The effect of diaries written by relatives for intensive care patients on posttraumatic stress (DRIP study): protocol for a randomized controlled trial and mixed methods study

ICU treatment often involves a prolonged period of ventilator dependency in a highly technical environment, resulting in a wide range of physical, cognitive and psychological impairments [1, 2]. Although many patients regain their strength within the first year, their quality of life is generally negatively affected [3‐5].

Psychological problems include symptoms of post-traumatic stress disorder (PTSD), anxiety and depression. Approximately 20% of ICU survivors have clinically significant PTSD symptoms [6], consisting of re-experiencing symptoms, avoidance symptoms, arousal symptoms and cognitive symptoms [7, 8]. Ehlers and Clark [9] suggest that PTSD is characterized by the disruption of meaningful connections between events. Moreover, risk factors include having frightening or unreal experiences in the ICU [6, 10, 11], whereas accurate memories of intensive care may protect against developing PTSD [11]. The prevalence of clinically significant levels of anxiety in ICU survivors is reported to be 6.9–18% [12], and the prevalence of depression is reported to be 28% [13]. Furthermore, 50% of patients surviving critical illness do not return to school or work 1 year post-ICU [14], and impairments have a substantial effect on patients’ and their partners’ financial situations [15].

A key part of restoring quality of life after critical illness is the need to understand what has happened and why it happened. This can be difficult, as memories of intensive care may be absent, incomplete or extremely disturbed due to the critical illness and the use of sedatives. However, research show that patients continuously strive to make sense of their ICU experience [16].

When witnessing critical illness in a close family member, relatives are confronted with uncertainty about the patient’s survival [17]. During this critical period, relatives need support from healthcare professionals to reduce family stress [18]. In the rehabilitation period, the patient struggles to rebuild strength and cognitive abilities, while relatives takes the position of informal caregivers [19, 20]. Supporting the patient influences relatives’ lives, restricting daily activities and limiting the chances of keeping one’s job [21, 22]. Family members’ responses to critical illness comprise a range of psychological issues, including symptoms of PTSD (median 21%, range 13–56%) [18], depression (median 23%, range 8–42%) and anxiety (median 40%, range 21–56%) [18]. Providing a follow-up visit to the ICU could be a way of meeting family needs after discharge. Relatives have found this helpful in dealing with the ICU experience [23]. Furthermore, diaries have been described as an initiative aimed at supporting both patients and their relatives in their efforts to deal with these problems [24].

Patient diaries written by critical care nurses have been used to help patients make sense of their ICU experience and to fill memory gaps [25, 26]. In previous research, diaries have been initiated when the length of stay was predicted to exceed 72 h and involved mechanical ventilation [27, 28]. Research has shown that nurses use their professional knowledge to create an empathic and personal story as imagined from the patients’ perspective [29].

Patients have described diaries written by nurses as helpful [29], but they particularly appreciate parts that reflect the presence of their relatives [30]. Patients and relatives have used a diary kept by nurses as a basis for dialogue about their time in the ICU [24]. This has helped relatives support patients and has promoted patient understanding of the ICU stay [24]. However, the optimal timing of the handover and reading of the diary to the patient can be difficult to determine, as the patient may feel uncomfortable if confronted with the diary too prematurely [24].

Diaries coauthored by relatives and staff have been shown to facilitate information and communication between staff and relatives [31], making ICU staff better understand relatives’ vulnerability [31, 32]. However, this has changed the objective of the diary from helping the patient to improving communication between staff and relatives. Moreover, as research has shown that relatives’ writing styles are far more emotional than nurses’ [33], relatives will be at risk of feeling exposed when co-authoring the diary with nurses. Diaries prompted by nurses but authored by relatives have enabled relatives to unload emotions, express their feelings and reflect on the critical situation [34], but patients’ perceptions of and use of such diaries have yet to be described.

Diaries have been proposed as a method to prevent or reduce patients’ psychological problems, including PTSD symptoms [35, 36], anxiety and depression [37], as well as the relatives’ problems [36, 38].

