Intensive care unit (ICU) delirium is an acute neurologic disorder characterized by inattention, disorganized thinking, and a fluctuating course of altered level of consciousness commonly observed during critical illness.1,2,3,4 Intensive care unit delirium is associated with a longer duration of mechanical ventilation, long-term cognitive impairment, increased healthcare costs, increased morbidity, and is an independent risk factor for mortality during hospitalization.4,5,6,7,8

ICU delirium management and prevention requires a multimodal approach including both pharmacologic and nonpharmacologic options.9 Guidelines endorse early detection of delirium using validated screening tools, such as the Intensive Care Delirium Screening Checklist (ICDSC)10 as well as the use of a multicomponent bundle to shorten or reduce delirium, such as the ABCDEF bundle (Assess, prevent, and manage pain; Both spontaneous awakening trials and spontaneous breathing trials; Choice of analgesia and sedation; Delirium assessment, management and prevention; Early mobility and exercise; and Family engagement and empowerment).3, 9

Nonpharmacologic strategies are feasible in critically ill patients including sleep promotion activities, orientation, early mobility, and activities of daily living.11, 12 Cognitive interventions, as part of a delirium prevention strategy, are specific therapies focusing on the domains of cognitive functioning impacted by delirium such as memory, attention and concentration, and executive function.1, 13,14,15 Components of cognitive interventions consist of cognitive training, stimulation, and rehabilitation.1, 14, 16, 17 However, there is a lack of evidence for the use of cognitive interventions in critically ill patients for the prevention or management of delirium.1

Occupational therapy includes interventions that treat physical, cognitive, emotional or psychological domains.18 Occupational therapists (OTs) have advanced education and training in the domains of cognitive assessment, treatment, rehabilitation and the relationship between a patient’s cognitive skills, performance, and environment and are uniquely qualified to deliver this specialized care to critically ill patients.19 Nevertheless, in most ICUs, the OT role is centred on physical mobility and rehabilitation rather than on cognitive complications of critical illness such as delirium; therefore, the potential and scope of the OT role may not be fully utilized.18

Our initial objective was to conduct a randomized controlled trial (RCT) to determine the effect of the standard of care ICU delirium prevention strategies (i.e., the ABCDEF bundle) with the addition of early OT-guided cognitive interventions titrated according to the Richmond Agitation and Sedation Scale (RASS) on delirium duration and prevalence.20 Because of the COVID-19 pandemic, the study was terminated before goal recruitment could be reached. Therefore, we report a feasibility analysis of the data. The primary outcome reported is feasibility of the protocol as measured by the proportion of eligible interventions delivered. Secondary outcomes included feasibility of goal session length from the patient’s perspective, participant clinical outcomes (delirium, cognitive status, functional status, quality of life [QOL], and ICU length of stay), and a description of methodological challenges and solutions for future research.

Methods

This feasibility RCT is reported according to the Consolidated Standards of Reporting Trials statement: extension to randomized pilot and feasibility trials, and the Guidelines for Reporting Trial Protocols and Completed Trials Modified due to the COVID-19 Pandemic and Other Extenuating Circumstances (Electronic Supplementary Material [ESM] eAppendix 1).21, 22 The study was terminated for the following reasons: 1) pandemic surges and high volumes of critically ill patients altered the baseline standard of care (i.e., usual delirium prevention practices) while strict visitation guidelines limited family presence, and 2) the projected effects of the pandemic on critical care resources and investigator/OT responsibilities to patient care (ESM eAppendix 1). Thus, the original target recruitment (N = 112) was not achieved. For these reasons, the trial was underpowered to report efficacy of the study intervention on delirium prevalence and duration, and a feasibility analysis was chosen. To mitigate the effects of lost data, funding, and time investment, we changed the trial objectives to measure feasibility of the protocol by describing the proportion of eligible cognitive interventions delivered by the OT.22 The study was conducted at South Health Campus, in a ten-bed medical/surgical ICU in Calgary, AB, Canada and was approved by the Conjoint Health Research Ethics Board, University of Calgary (REB19-1904; 10 December 2019), and the Health Research Ethics Board, University of Alberta (Pro00083213; 28 October 2018).

