Introduction

Critically ill patients commonly suffer profound weakness, pain, and delirium which are inextricably linked [1]. Limited physical function and adverse effects associated with immobilisation are likely to worsen if cognition is not intact [2]. Physical impairments related to these issues may last for years after discharge from the intensive care unit (ICU) [3].

Integrated pain, agitation and delirium (PAD) clinical practice guidelines for critically ill patients using pharmacologic and non-pharmacologic approaches have been developed following rigorous research and aim to prevent long-term detrimental patient outcomes [4]. To date, these approaches have not necessarily produced reductions in negative sequela for the critically ill [5]. Researchers agree that delirium and ICU-associated weakness (ICUAW) remain pervasive and under-recognised within our ICUs [6].

The ABCDE bundle focusses on a way to implement the PAD guidelines to reduce delirium and weakness related to over-sedation, prolonged mechanical ventilation and forced immobility in mechanically ventilated critically ill patients [7]. This bundle of cares as described by Balas and colleagues [7] comprises three inter-related components: coordination of awakening and breathing trials, monitoring and management of delirium, and early exercise. Clinicians are increasingly incorporating the ABCDE bundle into clinical practice [8] and investigating outcomes associated with its use in the ICU [9].

A prospective before–after study tested the effectiveness and safety of incorporating the ABCDE bundle into everyday practice [10]. Balas and colleagues [10] recruited 296 patients (146 in the pre-group and 150 in the post bundle implementation group) from five adult ICUs, a step-down unit, and an oncology special care unit in a large tertiary hospital in the USA. The authors found a number of significant patient benefits that resulted from incorporating the ABCDE bundle including decreased duration of ventilation, reduced delirium and increased episodes of mobilisation [10].

Although results of this study were promising, limitations of the study design warrant consideration. Balas and colleagues [10] acknowledged that education regarding the ABCDE protocol applied during the pre-implementation period may have influenced study results. There may have been contamination during the pre-test period with inadvertent implementation of aspects of the intervention. As with all pre-post studies, the effect of history on the internal validity cannot be disregarded. Rigorous RCTs are therefore required to adequately test the effect of this bundle. To the best of our knowledge, there are no published RCTs examining the ABCDE bundle and its impact on ICU survivor quality of life.

Objectives

In this study, we aimed to test the feasibility of conducting an RCT to examine the impact of the ABCDE bundle on quality of life in ICU survivors.

Our research question was: What is the feasibility of conducting a RCT using the ABCDE bundle compared with usual practice to improve ICU survivors’ quality of life? Feasibility outcomes included successful recruitment and retention of trial participants, intervention fidelity, identification of barriers to implementation of the intervention, and the feasibility of collecting outcome assessment data. The CONSORT extension statement checklist for pilot studies [11] was used as a guide to ensure complete and transparent reporting of our study (See Additional file 1).

Trial design

A prospective, single-centre, single-blinded, equally randomised (1:1), controlled feasibility study was conducted between April 2015 and March 2016 in an eight-bed, level two [12] ICU in south east Queensland, Australia. The protocol for this pilot trial is available on the Australian New Zealand Clinical Trials Registry website.

We were cognisant that a study involving the multicomponent ABCDE bundle is complex and that many trials involving complex interventions fail to prove a significant positive result [13]. This may be related to poor implementation of the intervention or substandard study design rather than genuine ineffectiveness. To ensure the effectiveness of our study protocol, we have utilised the principles of the United Kingdom Medical Research Council’s framework for the design and evaluation of complex health care intervention to prepare for a rigorous and appropriately powered future RCT [14].

The initial development phase of the framework required a comprehensive review of pertinent literature relating to the early rehabilitation of critically ill patients [15]. Our integrative review informed the modelling component of the framework which involved introducing the ABCDE bundle of cares into an ICU as detailed in the following method section. The results of the feasibility and piloting phase of the framework are presented in this manuscript. The outcome of this trial will determine if and how we can progress to an RCT to compare the ABCDE bundle with standard ICU practice.

Participants

Schweickert et al. [2] had reported improved functional outcomes in patients who had received physical therapy within 72 h of the initiation of mechanical ventilation. Thus, patients admitted to the ICU were included if they were aged over 18 years, had been mechanically ventilated for 48 h and were expected to require ventilation for at least a further 24 h.

Patients were excluded from the study if a premorbid functional or cognitive impairment precluded engagement in exercise. Exclusion criteria included an inability to mobilise 3 meters before the current ICU illness; were diagnosed with neuromuscular disease that could impair ventilator weaning; had suffered an acute stroke; were not for active resuscitation; had been readmitted to ICU within the current hospitalisation; or were not expected to survive the current ICU admission.