The effect of diaries on ICU patients’ psychological outcomes was studied in a randomized controlled trial (RCT) [35] wherein diaries were authored primarily by nurses, although relatives were provided with an option to participate. The diaries significantly reduced new cases of PTSD in patients (13.1% in the control group vs. 5% in the diary group, p = 0.02). However, this estimate may have been biased, as more cases of existing PTSD were excluded post-randomization from the intervention group (n = 8) than from the control group (n = 3). Furthermore, mean PTSD symptom scores did not differ significantly between groups [35]. A time-series-controlled study of a collaboratively authored diary showed no difference in PTSD, anxiety or depression in patients at 3 months post-ICU [36]. At 12 months post-ICU, patients in the diary group had fewer PTSD symptoms; however, the actual number of cases was not significantly reduced. Another smaller RCT [37] found that diaries authored by nurses led to a small but significant reduction in cases of anxiety and depression among patients in the diary group, whereas no significant changes were found in the control group.

The effect of diaries on relatives was investigated in a pilot study of patient diaries authored by nurses with optional participation for relatives [38]. This small study suggested a positive effect on relatives (median PTSS-14 scores increased from 26 to 28 in the control group and decreased from 24 to 19 in the diary group, p = 0.03) at 3 months post-ICU [38]. A controlled time-series study of collaboratively authored diaries [36] found a lower occurrence of PTSD (defined as the Impact of Event Scale-Revised > 22 points) among relatives using a diary than among a pre-intervention group and a post-intervention group (the diary group 31.7% vs. pre-intervention group 80% and post-intervention group 67.6%, p < 0.0001), but no difference at 3 months post-ICU.

Diaries authored by nurses may help the critically ill patient construct an illness narrative [24]. At a more theoretical level, a narrative is described by Ricoeur as a synthesis of heterogeneous elements that transform multiple incidents into one concordant story, moving the story to its conclusion despite its discordant elements [39]. Thus, a narrative is more than a sequence of unconnected events. It is a unity in which the different elements support the conclusion of the story. Driving the story toward a preferably tidy end is not always possible [40]. Critical illness represents a disruption and fragmentation in the lives of patients and their relatives. Therefore, the illness narrative can serve the purpose of restoring a new order or identifying a new purpose instead of the tidy end, which may no longer be available or appropriate [40]. This use of narratives has been described as a way to identify or plot a livable future for severely injured patients, thus giving them hope and something for which to live [41]. Plotting a narrative can also be described as an active way of processing and understanding what is happening.

According to Ricoeur’s tri-fold mimesis [42], writing a diary may be viewed as a creative process of configuring lived experience into text [34]. However, asking the relative to author the diary may produce a narrative in the diary that is not the patient’s story. More specifically, it is the relative’s configuration of the patient’s story. However, when the relative and the patient read and discuss the diary post-ICU, this supports the patient’s configuration of his or her story. Appropriation of the text through reading opens new understandings and meanings of critical illness for both the patient and the relative [34]. This can explain how the diary relieves psychological problems common in relatives of ICU survivors.

Overall, the purpose of the ICU diary has evolved, but it has not yet been fully explored. Investigating the effect of ICU diaries authored by relatives is therefore warranted. Furthermore, a more comprehensive understanding of how a diary written by relatives is perceived by patients and relatives is needed to determine if such diaries should be implemented on a larger scale.

The aim of the study is to evaluate the effects of Diaries written by Relatives of Intensive care Patients in relation to the psychological recovery of both the patient and their family (DRIP study) and gain a more comprehensive understanding. The study has the following specific objectives:

The study is a single-blind, two-armed, randomized controlled trial comparing a diary intervention with standard care as control (1:1). The randomized controlled trial will be combined with two hermeneutical phenomenological studies: an interview study addressing patients’ and relatives’ perception and use of the diary and an analysis of the diary content and structure. Thus, the study is a convergent mixed methods study [43], where all studies will be conducted simultaneously, analyzed separately but where results/findings will be synthesized to arrive at a more comprehensive and complete understanding of the diary intervention [43].

The study setting is four medical/surgical ICUs at two regional hospitals and one university hospital in Western Denmark. Levels of ICU certification among nurses range from 50 to 60% at the university hospital to more than 90% in the regional hospitals. A list of study sites can be obtained from the first author.

