Improving Sedation Practice in Critically Ill Adult Patients Using a Co-designed Sedation Protocol

Overall Aim and purpose: • The overall aim is to develop a co-designed analgosedation protocol for adult extracorporeal membrane oxygenation (ECMO) patients and test its acceptability with staff and patients. Objectives: Stage 1 * To describe and explore current practices of analgosedation (in terms of opioid and sedative choice, dose and assessment) and use of analgosedation protocols, and sedation holds in ECMO and non-ECMO ICU adult patients. * To describe patient outcomes (including those recommended by my PPIE advisory group): ICU mortality, ICU length of stay, and delirium incidence. * To determine clinical equipoise with the direct clinical care team on using a co-designed analgosedation protocol in a feasibility cluster randomised control trial. Stage 2 • To determine anticipatory acceptability of the co-designed analgosedation protocol. Study design: Stage 1: multi-centre observational cohort study reviewing analgosedation practices in UK ECMO centres comparing ECMO to non-ECMO patients The investigators will conduct a prospective multi-centre observational cohort study enrolling UK adult ECMO centres. The investigators aim to recruit all ECMO centres to participate. The investigators are recruiting sites via a well-established national pharmacist and physician ECMO groups. Analgosedation data will be collected for all ECMO patients and a representative sample of non-ECMO ICU patients either in the same unit or another representative ICU in ECMO centres with multiple ICUs. The investigators will collect outcome data to patient discharge from ICU. The investigators will organise meetings with consultant doctors at St Thomas' and Royal Brompton Hospitals to determine clinical equipoise of using a co-designed analgosedation protocol in a feasibility cluster randomised controlled trial versus no protocol. Data collection: Data to be collected Patients: Sites will record data using a printed case report form. The direct care team will de-identify identifiable patient data before uploading it on to the electronic case report form (e-CRF). The investigators will develop a e-CRF (in collaboration with with King's College London) to collect de-identified patient demographic characteristics, analgosedation management and patient outcomes. Sites will then transfer de-identified data on to the electronic data capture system (REDCap) and lock when completed to ensure patient confidentiality. The investigators will train the direct clinical care team, clinical pharmacists and research nurses at participating sites on data collection. Data collected will include baseline characteristics (age, gender, weight, height, APACHE II score and past medical history), continuous analgosedative (opioid and/or sedative) IV infusion hourly doses (excluding bolus doses), and enteral opioid and sedatives doses during a 5-day observation period documented in the patient medical record. The investigators will also document type and numerical values of pain score, sedation scores (Richmond, Agitation and Sedation Score (RASS)), delirium presence (assessed by Confusion Assessment Method for the Intensive Care Unit, CAM-ICU) and duration during the 5-day collection period Outcome data including ICU length of stay, ECMO duration and ICU mortality will be collected. Consultant doctors: The investigators will collect information on the number of consultant doctors who attend the meeting to review clinical equipoise, the total number of consultant doctors employed within each service, and the number of consultant doctors who confirm clinical equipoise. Data analysis: Quantitative analysis: The investigators will calculate the median (interquartile range) of daily (24 hour period) dose of each analgosedation drug. The investigators will convert all opioids (enteral and intravenous) to morphine dose equivalents; benzodiazepines (enteral and intravenous) into midazolam dose equivalents using opioid and benzodiazepine conversion tables. The investigators will calculate median daily dose by dividing the sum of all analgosedation doses received during the data collection period by the number of days of collected data. To compare analgosedation doses and patient outcomes of ECMO versus non-ECMO ICU patients, the investigators will test data for normality and use summary measures for central tendency (mean, median) and variability (standard deviation, interquartile range, minimum, maximum). The investigators will present categorical data as counts (%). Comparisons will be made using parametric t-test or non-parametric Wilcoxon rank sum test depending on data distribution. The investigators will use univariable and multivariable linear regression with a priori selected co-variates to examine the association between patient factors and analgosedation dosing including use of ECMO. Data will be analysed using R Studio. Stage 2 Study design The investigators will conduct a mixed methods study to determine anticipatory acceptability of a co-designed protocol using the Theoretical Framework of Acceptability questionnaire informed by the Theoretical Framework of Acceptability (TFA). Data to be collected The investigators will develop an interview guide addressing the seven domains (affective attitude, burden, ethnicality, perceived effectiveness, intervention coherence, self-efficacy and opportunity costs) of the TFA to qualitatively assess acceptability of the co-designed protocol. The investigators will administer the TFA Questionnaire at the start of the interview to obtain a quantitative score of acceptability. Interviews will be on-line (each lasting 45 minutes long) using Microsoft Teams, digitally recorded and transcribed by a professional transcription company. Patients and family (NHS and non-NHS participants) The investigators will collect data including name, age, contact details, gender, ethnicity, and job status for patients, family taking part. Healthcare staff (NHS participants) The investigators will collect data including name, age, contact details, gender, ethnicity, professional role, location of job, length of time in current position, and duration of intensive care career to date for healthcare staff taking part. Data analysis Qualitative analysis: The investigators will use The Framework Method using the TFA concepts as the backbone of the framework. The investigators will dual-code transcripts coding data within the domains of the TFA. The investigators will group and mark codes from the transcripts with one or more short descriptors to organise the raw data. The codes will then be grouped, summarised and categorised according to their meaning, similarities and differences. The investigators will use the data management tool NVivo to support the coding and the analysis of the qualitative data. The investigators will use the consolidated criteria for reporting qualitative studies (COREQ) validated checklist to ensure transparent reporting, rigor and credibility. Quantitative analysis The investigators will analyse the scores from the TFA questionnaire computing the total mean score of the 7 domains as a single acceptability score. Following analysis of both qualitative and quantitative datasets, the investigators will integrate both using established methodological triangulation to provide a mixed methods interpretation. This involves comparing both qualitative and quantitative datasets to confirm areas of convergence and divergence in the two datasets. Triangulation will help to increase validity of the research findings and explain research biases. The investigators will integrate findings using a convergence-coding matrix.