Psychological evaluation and support in COVID-19 critically ill patients: a feasibility study

To the Editor,

About 20% of acute respiratory distress syndrome (ARDS) survivors suffer from anxiety, depression or post-traumatic stress disorder (PTSD) [1] with a negative impact on long-term quality of life [2]. Higher rates—38.8%—of psychological sequelae were described in Middle East and Severe Acute Respiratory Syndromes’ outbreaks [3]. Recent data described about 40 to 48% of COVID-19 critically ill patients with post-intensive care disorder or acute stress disorder [4, 5]. The aim of this pilot study is to assess the feasibility of an early psychological evaluation and sustained support in COVID-19 patients admitted to intensive care unit (ICU) and describe their mental health outcomes during a 6-month follow-up. Every COVID-19 survivor was evaluated by a trained clinician psychologist in ICU at invasive ventilation weaning or when conversation was feasible for patients receiving high-flow oxygen. Psychological support was performed as needed according to standard care. Clinician psychologists met the patient and identified if psychological distress symptoms were present (i.e. anxious or depressive symptoms, sleep disorder, …). They met the patient as often as required for supportive interventions (to help patients to speak about their emotions and/or fears) and also explained to the patients the care they were given while being sedated. Clinician psychologists systematically met patients or called them by phone as preferred by the patient at day 7, week 6, 12 and 24 after ICU discharge and standardized evaluation occurred at those time points with psychometric evaluation (Hospital Anxiety and Depression Scale (HADS) Posttraumatic Stress Disorder Checklist for DSM-5 (PCL-5) and Insomnia Severity Index (ISI)). Univariate linear mixed regression models were used to identify predictors of mental health in follow-up. Thirty-seven patients were included (81% of men, mean age of 59 (± 11) from March to June, 2020 (Table 1). Three quarters of them (28/37) were intubated for an average length of 40 days (± 7). At discharge from hospital, 20/34 (59%) of them returned home while 14/34 (41%) went to rehabilitation centre. At day 7, 32 patients were evaluated: 5/32 (16%) had significant depression symptoms, 4/32 (13%) significant insomnia symptoms and 3/32 (9%) significant anxiety symptoms and at week 6, they were, respectively, 2/17 (12%), 3/17 (18%) and 2/17 (12%). At week 12, they were, respectively, 5/24 (21%), 3/22 (14%), 3/23 (13%), and 2/25 (8%) to have significant depression, insomnia, anxiety or PTSD symptoms and at week 24, respectively, 2/19 (11%), 2/19 (11%), 4/18 (22%), and 1/18 (6%) (Table 1). Insomnia and anxiety scores did not vary over time (Kruskal–Wallis, respectively, p = 0.76 and p = 0.95), whereas depression scores decreased at week 24 (Kruskal–Wallis, p = 0.04). Cumulative dose of midazolam, cumulative dose of clonidine and length of ICU stay were associated with insomnia in follow-up (respectively regression coefficient β: 1.14 [0.44–2.39], p = 0.007, β: 1.29 [0.24; 2.36], p = 0.02 and β: 1.02 [0.14; 1.90], p = 0.025) (Table 2). Cumulative dose of clonidine was associated with depression in follow-up (β: 0.57 [0.11; 1.03], p = 0.019). Severity of ARDS could be associated with occurrence of PTSD in follow-up despite non-statistical significance (p = 0.059) (Table 2). In conclusion, this study shows that, even in the context of a pandemic situation, it is possible to provide an early and sustained psychological support in critically ill patients. We report lower rates of post-intensive care psychological sequelae than what has previously been reported [4, 5] but the absence of control group prevent from drawing firm conclusions about the impact of psychological intervention. Nevertheless, these results strongly encourage future large randomized controlled studies to assess the efficacy of early psychological evaluation and personalized support in critically ill COVID-19 patients.

Table 1 COVID-19 critically ill patients’ socio-demographic, comorbidities and care characteristics

Table 2 Univariate mixed model with random intercept of factors associated with the occurrence of psychological sequelae during follow-up