This study will investigate the effect of diaries written by relatives for critically ill patients on the risk of posttraumatic stress symptoms.

Utilizing a randomized controlled design pairs of one patient and one relative will be assigned to either diary intervention or standard care. The effect of the diary will be evaluated using a questionnaire 3 months post charge.

The results of this study will add to the body of knowledge on how diaries may help critical care survivors and their relatives to proceed with their lives.

Diaries written for critically ill patients has previously been shown to protect patient and relatives from new cases of post traumatic stress (Jones et al. 2010, Jones et al. 2012, Garrouste-Orgeas 2012). Furthermore it has been described how relatives and patients use diaries written by nurses to process the time they stayed in the intensive care unit (Egerod et al. 2011). This study will investigate if there is a similar effect, when relatives write the diary for the critically ill patient instead of nurses doing it as patients find it very valuable when relatives has contributed to the diary or when relatives presence has been documented in the diary (Engström et al. 2009).

The investigators hypothesis is that a diary written by relatives for critically ill patients will reduce the risk of developing symptoms of post traumatic stress in relatives and patients.

Design: randomized controlled study. Relative and patient are randomized as a pair to either intervention group (diary intervention) or control group (standard care). Relatives will be recruited to the study while the patient is critically ill in the ICU, the patient will be recruited when he or she is able to receive information about the study.

Intervention: Nurses guide the nearest relative (spouse, partner, child, parent or friend) to write a diary during the patients stay in the ICU and up to a month post discharge. The diary is a notebook which also contains a written instruction on how to write and use the diary during and after the time in the intensive care unit. At least two photographs of the patient is taken by the staff and added to the diary upon the patients consent.

Blinding: Due to the nature of the intervention neither caregivers nor participants can be blinded to the intervention. However, steps will be taken to ensure that the outcomes assessor is blinded.

Setting: the study will take place in the two ICUs (10 bed and 7 beds respectively) of the hospital unit Hospitalsenheden Vest, a 441 bed hospital serving 300.000 citizens. Both ICUs are mixed medical/surgical units.

Primary outcome: Risk of post traumatic stress symptoms (PTSS-14 (Twigg et al 2008).

Secondary outcomes: Anxiety and depressionsymptoms (HADS (Zigmond and Snaith 1983)) and general physical and psychological wellbeeing (SF-36 (Bjorner et al. 1998)).

Baseline data: To compare groups the following parameters will be recorded on all eligible patients. Patients: age, sex, education, marital status, diagnosis, SAPS-II score (Le Gall et al. 1993), hours on ventilator, LOS-ICU. Furthermore the following data will be recorded on participating relatives: age, sex, relation to patient, education and working status.

Sample size: Based on findings from a previous study (1) where the median PTSS-14 score was 24 (SD 12), alfa 5% and beta 80% and a minimal relevant difference of 4 points, 71 relative-patient pairs needs to be included in each group. To compensate for dropouts due to critical illness and death, 100 pairs will be included in each group. The data collection period is anticipated to last 24 months.

Follow-up: 3 months post ICU discharge all participants receive a questionnaire to be completed and returned by mail or email.

Statistical analysis will cover a comparison of mean and median values in the two study groups, supplemented by a stratified analysis to estimate risk reduction. Parametric and non-parametric test will be applied according to the nature of data.

Dissemination: Results, positive, negative or inconclusive will be published in an international peer-reviewed journal.

Inclusion Criteria:

Patients:

• Expected to stay in ICU > 48 hours
• Expected to be on ventilator > 24 hours
• Has a close relative > 18 years old
• Speaking and understanding Danish.
• Expected to be able to deliver an informed consent later

Relative:

*Speaking and understanding Danish

Relative and patient are included in the study as a pair.

Exclusion Criteria:

• Patient staying less than 48 hours in ICU
• Patient less than 24 hours on ventilator
• Patient not able to give an informed consent to participate in study or patient dies in ICU (both relative and patient are excluded).