Impact of post-traumatic stress symptoms on the health-related quality of life in a cohort study with chronically critically ill patients and their partners: age matters

Survivors of an acute critical illness with continuing organ dysfunction and uncontrolled inflammatory responses are prone to become chronically critically ill [1]. Although there is no consensus about the precise definition, chronic critical illness represents a syndrome comprising multiple clinical features. Primary characteristics are prolonged mechanical ventilation (> 72 h) and the need for elective tracheotomy [2, 3]. Minor criteria are neuromuscular weakness, immunodeficiency, endocrinopathy, malnutrition, anasarca, and psychological distress [4‐6]. After the discharge from the intensive care unit (ICU), these patients require a continued high level of intensive nursing care and are at an increased risk for ongoing functional impairments, hospital readmissions, and a high mortality rate [4, 7]. The affected patients are exposed to a protracted critical situation which implies a severe emotional stressor for the whole family system (for systematic reviews, see [8, 9]). As a consequence, a cluster of adverse mental sequelae may occur in family members, including depression, anxiety, and acute/post-traumatic stress disorder (PTSD). These have been referred to as Post-intensive Care Syndrome-Family (PICS-F) [8, 9]. According to the Diagnostic and Statistical Manual of Mental Disorders (DSM-V), the mere learning that a relative or satisfying) to 10 (very satisfying) to 10 (satisfying friend was exposed to a trauma is sufficient to pioneer the etiopathogenesis of PTSD [10].

Current findings in family members of chronically critically ill patients reveal that 16% show clinically relevant symptoms of post-traumatic stress within up to 6 months following the ICU stay. Additionally, these family members show a significantly diminished health-related quality of life (QoL) compared to a normative sample [11]. According to the Theory of Dyadic Illness Management [12], the health outcomes of both patients and their partners may influence each other and necessitate the treatment of the dyad as an interdependent team. In this regard, intercorrelations of post-traumatic stress symptoms between patients and their partners have been found in general ICU patients [13] and following severe sepsis [14]. Furthermore, psychological symptoms of patients and their partners or spouses affected not only their individual health-related QoL (actor effects) but also that of the respective other (partner effects) [14, 15]. Thus, studies including patients and partners should treat the patient-partner dyad as interacting and non-independent unit of analysis using appropriate statistical strategies such as the Actor-Partner-Interdependence Model (APIM) [16]. Currently, there is a lack of studies on mental long-term sequelae in chronically critically ill patient-partner dyads using appropriate dyadic analysis strategies. Besides, there is evidence that several contextual factors may represent risk or protective factors on the dyadic appraisal and management behaviors of the chronically critically situation [12]. Within the aftermath of ICU treatment, psychiatric disorders are more common in partners than other kinship relations (e.g., [17]). Furthermore, there is evidence that especially partners of younger patients [18, 19], of patients with sepsis, and female partners are prone to an increased risk for mental distress [14, 20, 21] (for a review, see [22]). However, these findings have not been unequivocally clarified for chronically critically ill patient-partner dyads [14].

Addressing the psychological needs of the family members, particularly the partners, is of clinical relevance, since high levels of post-traumatic stress symptoms may increase the risk for a full syndromal PTSD. In the long-run, the partners might not appropriately fulfill their role as informal care-givers and surrogate decision makers. As a consequence, irrational or uninformed decision-making may occur, leading to prolonged ICU stays and diminishing the rehabilitation outcome in chronically critically ill patients [22, 23]. Therefore, the main aim of the present study was to investigate the dyadic relation between post-traumatic stress and health-related QoL in a homogeneous sample of chronically critically ill patients and their partners, applying a straightforward time schedule of 6 months after being discharged from acute care ICU. As analysis strategy, the APIM approach was used, considering differential effects of the patients’ and partners’ age and gender.

Of the N = 207 enrolled chronically critically ill patients, data of n = 70 patient-partner dyads could be collected 3 or 6 months following the transfer to post-acute ICU (see flow chart, Additional file 2: Figure S1). The partners of the chronically critically ill patients were assessed with respect to post-traumatic stress and health-related QoL at a median time of 4.8 months (IQR 3.9–6.5) following the ICU discharge. The partners were on average 1 year younger than the patients (median 61.6, IQR 56.1–66.1). 75.7% of the partners were female. The patients stayed a median time of 62.5 days in ICU (IQR 45.5–99.5). The median time of mechanical ventilation was 48.5 days (IQR 28.8–76.0). The dyadic relationships had a median length of 37 years (IQR 26.5–42.5). The partners evaluated their relationships as very satisfying or satisfying) to 10 (very satisfying) to 10 (very close (median 10.0, IQR 8.0/9.0–10.0). For a detailed description see Table 1.