During the week, potential participants were identified during the 8 am ward round by a member of the research nursing team. Senior medical or nursing staff contacted the on-call member of the research team on the week-end if a suitable participant was identified.

Consent was initially provided by the substitute decision-maker as the participant lacked capacity related to altered consciousness, sedation and medical condition. Once able, the participant was asked to provide deferred consent to continue in the study.

Patient flow through the research process is depicted in the CONSORT flow diagram (Fig. 1).

Sample size

Thirty participants were recruited with 15 in each group. This number was consistent with recommendations for pilot and feasibility studies where samples of 10 to 20 participants per group have been deemed adequate to assess feasibility outcomes [16, 17].

Randomisation

Participants were randomised without stratification to either the intervention or control group using a computer-generated allocation sequence using permuted blocks of two (www. randomizer.org). The assigned group allocations were concealed in opaque (when held to the light) envelopes that were consecutively numbered and sealed. The envelopes were prepared by personnel not connected to the study and were stored in a secured office within the ICU. An envelope was opened by a member of the research team, and a participant was assigned to the allocated group once eligibility criteria were met and informed consent obtained.

Blinding

It was not possible to blind the research team or the participant to group assignment. However, participant outcome measurement assessors were blinded to participant assignment as they were separate to the health care providers who performed the interventions and did not normally work in the ICU.

Intervention group

The intervention group received the ABCDE bundle. The completed Template for Intervention Description and Replication (TIDieR) [18] provided in Additional file 2 provides a detailed description regarding the intervention for replication in future studies.

In brief, the ABCDE bundle education program commenced 2 months prior to study commencement. Multimodal education was delivered via unit-based presentations and simulation in order to familiarise staff with procedures and to practice emergency responses. Ongoing support was provided during the trial. Nursing, medical and allied health staff who provide exercise interventions in the trial were experienced critical care clinicians.

The ABCDE bundle was a standardised and protocolised complex care bundle that was integrated into daily patient care activities and delivered by the appropriate member of the treating multidisciplinary team at various times each day. The implementation of the bundle was facilitated by using computer software that provided staff care prompts. That is, the protocol components were embedded into the patient record computerised information system (CIS). Each component of the bundle was prescribed by nursing or physiotherapy staff following completion of a safety screen. The CIS provided an alert when bundle components were due for action. A 1-h window was provided so that other ICU activities could be considered. An electronic signature was required to confirm that each component of the bundle had or had not been completed.

The Awakening and Breathing Coordination (ABC) component of the bundle required the completion of a safety screen within the CIS to determine whether it was safe to interrupt sedation and commence a spontaneous awakening trial (SAT). If the SAT was successful, a Spontaneous Breathing Trial (SBT) safety screen was performed prior to testing pressure support ventilation or a T-piece trial. Mechanical ventilation was recommenced if the patient failed the SBT. The intensivist considered extubation if the SBT was passed within 30 min.

The Delirium monitoring and management component ensured all patients received routine pain, sedation and delirium assessment using standardised and validated assessment tools [19, 20]. Pain was assessed every 2 h with the Numeric Rating Scale (NRS) if conscious or the Critical Care Observation Tool (CCOT) if unconscious. Continuous infusions of Fentanyl and/or Remifentanil were titrated to keep NRS less than 4 or CCOT less than 3. Level of alertness was monitored every 4 h with the Richmond Agitation and Sedation Scale (RASS). Sedation was optimised by keeping the RASS between light sedation (− 2) to alert and calm (0). Propofol and/or Dexmedetomidine were recommended for the patients in the intervention group. The CAM-ICU was performed every 12 h (8 am and 8 pm) to determine if the patient was delirious.

The early exercise and mobility component of the bundle required the completion of a screen within the CIS to ensure the patient met safety criteria. If the patient did not pass the safety screen, they received passive range of motion exercise and sitting position three times a day in bed. The nurse or physiotherapist determined the patient’s capacity for independent movement. Patients progressed through four levels of progressive activity, receiving the highest level they could manage. The nurse created a nursing order in the CIS which would appear in the CIS at the appropriate time throughout the day. The nurse was provided a 1-h window either side of the prescribed time to allow other patient therapies to occur.

