Effect of a condolence letter on grief symptoms among relatives of patients who died in the ICU: a randomized clinical trial

Kentish-Barnes. Intensive Care Medicine 2017; first published on line 14th Feb. DOI 10.1007/s00134-016-4669-9

Clinical Question

  • Among relatives who died in the Intensive Care Unit (ICU), does a condolence letter from a clinician, compared to no condolence letter, effect grief symptoms?


  • Studies have shown that over half of bereaved relatives experience grief symptoms, characterised by variably combined symptoms of anxiety, depression, stress and complicated grief
  • Two previous RCT’s have shown that improving communication during end-of-life care alleviated grief symptoms in relatives of patients who had died in the ICU
  • This RCT evaluates the effect of a simple implementation, a condolence letter, to try and improve communication and grief symptoms after death


  • Multicenter randomised control trial
  • Randomisation occurred with 24 hours of the patient’s death
  • Centralised internet-based randomisation procedure was used
  • Stratified according to study centre with concealed sized permuted blocks
  • Single relative per patient was included
  • Informed consent was sought in the hours following patient’s death
    • the patient information sheet did not specify the nature of the intervention, but that a strategy to improve communication was being evaluated
  • Intention to treat analysis
  • Power calculation: 240 patients would give the study 90% power to detect a  decrease in the prevalence of anxiety and depression from 60% to 42%, and a 30% decrease in the Hospital Anxiety and Depression Scale Score


  • 22 ICUs in France
  • Dec 2014 – Oct 2015


  • Inclusion: age at least 18 years; at least one family visit prior to death; and an ICU stay of at least 2 days
  • Exclusion: pregnant and family not fluent in French
  • 242 relatives of patients that died in ICU
  • Baseline characteristics were similar for key confounders apart from:
    • more of the follow up interviews occurred with spouses in the intervention group (38.5% vs 32.3%)
    • interviews with children were similar; ‘others’ were higher in the control group (26.2% vs 22%)
    • more family-clinician or intra-team conflicts in the control group, although small numbers 13 vs 9
    • more patients were intubated at time of death in the intervention group (58.5% vs 54.6%), and sedated at time of death (78.9% vs 76.5%)
    • preferred role of relatives of being only informed of end-of-life decision, lower in intervention group 25% vs 34%


  • Condolence letter
    • handwritten by the physician and nurse in charge
      • covered the following 5 domains: recognise the death and name the deceased; mention a personal impression of the deceased; recognise the family member; offer help; and express sympathy
    • sent by standard mail 15 days after the patient death


  • No condolence letter
In both groups
  • Symptom control, timing of communication at the end-of-life and implementation of treatment-limitation decisions were at the clinician’s discretion
  • After randomisation, the clinicians recorded all reactions or feedback (telephone calls, letters, visits, or other) from the relatives within 4 months following death
  • Telephone interviews with family members occurred at 30 days and 6 months after the patient’s death. These were conducted by psychologists, sociologists, and research nurses blinded to study group


  • Primary outcome: at one month. Expressed as median (25th-75th percentile) unless stated otherwise
    • Hospital Anxiety and Depression Scale (HADS) Score: higher in the intervention group but did not reach statistical significance based on P value
      • 16 [10-22] vs 14 [8-21]; P=0.36, mean difference 0.77 (95% C.I. -1.7 to 3.3)
    • HADS-depression sub-scale scores: higher in the intervention group but no significant difference
      • 8 [4-12] vs 6 [2-12]; P=0.10, mean difference 1.1 (95% C.I. -0.5 to 2.6)
    • Relatives with symptoms of depression (sub-scale score >8): higher in the intervention group but no significant difference
      • 56% vs 42.4%; P=0.054, relative risk = 0.76 (95% C.I. 0.57-1.00)
    • HADS-anxiety sub-scale scores: no significant difference
      • 7 [4-11] vs 7 [4-12], P=0.92
    • Relatives with symptoms of anxiety (sub-scale score>8): no significant difference
      • 47.7% vs. 45.5%, P=0.97
  • Secondary outcomes: at 6 months, and were all worse in the intervention group. Expressed as median (25th-75th percentile) unless stated otherwise
    • HADS score
      • 13 [6–19] vs. 9 [4–17]; P=0.04
    • HADS-depression sub-scale scores
      • 6 [2–10] vs. 3 [1–8]; P=0.01
    • Relatives with symptoms of depression (sub-scale score >8)
      • 37% vs. 25%; P=0.05
    • Impact of Event Scale Revised (IES-R) score
      • 28 [15-38] vs 20 [10-37]; P=0.09
    • Relatives with significant post traumatic stress disorder (PTSD) related symptoms (IES-R ≥26)
      • 52.4% vs 37.1%; P=0.03
    • Inventory of Complicated Grief (ICG) score
      • 16 [8-30] vs 13 [4-27]; P=0.07
    • Relatives at high risk for complicated grief (ICG ≥25)
      • 38% vs 27%; P=0.11
  • Other
    • Feedback was received from 55 (44.7%) intervention- group relatives and 8 (6.7%) controls; P < 0.0001
      • Of the 55 intervention-group relatives who gave feedback, 50 did so to thank the clinician for the condolence letter
      • No relatives complained about the condolence letter

Authors’ Conclusions

  • In relatives of patients who died in the ICU, a condolence letter failed to alleviate grief symptoms and may have worsened depression and PTSD-related symptoms


  • A relevant and important study evaluating the impact of a simple measure to support families of patients that have recently died on ICU
  • Multicentre
  • RCT including university and non-university hospitals
  • Interviewers were blinded
  • All letters complied with the study guidelines
  • Trial registered on


  • All participating ICUs were in France, which may limit the general applicability of the findings, as end-of-life care and bereavement follow-up vary widely across countries
  • All ICUs involved had a special interest in family-centred care
  • Intervention was not standardised. Whilst the clinicians followed guidelines for the letter, they were otherwise free to write what they wanted to
  • The psychological status of the relatives at the time they received the condolence letter was not evaluated
  • The outcome measure were assessed at one and six months. This may have been too short to demonstrate any longer term benefit
  • 14% relatives (34 patients) did not complete the HADS at one month

The Bottom Line

  • This is an important study evaluating a simple intervention to try and improve grief symptoms amongst family members. It unexpectedly showed that condolence letters may increase depression and PTSD symptoms amongst family members at 6 months
  • Grief is a complex process with many confounders, including individual adaptive response and external influences. These nuances and impact are difficult to evaluate
  • Our current departmental policy of sending out condolence cards will continue. However, we will review the timing, content and potential deleterious effects

External Links


Summary author: Steve Mathieu
Summary date: 21st June 2017
Peer-review editor: David Slessor

One comment

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.