Participant selection and description

Informed consent was obtained from study participants or their substitute decision makers (SDM). All patients aged 18 yr or older admitted to the ICU were assessed for eligibility. Exclusion criteria included a diagnosis of primary direct brain injury, a prior diagnosis of dementia-related illness, a prior diagnosis of developmental disability, pre-existing cognitive impairment, patients requiring palliative care, an anticipated ICU stay less than 48 hr, non-English speaking patients, patients with severe communication disorders, plasmapheresis patients, patients with severe hearing or visual impairment, and patients transferred from another ICU. Based on departmental guidance to minimize exposure, patients with COVID-19 were added as an exclusion criterion in March 2020.

Randomization and blinding

Patients were randomized into the control arm (i.e., standard of care) or the experimental arm, which included usual delirium prevention practices plus OT-guided cognitive interventions (Fig. 1; see ESM eAppendices 2 and 3 for institution delirium prevention practices). A computer-generated list of random numbers with simple 1:1 randomization was used to allocate participants to the control group or the experimental group. The allocation sequence was concealed from study investigators and research assistants enrolling participants in the trial. Allocations were placed in sequentially numbered, sealed opaque envelopes, which were only opened following enrolment. Due to the nature of the intervention, it was not possible for study participants or the OT to be blinded. Study investigators and research assistants conducting outcome measures were blinded to the patient assignment.

Fig. 1
figure 1

OT-guided cognitive intervention protocol. ADL = activities of daily living; IADL = instrumental activities of daily living; ICU = intensive care unit; OT = occupational therapist; RASS = Richmond Agitation and Sedation Scale20

Full size image

Study intervention

The interventions were conducted Monday to Friday twice daily for 20 min, starting on the first weekday after enrolment, via face-to-face contact with an OT. Sessions were terminated if the patient became agitated or did not meet physiologic parameters such as ordered heart rate, respiratory rate, or blood pressure goals.

The four main components of the protocol were family or loved one-directed interventions; cognitive stimulation; cognitive training; and cognitive rehabilitation (Fig. 1). The intervention was titrated to the patient’s level of sedation using the RASS (ESM eAppendix 4).20 The following interventions were conducted at all levels of sedation: 1) family education and participation in interventions, 2) stress management strategies for the patient and family, 3) cognitive stimulation activities, and 4) informal cognitive assessment. Informal cognitive assessment was completed throughout the session, guiding the cognitive training, stimulation and rehabilitation activities chosen by the therapist based on the patient’s ability.

Data collection

Patient data were obtained from hospital electronic medical records validated for use in research (Sunrise Clinical Manager version 18.4, Allscripts Healthcare, LLC, Chicago, IL, USA; eCritical MetaVision version 5.46, iMDsoft, Tel Aviv, Israel).23 Study data were collected and managed using REDCap electronic data capture tools hosted at the University of Calgary, Calgary, AB, Canada.24

Outcome measures

Delirium prevalence and duration were measured using the ICDSC10 (ESM eAppendix 5). The ICDSC was conducted twice daily (once per shift) by the primary registered nurse (RN). All critical care RNs in our department are trained to score delirium using the ICDSC. Moreover, a provincial delirium initiative developed an ICDSC dashboard used provincially to ensure that delirium is measured consistently among RNs. Patients were considered as having delirium if they had an ICDSC score ≥ 4 at least once during their ICU stay (i.e., ever delirium). Delirium duration was measured by determining the total number of days with delirium from ICU admission to ICU discharge for each patient. Additional outcomes included cognitive function measured using the Johns Hopkins’ Adapted Cognitive Exam,25 physical function measured using the Functional Status Score for ICU,26 QOL measured using the Euro-Qual 5 Dimensions,27 ICU length of stay, days of mechanical ventilation, and hospital length of stay (see ESM eAppendix 6 for additional instruments).28,29,30,31

Statistical analysis

A previous systematic review of nonpharmacological interventions reported an average reduction of delirium prevalence of 24.7%.32 To detect this difference in delirium prevalence (assuming 49% without the intervention based on local data) at 80% power and a two-sided α = 0.05, we determined that 56 participants would need to be enrolled per arm (N = 112 total). Patient characteristics and clinical outcome data are summarized using frequency with percent for categorical data and mean with standard deviation (SD) or median with interquartile range [IQR] for continuous data.

Results

Enrolment began on 13 June 2019, but was paused on 19 March 2020 and resumed on 28 July 2020 because of the COVID-19 pandemic. Lack of capacity of the OT and investigators (who are front line providers) to provide critical care and conduct study activities resulted in termination of the study on 30 November 2020, before the original goal enrolment number (N = 112) was reached (ESM eAppendix 1).