Nursing, medical and allied health staff involved in providing interventions and those involved in performing outcome measurements were experienced critical care clinicians and were deemed competent to perform the rehabilitation strategies. Safety guidelines especially with regard to management of endotracheal tubes and invasive lines during exercise and mobility were provided to all ICU staff involved in patient care. Safety strategies were effective as no adverse events occurred during the trial.

Control group

The control group received standard medical, nursing and allied health care with routine pain, sedation and delirium assessment using the Numeric Rating Scale (NRS) if conscious or the Critical Care Observation Tool (CCOT) if unconscious, the Richmond Agitation and Sedation Scale, and the Confusion Assessment Method-ICU (CAM-ICU).

Interruption of sedative and opiate infusions was performed at the discretion of the ICU consultant on duty and would generally occur following the morning handover round (8.00 am). The decision to progress to cessation of sedation was based on information regarding the patient’s clinical progress during the previous 12 to 24 h. There were no standardised procedures to guide spontaneous breathing trials. Nursing staff would closely monitor the patient’s neurological, haemodynamic and respiratory status. The intensive care specialist would make the decision whether to progress to extubation or to resedate the patient.

There were no standardised approaches to the provision of exercise and mobility at this site. Nursing staff did not routinely provide patient exercise. Physiotherapists provided physical exercise for each patient on an ad hoc basis. The degree of exercise ranged from no exercise to passive range of motion to sitting out of bed and was delivered at the discretion of individual physiotherapists.

Primary outcome assessment data

The Functional Independence Measure (FIM) quantified the participant’s functional and cognitive status at both ICU and hospital discharge. This tool is validated for use in the critically ill population [21, 22] and provides reliable information regarding patient functional change during rehabilitation across various hospital settings [23].

Secondary outcome assessment data

The Physical Function ICU Test-scored (PFIT-s) is a reliable and responsive measure of the physical function and potential patient physical limitations of critically ill patients [24]. This test was developed in Australia [24] and has recently been validated for patient cohorts in the United States (US) requiring mechanical ventilation for 4 days or longer [25]. In our study, the PFIT-s was used in addition to the FIM to provide a specific functional test for ICU patients at ICU discharge.

Physiotherapists who performed the PFIT-s and FIM assessments and Occupational Therapists who performed the FIM assessments were trained and credentialed in the use of the tools. Assessments were timed to be performed within 24 h of both discharge from the ICU using FIM and PFIT-s and discharge from the acute hospital using FIM.

The Short-Form (36) Health Survey (SF-36v2™) provided a baseline and post discharge measure of participants’ health-related quality of life (HRQOL). The SF-36v2™ is a validated and reliable tool [26] and has been used to assess the health status of ICU patients prior to hospitalisation and after discharge [27]. Study data collectors delivered the questionnaires during one-on-one personal interviews with participants or their next of kin when an inpatient of the ICU and by telephone interview at the 90-day post discharge follow-up. Responses provided by next of kin to HRQOL questions within the SF-36 have been validated [28] with a significant correlation shown between responses provided by patients and their next of kin [29].

Whilst the main purpose of the study was to test feasibility, the research team also collected health outcome data including duration of stay in the ICU and hospital, duration of mechanical ventilation, and mortality.

Trial feasibility - Participant recruitment and retention rates

Enrolment logs were recorded for all patients who met eligibility criteria. Pre-screen failure logs were kept for patients who met the inclusion criteria but were unable to be enrolled. Reasons for pre-screen failure were recorded and data entered into an excel spreadsheet. Recruitment success was defined as 80% of eligible participants agreeing to be enrolled in the study. Successful retention was defined by less than 10% attrition rate for those participants who had survived to the 90-day post discharge assessment.

Trial feasibility - Intervention fidelity

The bedside nurse recorded each component of the ABCDE bundle as it was delivered within the CIS. The research assistants retrospectively recorded this data on a case report form. Successful adherence to the protocol was defined as the administration of the entire prescribed ABCDE bundle on at least 80% of ventilated days.

Awakening and Breathing Coordination: Participant records were checked for the presence of completed Spontaneous Awakening Trial (SAT) and Spontaneous Breathing Trial (SBT) safety screen, and whether a SAT and SBT was performed if appropriate. Full compliance was considered if a SAT and SBT if appropriate had been administered each day.

Delirium monitoring and management: Assessment of sedation and delirium status using the Richmond Agitation Sedation scale [20] and the CAM-ICU [19] were recorded.

Early Exercise: Protocol adherence was determined when the participant performed an appropriate level of exercise each day.

Trial feasibility - Barriers to the implementation of the intervention

Reasons for not completing components of the ABCDE bundle were recorded by the bedside nurse. These records were stored at the bedside until collection by the research team at participant discharge from the ICU.