The patients who dropped out were more often single or widowed showed a lower level of education and had a significantly higher Barthel index at discharge from the rehabilitation hospital than the patients who were followed up (see Additional file 3: Table S2). Three quarters of the patients had the diagnosis of an acute respiratory insufficiency, and more than one third had a coronary heart disease or diabetes. There were no significant differences between patients who were included or dropped out with respect to medical comorbidities (see Additional file 4: Table S3).

The primary aim of the present study was to investigate the association between post-traumatic stress and health-related quality of life (QoL) simultaneously in chronically critically ill patients and their partners following the discharge from the intensive care unit (ICU). In summary, our main results revealed that nearly every fifth patient (17.1%) and partner (18.6%) showed clinically relevant symptoms of post-traumatic stress. Additionally, the patients reported a significantly decreased health-related QoL compared to their partners and the general German population. Moreover, more severe post-traumatic stress symptoms were associated with a lower health-related QoL in both the patients and their partners (actor effect). The impact of post-traumatic stress symptoms on the health-related QoL of the respective other (partner effect) was only present in partners of dyads aged 57 years or younger. Finally, male partners showed higher actor effects than female partners and female patients (Additional file 7).

Our rate of clinically relevant post-traumatic stress symptoms is in the range of 13% or 35 to 57% reported by Davidson et al. [8] and van Beusekom et al. [9] for relatives of critically ill patients. With respect to family members or spouses of chronically critically ill patients, there is only one study currently available, which addresses post-traumatic stress symptoms as primary outcome. Rosendahl et al. [14] found 69% and 62%, respectively, of clinically relevant post-traumatic stress symptoms in severe sepsis survivors and their spouses following an average of 55 months after sepsis. However, these heterogeneous rates should be evaluated critically in the context of different samples, assessment methods, and cutoff values used to define the clinical relevance of post-traumatic stress symptoms.

The decreased health-related QoL found in the present study confirms former results (e.g., [9, 14, 35‐38]) and adds insight into the kind of special burden chronically critically ill patients, and their partners are suffering from following long-term ICU treatment. These patients belong to a patient cohort requiring prolonged mechanical ventilation, being faced with a high risk for complications during their ICU stay as well as an ongoing high mortality rate, and lasting physical limitations following the ICU discharge [39, 40]. The partners of chronically critically ill patients are often mainly involved in informal caregiving in the aftermath following ICU treatment. In the present study, about two thirds of the partners reported being the caregiver for the chronically critically ill patient at the moment. In line with the present literature, informal care-giving is often associated with decreased physical and emotional health, especially when the patient does not recover fully like in chronically critically ill patients (for a systematic review, see [22, 41, 42]).

In the APIM, significant actor effects in both the patients and their partners were demonstrated, which is in accordance with existing findings [14, 15]. A significant effect of a person’s post-traumatic stress symptomatology on the respective other’s health-related QoL (partner effect) could not be shown, contrasting existing findings in survivors of severe sepsis and patients with heart failure [14, 15]. However, the previously mentioned studies revealed significant partner effects in only one direction. Lacking partner effects in our study might be explained by differences in sample characteristics between studies. In our sample, not only spouses but also life partners were included. Above, Chung et al. [15] only considered spouses who were identified as primary caregivers in patients with heart failure. In our sample, only about two thirds of the partners reported to be involved in the chronically critically ill patient’s care at the moment which may have led to a lack of partner effects. Furthermore, differences between studies according to the measurement of the health-related QoL should be taken into account in order to understand the inconsistency of the findings. While Chung et al. [15] used a disease-specific instrument (the Minnesota Living with Heart Failure Questionnaire), Rosendahl et al. [14] applied the mental component summary of the Short-Form-12 (SF-12) Health Survey. While the latter shows considerable overlap with affective symptoms of post-traumatic stress [43], the EQ-5D-3L used in the present study, only includes one item to evaluate a person’s mental situation (e.g., “I feel extremely anxious or depressed.”). This may have led to a lack of partner effects.

