethical climate, intention to leave the ICU, and the feeling of work safety

Setting

• 370 units from 60 countries were randomised
• Baseline data were collected between Sept 1 and Oct 10 2024
• Following the four week intervention period, data were collected between Nov 15 and Dec 15 2024

Population

• Inclusion:
  • ESICM affiliated ICUs were invited
  • Units obtaining institutional review board (IRB) approval and completing baseline assessments were randomised
  • All HCPs working in these units were invited to participate
• Exclusion:
  • Healthcare professionals who did not complete the assessments or did not consent to the use of their data were excluded
• 679 centres expressed interest > 434 randomised > 370 in final intention-to-treat analysis
  • Comparing baseline ICU characteristics of intervention vs control
    • Number of centres enrolled 192 vs 178
    • University hospital 67% vs 68%
    • Median number of ICU beds 19 vs 23
    • ICU mortality 17% vs 19%
    • Similar geographical distribution across units
• Number of eligible HCPs not reported. 15,527 completed baseline assessment > 9568 completed follow up and were included in analysis (59% response rate)
  • Comparing baseline HCP characteristics of intervention vs control
    • Number of HCPs 4966 vs 4602
    • Median age 34 vs 36
    • Female % 68 vs 69
    • Median ICU experience 36 months vs 30 months
    • Role: nurse 60% vs 61%
    • Role: physician 30% vs 28%
    • Baseline burnout prevalence similar: 58.6% vs 60.5%

Intervention

• Six item intervention targeted at individual and system level factors:
  • Posters, including the word Hello in all country languages designed to encourage saying Hello
  • Weekly e-mail reminders about the importance of teamwork
  • Greetings during morning huddles to set a positive tone and enhance teamwork
  • A box in which staff placed positive messages for their colleagues
  • Role modelling by leaders to promote positive communication
  • Noticeboards posting positive messages, drawings etc. directed to all ICU HCPs
• Four week intervention period

Control

• Standard care
• Units were waitlisted to receive the intervention Jan 2025 following completion of the trial

Outcome

• Primary outcome: burnout prevalence, as assessed by the Maslach Burnout Inventory. Burnout significantly lower in the intervention group:
  • 52.2% vs 63.3%
  • n=2592 (52.2%; 95% CI 50.8–53.6) vs n=2888 (63.3%; 95% CI 61.9–64.7), adjusted OR 0.56 (0.46–0.68), p<0.0001
• Secondary outcomes:
• Favouring the intervention group
  • Change in prevalence of burnout between study initiation and completion (pre and post analysis)
    • 58.6% vs 52.2% and 60.5% vs 63.3%
    • approximated adjusted difference-in-difference -7% (95%CI -4 to -10%), p<0.001
  • MBI subscale scores
    • Emotional exhaustion
      • 22.8 ± 12.95 vs 25.5 ± 12.88
      • Adjusted mean difference -3.54 (95% CI -4.88 to -2.19), p<0.0001
    • Depersonalisation
      • 9.28 ± 6.94 vs 10.88 ± 7.26
      • Adjusted mean difference -1.79 (-2.43 to -1.14), p<0.0001
    • Personal accomplishment
      • 34.93 ± 7.50 vs 34.32 ± 7.50
      • Adjusted mean difference +1.15 (0.44–1.86), p=0.002
  • Visual analogue scales (scored 0-10)
    • All six VAS-based secondary outcomes: work satisfaction, patient-centred care, family-centred care, ethical climate, intention to leave the ICU, feeling of safety at work (post-hoc analysis), favoured the intervention
    • Differences were small but statistically significant
• No significant difference in number of colleagues HCPs felt in conflict with
• Exploratory/post-hoc analysis
  • Adherence with the intervention significantly interacted with the intervention effect on burnout prevalence (p=0.04), although no dose-response relationship was observed
  • Substantial heterogeneity in burnout risk demonstrated across centres and countries: median OR was approximately 2.0 for centre effects and 1.84 for country effects
  • Effectiveness of the intervention varied significantly by country and geographic region (p<0.001 for both)
  • No significant interaction between HCP job category and effectiveness (p=0.32)

Authors’ Conclusions

• A simple, inexpensive four-week unit-based program improved work climate and reduced burnout among ICU staff within one month across diverse settings. Results support implementing pragmatic low burden, team focused strategies to address occupational burnout in critical care.

Strengths

• Large, multicentre trial with good global coverage
• Addresses an important problem, and one which is lacking trials of this size
• Pragmatic and low-cost intervention with high scalability
• Pre-registered on ClinicalTrials.gov
• Pre-published protocol with no deviations reported
• Appropriate use of cluster randomisation at the ICU level, well matched to the intervention, which targeted unit culture and communication rather than individual behaviour

Weaknesses

• Unblinded design with self-reported outcomes, open to social desirability bias especially in a behavioural intervention focused on positive communication
• Only 59% of HCPs completed the post-intervention data collection
• Response rates are not reported by trial arm, so differential response bias cannot be excluded
• Responses were de-identified and therefore individual HCPs could not be followed from baseline to follow-up
• Baseline data are incompletely reported for the secondary outcomes, limiting assessment of baseline balance and causal inference for these endpoints
• Although substantial between-centre heterogeneity in burnout risk was demonstrated, no analysis was performed to determine whether centres with higher baseline burnout derived greater benefit from the intervention
• Subgroup and heterogeneity analyses were conducted at broad geographic levels (e.g. Northern Europe), which may limit applicability to individual countries or units
• While statistically significant at the population level, reductions in burnout scores were modest at the individual level
• Intervention and follow up periods were short
• Follow up is needed to determine whether improvements translate into objective outcomes, e.g. staff retention, absenteeism, patient outcomes etc. (6 month follow up data is coming)

The Bottom Line

• This brief, low-cost intervention promoting positive communication led to modest, short-term improvements in self-reported burnout scores and workplace climate
• While unlikely to cause harm, I look forward to seeing the longterm data as concern that improvements may not be sustained in the absence of structural improvements in staffing, workload, and organisational support. HELLO may help, but should not distract from the harder work required to address burnout in intensive care

External Links

• article Positive communication for decreasing burnout in intensive-care-unit staff: a cluster-randomized trial
• protocol HELLO: a protocol for a cluster randomized controlled trial to enhance interpersonal relationships and team cohesion among ICU healthcare professionals

Metadata

Summary author: Conor Foley
Summary date: 22nd December 2025
Peer-review editor: David Slessor

Picture by: Oles Kanabckuu / StockSnap