Relatives randomized to the intervention group will receive a good quality hardcover notebook for diary use, including written information and guidance from nurses. Relatives and patients randomized to the control group will receive standard care (no diary). Written information for the relatives on how to author the diary is enclosed in the diary (see Table 1). The instructions describe the purpose of the diary as a way for the relative to help the patient after their ICU stay, as well as the potential value the diary may have for the relative. To help the relative begin writing, a list of possible topics is provided for inspiration [36] (see Table 1). The relatives are encouraged to write often and to use their own way of expressing themselves. But to avoid harming the relatives, they will be warned against writing thoughts and feelings too private or personal to be shared [44]. Finally, the leaflet will provide advice for when and how to share the diary with the patient, underscoring the need to wait until the patient agrees to see the diary to prevent the patient from being confronted with their difficult time in the ICU until he or she is ready [45].

The primary measure of outcomes in the randomized controlled trial is the Post-traumatic Stress Symptoms-14 inventory (PTSS-14) [48]. Secondary outcomes are measured by scores on the Hospital Anxiety and Depression Scale (HADS) [49] and the Medical Outcomes Study Questionnaire Short Form (SF-36) [50]. The qualitative studies focus on the meaning that the diary holds to relatives and patients as well as describing the structure and content of the diaries [51].

For the randomized controlled trial, a minimum of 71 patients in each group will be required, assuming a significance level of 5%, a power of 80%, an MCID of 4 points and a median PTSS-14 score of 24 (SD12) [35]. The Danish National Patient Registry of 2014–15 identified 3325 admissions to the study sites, out of which 1016 patients received ventilator treatment. ICU mortality was close to 30%. Based on this, we aim to include 100 patients/relatives pairs in each arm of the study. We estimate that 50% of the mechanically ventilated patients will be eligible for the study, therefore inclusion in the project is estimated to run for 30 months, starting March 2015 (see Fig. 1).

For the qualitative studies, 12 patient-relative dyads who have received the intervention will be sampled among participants in the intervention group. Patient-relative dyads will be selected to achieve phenomenological variation. This sampling strategy aims to achieve variations of experience of the phenomenon rather than variations of specific demographic variables [52]. In practice, we will approach patients and relatives consecutively as each individual has their own unique experience to share. Sampling diaries within the interview group will create strong coherent qualitative data.

Relatives will be randomized at a 1:1 ratio to either the intervention group or control group by the investigator or by study nurses using sealed, opaque envelopes. Block randomization ensures that each site randomizes equally to both groups. Because this is not feasible, the investigator, study nurses and participants will not be blinded to group allocation. However, to reduce the risk of detection bias, the investigator will be blinded to group allocation during data entry. The identity of each respondent will be masked and can only be re-identified using a separately kept key-file.

Patient baseline data include date of birth, age, sex, LOS-ICU, hours on ventilator, reason for admission to intensive care (medical/surgical), SAPS-II score [53] and level of education (higher, medium/low or none). Relative baseline data include age, sex, relationship to patient and level of education. Baseline data will be managed using Epidata software.

