The following questions were assessed: 1) do the eligibility criteria provide a representative sample of the emergency abdominal surgery population? 2) will recruitment rate be sufficient to complete the trial within reasonable time? 3) is data collection feasible? 4) are the postoperative physiotherapy protocols feasible to deliver? 5) is there a significant difference in treatment dosage between groups? 6) are the intervention arms safe compared to the benchmark of adverse event rates during physiotherapy sessions to critically unwell patients?
Design and setting
From June 2015 to October 2016, we conducted a prospective single-centre double-blinded (patient and assessor) parallel-group randomised controlled trial at a tertiary government-funded regional hospital in Australia. Trial conduct, processes, procedures, and treatment protocols were piloted for feasibility and safety to inform modifications required prior to rolling into the intended full-scale multicentre phase of ICEAGE. The pilot phase of ICEAGE was granted local ethical approval (Human Research Ethics Committee (Tasmania) Network, Tasmania, Australia (H0013666) and was prospectively registered on the 8th of April 2015 in the Australian and New Zealand Clinical Trial Registry (ANZCTRN 12615000318583)25. This manuscript is reported in accordance with the CONSORT guidelines for randomised pilot and feasibility trials26.
Eligible participants were adults (≥18 years) admitted for emergency abdominal surgery requiring an open incision above, or extending above, the umbilicus. Patients were ineligible if they did not provide consent to participate, were preoperatively unable to walk for one minute without a seated rest, unable to understand English, received preoperative education and breathing exercise training from a physiotherapist13, had undergone elective abdominal surgery in the previous seven days, assessed by senior medical staff as approaching imminent death within 48 hours of surgery, or had medical orders not to participate in early active rehabilitation. Patients were also excluded if the research team were unable to approach the eligible patient within 48 hours of surgery.
New admissions to surgical wards and the intensive care unit were screened for eligible patients by the research team on weekdays. Eligible patients were provided with a single standardised physiotherapy session at the earliest possible occasion within the first two days after surgery. Following this session, eligible patients were approached at the earliest convenient time by the chief investigator (IB) or principal site investigator (KS). Written and verbal information about the clinical trial was provided and signed consent gained if the patient was willing to participate. If patient was incapable of consenting (i.e sedated, delirium, severe pain, cognitively impaired) the next of kin was contacted.
Following consent being gained, the treating physiotherapist opened a sequentially numbered concealed envelope pre-prepared by an independent administration assistant. Allocation order was by computer generated blocked random number table (1:1). The participant’s name was written on the envelope to record that recruitment occurred in order of eligibility. Participants, hospital doctors, nurses, and assessors were masked to group allocation to group allocation. Participants were withdrawn if they withdrew consent.
Prior to recruitment all participants received an initial standardised physiotherapy treatment session on the first available occasion within the first two postoperative days (Figure 1). This session consisted of education on PPC prevention with early ambulation and breathing exercises, assisted mobilisation out of bed, and a single session of coached deep breathing and coughing exercises. Participants were encouraged to continue performing their own breathing exercises every hour during the day until they were frequently ambulant. An information booklet was provided to reinforce the information and coaching provided. Education content, information booklet, and breathing exercises were based on those previously shown to be highly valued by patients27 and effective in halving PPC incidence after elective abdominal surgery13. Following recruitment into the trial, participants were assigned to receive ongoing standard-care physiotherapy (control) or intensive physiotherapy (intervention).
Standard-care physiotherapy (Control): Following the standardised initial physiotherapy session, control group participants were provided with once-daily assisted ambulation with a physiotherapist or with a therapy assistant working under the direct supervision of a physiotherapist. Assisted ambulation was delivered using a protocol utilised previously in elective abdominal surgery13, 27-29 (Figure 1). The aim was to assist the participant to walk away from the bedside for at least three minutes to a target of 10 minutes, but no longer than 15 minutes. The total assisted ambulation per-protocol time over the first five days was 50 minutes. Assisted ambulation sessions were ceased once a threshold score determining readiness to discharge from physiotherapy was met30.
Intensive physiotherapy (Intervention): In addition to standard-care physiotherapy, intervention group participants received additional coached breathing exercises, functional rehabilitation exercises, and education (Figure 1). Participants were provided with twice-daily coached deep breathing and coughing sessions for the first two days and once-daily assisted ambulation and targeted functional and strengthening exercises. Exercise prescription was according to a descending scale dependent on patient acuity, conscious status, and functional capacity. Exercises were individualised to the participant’s functional status and progressed daily by the physiotherapist. Total daily exercise therapy duration target was at least 30 minutes. The target total assisted rehabilitation per-protocol time over the first five days was 150 minutes. The ICEAGE rehabilitation protocol was modelled on previous critical care rehabilitation trials31,32. Specific exercises, repetitions, intensity, and sets were chosen pragmatically by the treating physiotherapist using their clinical judgement. Additional education regarding preventing PPCs, improving recovery, and prompts to continue with breathing exercises and physical activity were provided daily at the discretion of the treating physiotherapist. Intensive physiotherapy was provided each day for a minimum of the first five postoperative hospital days and then continued until a threshold score was met30, or day of hospital discharge, whichever came first.