Trial feasibility - Adverse events

Participants who suffered an adverse event were to be immediately reviewed by the intensivist and the event recorded in detail. Adverse events were to be reviewed by an independent safety monitor at the recruitment of participant 10 and 20 or before if indicated.

Trial feasibility - Collection of outcome data.

Primary and secondary outcome measures were performed at four time points during the trial. A baseline assessment occurred at enrolment into the study, with further assessments performed within 24 h of ICU discharge, hospital discharge and the final Health-Related Quality of Life (HRQOL) assessment performed at 90 days post discharge from hospital.

Statistical analysis

The feasibility study was designed to test our procedures and estimate the proportions of our participants who would meet our feasibility objectives in a powered RCT.

Demographic and participant characteristics were summarised by using mean and standard deviation for continuous variables and number and percent for categorical variables. Descriptive statistics were used to report feasibility outcomes. Participant recruitment and retention rates, the adherence to the trial intervention, barriers to the implementation of the intervention, and the feasibility of collecting outcome assessment data were summarised and reported as frequencies and proportions or as free text. Functional and quality of life outcomes were expressed as mean and standard deviation.

Between groups inferential comparisons were not performed as the study was not powered for this analysis. All analyses were based on the intention-to-treat principle using complete case data only and were performed using IBM SPSS software, version 21.0.

Eligibility

Between May and December 2015, 433 patients were screened for suitability for inclusion in the trial. A total of 37 patients met the inclusion criteria of which three were not enrolled as they were outside the time for recruitment, one declined consent, one was homeless and could not be followed up, an interpreter was unable to be found for a non-English speaking patient and one patient was related to a research team member. A summary of patient eligibility is illustrated in the CONSORT flow diagram (Fig. 1).

Recruitment and consent

Recruitment and consent processes performed by dedicated research personnel were deemed successful. Ultimately, 30 eligible patients (81.1%) consented and were randomised either into the intervention (n = 15) or control group (n = 15).

Randomisation procedures

Randomisation processes worked smoothly with an unpredictable assignment to comparison groups. However, there was a noted imbalance in demographic data with a disproportion in age and gender assignments between the groups.

Blinding procedures

The outcome assessors who completed the FIM and PFIT-s remained blinded to group assignment throughout the course of the trial. It was not possible to blind the research team, the treating multidisciplinary team or the patient to group assignment.

Adherence to the intervention

Adherence to this complex intervention exceeded the defined success rate of 80% of sessions received each mechanically ventilated day.

Acceptability of the intervention–barriers

Outcome assessment

A total of 25 (83.3%) FIM assessments were performed at ICU discharge and 24 (80%) at hospital discharge.

Reasons for non-completion of the assessment included mortality, with a total of three deaths at ICU discharge and a further death at hospital discharge; two participants were transferred to other facilities, one participant refused the assessment at hospital discharge and one participant remained a hospital in-patient at the time of project completion.

Twenty-four participants (80%) completed the PFIT-s assessment at ICU discharge. Reasons for non-completion were mortality (n = 3, 10%); participant was transferred to another facility (n = 2, 6.7%) and one participant refused to be assessed (n = 1, 3.3%) due to fatigue.

HRQOL was assessed with the SF-36 v 2.0 at baseline and again at 90 days post discharge from hospital. Twenty-nine (96.7%) baseline assessments were completed during the initial ICU admission. One participant died during the admission, and it was deemed inappropriate to seek the baseline assessment. At 90 days post discharge, 23 (76.7%) follow-up assessments were complete. A total of 5 (16.7%) participants died during the trial period related to their clinical condition, whilst a further 2 (6.7%) refused to participate. Despite longer length of stay and duration of ventilation, the intervention group mean scores indicated better functional and quality of life results at discharge. However, the sample size was too small to provide reliable interpretation.

Selection of most appropriate outcome test

The outcome assessment tests provided robust clinimetric measurements of the participant’s physical, functional and HRQOL.

Trial feasibility—retention

Five (16.7%) of the 30 participants had died within 90 days post hospital discharge. Twenty-three (92.0%) of the remaining 25 participants, were successfully followed up. Two participants were lost to follow-up due to non-contactability (n = 1) and withdrawal of consent (n = 1).

Synergy of the components of the trial

The study protocol was robust evidenced by successful implementation of a new complex intervention at the study site which met study objectives. This suggests that the study protocol requires very minor adjustment to provide evaluation of the ABCDE bundle in an appropriately powered trial.