When age groups were formed, particularly younger partners of chronically critically ill patients revealed a significant partner effect on the patients’ health-related QoL. This result confirms findings by Anderson, Arnold, Angus, and Bryce [44] and Gries et al. [45], showing that a younger relative’s or patient’s age displays a major demographic risk factor for the development of the PICS-F. Our finding extends the existing literature from acutely ill patients treated in the ICU to chronically critically ill patients and their partners. Pankrath et al. [46] also showed moderating effects of age in a cohort of chronically ill patients with hematological cancer. Younger patient-partner dyads may be more affected by the chronically critically state of the patients than older patient-partner dyads. The present results suggest that this effect seems to be independent of the length of the partnership. Probably, younger patient-partner dyads (≤ 57 years) showed a greater emotional interdependence due to the major problems evolving from the chronic critical illness during this specific phase of the life span (including, e.g., occupational disability, financial burden, care for minor children). In contrast, older patient-partner dyads are more often adapted to chronic disabilities, and thus, role changes, financial issues, and disruptions of daily activities can be dealt with more effectively than in younger patient-partner dyads [46, 47]. Moreover, younger partners are more often solely involved in the informal caregiving than older partners are. This reinforces satisfying) to 10 (very satisfying) to 10 (very close interactions and emotional transmissions leading to the induction of feelings of empathy in the members of the dyad.

With respect to the impact of gender, our study could not confirm former results for chronically ill patients, showing a transmission of emotional stress only from male patients to their female partners [20]. A larger actor effect of a person’s post-traumatic stress on his/her individual health-related QoL could be shown for male partners compared to female partners and patients. However, this finding should be regarded as rather exploratory, since only about one quarter of our patients were female. However, it may be assumed that men probably perceive the chronic critical illness of their female partners as more profoundly incisive than women, with respect to role changes and disruption of daily routines. Future research is needed, taking into account the perceived satisfaction and satisfying) to 10 (very satisfying) to 10 (very closeness of the partnership as moderating variables. Moreover, future studies should focus on the impact of dyadic coping (e.g., supportive dyadic coping vs. lack of emotional involvement) on the patients’/partners’ outcomes and relationship satisfaction [46].

Although the present study has several strengths such as prospective data assessment, the investigation of a homogeneous sample of chronically critically ill patients and the use of a straightforward time-frame of 6 months, our results should be carefully evaluated in the context of methodological shortcomings. First, we studied a convenience sample of chronically critically ill patients recruited during their weaning in a large rehabilitation hospital. On the whole, this sample is similar to other cohorts recently described in the literature in that the participants and their partners were in their early 60s, most of them were married and had an education of about 10 years [2, 14]. Moreover, our mortality rate is similar to that of existing findings in chronically critically ill patients [2]. Nevertheless, some peculiarities of our present sample need to be pointed out, e.g., males and females were not evenly divided. Beyond that, our patients had stayed in the ICU for about 8 weeks and had been ventilated for about 6 weeks. However, current evidence in chronically critically ill patients is based on patient samples receiving mechanical ventilation/length of ICU stay between 3 and 4 weeks or shorter [2, 14, 36, 48]. Thus, the generalizability of our results is restricted to a highly specialized cohort of patients in need of protracted critical care.

Second, PTSS-10 was used instead of the PTSS-14. The latter presents a more reliable and valid screening instrument for the assessment of post-traumatic stress symptoms since it also represents the diagnostic criteria re-experiencing and numbing [49]. Third, the correlative nature of the present results should be considered since post-traumatic stress and health-related QoL were assessed concurrently. Information regarding physical comorbidities, post-traumatic stress, and health-related QoL before admission to post-acute ICU were not available in both the patients and their partners. Thus, a causal attribution of the present data to, e.g., the protracted treatment on ICU is not possible.

Fourth, the present results should be carefully evaluated in the context of the rather high drop-out rate. Although, our rate of 66.2% mirrors the common clinical situation in chronically critically ill patients [2, 14, 50], more severely ill patients could not be followed up. This may have led to an underestimation of the patients’ and their partners’ psychological distress as well as the partner effects following the ICU treatment. Accordingly, a limited sample sizes of N = 70 dyads could have been analyzed. A post hoc power analysis revealed that with 1-ß = 72%, our study was underpowered and a minimum sample size of n = 84 dyads would have been necessary to reach a power of 80%.

To our knowledge, this is the first study on the impact of post-traumatic stress symptoms and the health-related QoL in chronically ill patients and their partners following prolonged intensive care. In summary, nearly every fifth chronically critically ill patient and his/her partner suffered from clinically relevant symptoms of post-traumatic stress with a significantly negative impact on the health-related quality of life, especially in male partners. Of utmost importance is the age-dependent partner effect: the severity of post-traumatic stress symptoms of particularly younger partners exerted a significant influence on the patients’ health-related QoL. Further research is required to identify partners of chronically critically ill patients at highest risk for post-traumatic stress [51]. Future studies should therefore allow a profound assessment of the special needs of patient-partner dyads, taking into account the impact of age and gender on the dyadic association between post-traumatic stress and health-related QoL.