Feasibility of recruitment
Recruitment feasibility was measured with the following metrics; number of patients having emergency abdominal surgery, number fulfilling eligibility criteria, number of occasions and reasons for when eligibility criteria were queried by the investigating team as possibly considered unsuitable/unrepresentative and recommended for modification, number and reasons for not being able to approach patients to gain consent to enrol in study, number of patients approached within the first two postoperative days, number who consented, number who withdrew.
A convenience sample of 20 consecutive participants provided feedback regarding their experience of the recruitment process. These participants were interviewed by the chief investigator (IB) within seven days of surgery, face-to-face or by telephone, using a semi-structured interview. Interviews were collated and interpreted thematically and by proportional responses.
Feasibility of data collection
Number lost to follow-up at the 14-day primary trial outcome of PPC and 90-day secondary trial outcome of health-related quality of life.
Feasibility of treatment protocols
Feasibility of delivering the protocol was assessed by recording reasons for not being able to provide treatment protocols within the first five postoperative days and number of protocol violations.
Separation between groups
This was assessed using time from surgery to first successful assisted ambulation event >3 mins, total duration of exercise sessions provided over the first five postoperative days, exercise time provided as a proportion of total protocolised time (control group 50mins; intervention group 150mins), and total number of coached breathing exercise sessions provided within the first three postoperative days. Time to receive the initial physiotherapy assisted ambulation session was intended to be similar whilst a separation between groups was intended for number of coached respiratory sessions and duration of assisted physical activity over the first five postoperative days.
Safety of treatment protocols
An adverse event was defined as a physiological event that occurred in direct response to delivering the treatment protocol. These included, a) symptomatic blood pressure or heart rate change 20% from resting heart rate, new cardiac arrhythmia, and drop in oxygen saturation > 10% from baseline which fail to recover within two minutes of ceasing activity, b) requirement to increase sedation or inotropic support, c) line detachment, d) fall, or e) or severe nausea. The proportion of adverse events caused directly by the ICEAGE treatment protocols was compared to the adverse event rate of 2% in surgical critical care patients provided with assisted early mobilisation32.
No a priori hypotheses of trial conduct success or failure were set. The decision to progress with the trial was made pragmatically based on the analysed pilot data and anticipated prospective recruitment rates given appropriate modifications, if necessary, to the study protocol.
For the pilot phase of ICEAGE, between-group estimates of treatment effect to the trial’s primary outcome, PPC, and key secondary outcomes were not undertaken unless there was a decision to cease the trial for futility. This was to deliberately avoid a possible detection bias that could influence ongoing trial conduct given a decision to proceed with the trial.
The primary aim of this pilot phase was to investigate trial feasibility, safety, and trial processes rather that to estimate a treatment effect. The sample size for the ICEAGE pilot phase was pragmatically set at 10025, 33. This was the expected throughput of eligible patients at the participating centre within a 12-month period. After 12-months this size sample had not been attained. The pilot phase continued for a total of 15 months until a revised target sample of 50 was achieved.
Data was analysed with intention-to-treat principles. Recruitment data was reported in numbers and proportions. Imbalances between-groups in baseline measures was assessed to determine covariates for inclusion in comparative testing for group treatment separation outcomes. Continuous measures were compared with independent samples t-tests or Mann Whitney U test depending on normality (Shapiro-Wilk test). Fishers exact test was used for proportions. With equivalence established, outcomes were assessed using simple main effects comparisons between-groups with the same tests. Between-group differences in the duration of supervised physical activity each day over the first five postoperative days was analysed using two-way mixed ANOVA with non-parametric data transformed with square root functions, Bonferroni adjustment for 95% confidence intervals, and p-values estimated with Greenhouse-Geisser adjustment for non-sphericity. Simple main effects for differences between-groups on each postoperative day separately were analysed with one-way ANOVAs. Values are reported as means with standard deviations (SD) for normally distributed variables, medians with interquartile ranges for non-parametric continuous variables, and frequencies and percentages for categorical values. P values are two-sided with <0.05 considered statistically significant.