After screening 408 patients for eligibility, 70 patients were enrolled with 37 randomized to the control group and 33 to the intervention group (Fig. 2). Forty-seven patients were enrolled before the pandemic and 23 patients were enrolled during the pandemic. One patient withdrew from the study after randomization, and 69 patients were included in the final analysis. Three patients were lost to follow up and 27 patients received the study intervention (at least one treatment). Five patients (16%) admitted on a holiday or weekend did not receive cognitive interventions because an OT was not available and were excluded from feasibility analysis. The most common reason for study exclusion was expected ICU stay less than 48 hr (n = 143/408; 35%). An average of six patients (69/11.6 months) included in final analysis were enrolled per month and an average of 35 patients were screened per study month.

Fig. 2
figure 2

CONSORT enrolment flow chart

Full size image

Differences between the control and intervention group were owing to an incomplete sample size. Baseline characteristics of study participants are shown in Table 1. The median [IQR] acute physiology and chronic health evaluation (APACHE)-II score was 22 [15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31] in the intervention group and 19 (14–24) in the control group. Patients in the intervention group were older (median [IQR] age, 63 [56–71] yr) than those in the control group (median [IQR] age, 53 [43–64] yr). The results of the study are summarized according to methodological issues reported in feasibility studies as discussed in Sosnowski et al.33 and Shanyinde et al.34 (Table 2).

Table 1 Summary of baseline patient characteristics
Full size table
Table 2 Summary of findings of methodological issues for feasibility research33, 34
Full size table

Proportion of eligible cognitive interventions delivered

The majority (110/137, 80%) of eligible cognitive intervention sessions were delivered. The mean (SD) number of sessions per patient was 4.1 (3.8). The most common reason for omitted interventions was patient test(s) or procedure(s) (19/137 sessions, 14%), patient refusal (5/137 sessions, 4%), family member refusal (2/137 sessions, 2%), healthcare provider refusal (1/137 sessions, < 1%) and one medical emergency (< 1%). Additionally, five patients (16%) enrolled in the intervention arm were discharged from ICU after a short length of stay (median ICU days, 2.7; range, 2.4–2.7) and prior to receiving any OT interventions on a weekend or holiday. There were no adverse events associated with the protocol.

Individual session length

The mean (SD) session length was 19.8 (3.1) min. The median [IQR] time to initiation of first OT intervention from the time the patient was admitted to ICU was 46 [36–60] hr. Thirteen sessions of 110 (12%) were terminated early because of patient request or fatigue (8/110, 7%), healthcare provider request (4/110, 4%); and a medical emergency (1/110, < 1%).

Description of interventions conducted

Interventions were titrated according to RASS with the majority of interventions conducted at a RASS of 0 (n = 41, 37%), followed by a RASS of -5 (n = 18, 16%).20 The OT did not observe RASS scores of +2 to +4 in patients during study intervention periods, so no interventions were conducted at those levels of agitation.

Of the 110 OT-guided sessions, the most frequent cognitive interventions conducted were cognitive stimulation exercises (91/110, 83%), family-based intervention (78/110, 71%), cognitive training (27/110, 25%), and cognitive rehabilitation (12/110, 11%) (ESM eTable). We identified eight major cognitive domains targeted during each cognitive intervention session: sensation, perception, motor skills, attention and concentration, memory, executive functioning, processing speed, and language skills.15 The most frequently targeted cognitive domains were sensation (99/110 sessions, 90%), attention and concentration (62/110, 56%), and memory (53/110, 48%).

Clinical outcomes

Overall, 36/67 (54%) of patients enrolled experienced delirium. A description of delirium outcomes, cognitive function, functional status, QOL, and length of stay is presented in Table 3.

Table 3 Clinical outcomes stratified by group assignment
Full size table

Discussion

In this feasibility RCT, we have shown that OT-guided cognitive interventions in critically ill patients are feasible. Despite the severity of illness experienced by ICU patients, the length of the interventions was also deemed feasible. Additionally, we were able to measure clinical outcomes relevant to future RCTs including delirium prevalence and duration, cognitive status, functional status, QOL, and ICU length of stay. We also identified several methodological challenges with recruitment, refusal of consent, OT weekend support, and ICU admission outcome measures. The results of this study will provide much-needed data to help plan for future large RCTs of OT-guided cognitive interventions.