The primary outcome is measured on the PTSS-14 scale [48] developed for critically ill patients and used in studies evaluating the effect of a nurse-written diary [35, 38] on patients and relatives. The PTSS-14 is an update of the existing PTSS-10 scale based on the DSM-lll criteria, including items related to numbing and re-experiencing corresponding with the DSM-lV [48]. The PTSS-14 has been validated by Twigg et al. [48] in a cohort of patients from general ICUs in the UK. Content validity was assessed by a multidisciplinary group of experts in PTSD [48]. Criterion validity was assessed by comparing the PTSS-14 to the Post-traumatic Stress Diagnostic Scale (PDS), showing good concurrent validity with Pearson’s correlation of 0.86 at 3 months post-ICU discharge [48]. An ROC curve analysis determined 2 months post-ICU as the most favorable time to administer the PTSS-14 (AUC 0.950) [48]. However, for the diary to have time to show an effect, we will administer the PTSS-14 at 3 months (AUC 0.936, sensitivity 100%, specificity 84%) [48]. The PTSS-14 has previously been used in the RACHEL study [35, 38]. For that study [35], the PTSS-14 was translated into Danish. The process of translation involved translation back and forth, but no further validation of the Danish version. The PTSS-14 has been shown to be easy to administer to the fragile population of former ICU patients [48]. However, the fact that it is neither based on the newer DSM-V criteria nor fully validated in Danish poses some limitations. Secondary outcomes are determined by using the HADS [49] and SF-36 [50]. The HADS assesses anxiety and depression. The two scales have moderate-to-strong correlation [54], and optimal balance between sensitivity and specificity is achieved using a cut-off value of 8 for both subscales. This was done by Knowles and Tarrier [37], whereas actual cases are more conservatively estimated using a cutoff value of 11 [55]. Finally, the SF-36 will be used to measure health-related life quality. The SF-36 measures health in 8 domains [50]. Although a newer version exists (SF36 v.2), the original version will be used, as Danish reference values for this version exist [50]. Using scales to assess psychological problems like PTSD, anxiety and depression have the inherent limitation that no actual diagnosis is made; instead, we focus on reducing symptoms. Unfortunately, no general agreement exists on which outcomes to measure in trials regarding interventions to improve rehabilitation for ICU patients [55], so to ensure comparability with other studies, the PTSS-14, HADS and SF36 have been chosen. All inventories are available in Danish. They will be administered at 3 months post-ICU discharge by mail or e-mail (SurveyXact), depending on the choice of each participant. Paper questionnaires will subsequently be entered into SurveyXact by the investigator and pseudonymized.

The photocopied diaries will be transcribed verbatim, noting drawings, changing of pages, underscoring and other non-text elements. Pictures will be described in terms of content.

Interviews will be conducted at the homes of the patients and/or relatives in order to keep relatives and participants as comfortable and safe as possible. Open-ended questions will be used to explore the participants’ experiences of the diary allowing participants to recount these experiences in their own language [56]. The interviews will be digitally recorded and transcribed verbatim. Expressions of feelings will be put in parenthesis (e.g., sounds of laughter, crying, loud or muffled voices) and pauses indicated by ellipses in the transcripts [56].

From the questionnaires, the total PTSS-14 mean score will be calculated for each group and compared using students’ t-test. If necessary, logarithmic transformation will be undertaken to achieve normality. The two HADS subscales describing anxiety and depression will be analyzed separately. Both subscales will be dichotomized using a cut-off value of 11. The number of cases in each group will be compared using a Chi-square test. The eight domains of the SF-36 will be calculated for each group and compared using Wilcoxon’s rank-sum test. The statistical analysis will be managed using STATA software.

In the analysis of the qualitative data (diaries and interviews) the meaning will be sought through a three-step analytic process, drawing on the interpretation theory of Ricoeur [51]. Thus, the data material will first be read thoroughly to become familiarized with it. A naïve interpretation will be formulated for each interview or diary, representing an overall naïve understanding [51] of the data. The structural analysis within the text will explain the text in terms of its smaller parts, moving from what the text says to what may be understood from the text - its reference to the world [51]. The process of explaining the text in terms of its internal relationships will serve as a validation or modification of the initially formulated naïve interpretation. This will allow the most significant interpretation to stand out in the critical interpretation. Thus, Ricoeurs’ interpretation theory will allow us to move from the specific situation of the participants to a more general and comprehensive understanding of diaries authored by relatives [51].

Subsequently, the findings of the qualitative studies and the results of the RCT will be compared. The extent to which they relate to each other will enable a more comprehensive and nuanced interpretation of the diary intervention to be made.

In order to assess the fidelity of this very complex intervention, a number of key processes will be monitored [57]. The number of eligible relative/patient pairs approached for the study will be used in monitoring recruitment to the study. After randomization, the number of pictures taken of patients in the intervention group will be documented and monitored. As part of the follow-up, the questionnaire will contain questions assessing the delivery of the intervention and its use by relatives and patients and finally, the interview study and the study of diary content and structure will also function as a sample for a qualitative evaluation of fidelity [57]. A data-monitoring committee has not been considered necessary due to the lack of competing interests and due to the small scale of the study.

Should any unintended effects or adverse events be observed by nurses or reported by participants, support will be offered by the medical staff at the involved study sites. After discharge from the ICU, relatives and patients will be advised to contact the ICU or the investigator in case they experience any harm.