The majority of eligible cognitive interventions were delivered and only 4% of sessions were refused by patients, which shows feasibility of the intervention. We found that the goal session length of 20 min was frequently achieved with a low rate of early termination due to patient fatigue. Similarly, in a feasibility trial of combined cognitive and physical therapy in critically ill patients, Brummel et al. found that cognitive interventions conducted by physicians or nurses for 20 min twice daily was feasible and safe.35 In another pilot study of non-mechanically ventilated patients, OTs conducted a longer duration of intervention (40 min twice daily for five days) and found a positive impact on delirium incidence and duration.36

Methodological challenges and proposed solutions

Participant consent

We identified several methodological challenges outlined in Table 2. Many patients or their SDM refused consent, which could be attributed to early study enrolment during periods of high stress or the belief that cognitive interventions may be burdensome during critical illness. However, similar to the concept of early mobility in ICU patients, our protocol aims were for early cognitive engagement, so we view early study enrolment as beneficial. Nevertheless, consent timing could be extended to 72 hr in future trials to ensure inclusion of patients admitted on a weekend or holiday. Finally, highlighting family participation in delirium management as part of the study design may provide family members with a sense of purpose and autonomy.

Randomization

A failure to complete goal study enrolment due to early termination resulted in imbalances between groups. The intervention group had a higher proportion of delirium, which may be attributed to the group’s higher median age and higher severity of illness. Both age and higher APACHE-II scores have been independently associated with delirium in critical illness and thus, future RCTs may stratify randomization and include a subgroup analysis.5, 37, 38 It should be noted that Brummel et al. trialled their cognitive plus physical therapy intervention protocol on patients with higher APACHE-II scores (median score, 25) than our sample intervention population (median score, 22) and yet the intervention was still deemed feasible.35 It is worth investigating the demands, time constraints, and level of coma in sicker patients (i.e., those with higher APACHE-II scores) and the potential impact on the number or duration of interventions. These variables (along with other factors) and treatment effect modification should be considered in a larger hypothesis-testing RCT.

Occupational therapist guided cognitive interventions

A mean of four sessions per patient were delivered. Nevertheless, the optimal “dose” or number of OT interventions should be investigated in future trials. The analysis of delirium outcomes, with a cost effectiveness analysis using just once daily interventions would be important for resource-limited institutions. Furthermore, the indirect supervision of other providers, such as therapy assistants, by OTs may be acceptable to maximize the number of interventions. Occupational therapist or assistant delivery of delirium prevention practices may help to ameliorate nursing workloads by allowing this specifically focused therapy to be conducted by a separate provider; therefore, future trials may measure nursing satisfaction with the protocol. Finally, determining the effects of interventions at each level of sedation on outcomes such as delirium, cognitive scores, hospital length of stay, outpatient cognitive function, or memory formation at various levels of sedation would delineate the most efficacious portions of the protocol and assist with further refinement.39

Limitations

There are several limitations to our study. This study was underpowered to determine effects of the intervention on delirium prevalence and duration, and a failure to complete target study enrolment because of early termination resulted in imbalances between the groups. A well-powered trial is needed to equally distribute confounding variables and address the protocol’s effect on delirium prevalence and duration.

The COVID-19 pandemic greatly impacted the target recruitment number and may have impacted our cognitive intervention protocol because of family visitation restrictions and continuous masking, which interfered with communication and facial recognition. Although the study was terminated before a major pandemic surge, the standard of care delirium prevention practices may have been affected by higher staff workloads. This may have included missing delirium scores in the electronic medical record. Future studies should consider using trained study team members to conduct delirium assessments to ensure complete data collection. Additionally, the use of sealed opaque envelopes to conceal randomization, while cost effective and practical, may have increased the risk of bias; therefore, future trial designs may include the use of a web-based allocation concealment method.40 Finally, this protocol was not designed for neurocritical care patients, and enrolment at a single site limits generalizability of feasibility findings to other ICUs and to patients with COVID-19.

Conclusion

Intensive care unit delirium remains a frequent and severe consequence of critical illness with the potential for long term cognitive impairments. Using a novel approach, this OT-guided cognitive intervention protocol is feasible in critically ill medical/surgical patients and can be titrated according to the deepest levels of sedation while incorporating family members. A larger RCT is required to determine the impact of such a protocol on delirium prevalence or duration.