Background: Endotracheal suction (ETS) is one of the most common airway interventions performed on children requiring invasive mechanical ventilation. It is an essential airway clearance strategy to prevent retained secretions from occluding the endotracheal tube (ETT) or causing pulmonary complications such as diffusion impairment. More than 40% of children admitted to the paediatric intensive care unit (PICU) will require ETS, which in Australia is predominantly a nursing responsibility. Despite the ubiquity of the procedure, adverse events (AEs) such as oxygen desaturation occur in approximately one quarter of all ETS events. Additionally, ETS practice is varied across clinicians, intensive care units and healthcare institutions. Complications arising from the inappropriate application or failure to apply interventions, such as normal saline instillation (NSI) or lung recruitment manoeuvres (RMs), may contribute to significant patient harm and result in a longer duration of mechanical ventilation or PICU admission. Evidence-based nursing practice is vital for the prevention of AEs associated with ETS and to improve patient outcomes. Aims and objectives: The overarching aim of this PhD research was to investigate two ETS interventions— NSI and RMs—which may optimise ETS practice and outcomes in mechanically ventilated children. Four objectives guided the three research phases: 1. Identify current ETS practice and establish a baseline of NSI and RM application.; 2. Determine risk factors associated with ETS AEs.; 3. Explore PICU nurses’ experience with NSI and RMs with ETS.; 4. Evaluate the feasibility of conducting a full-scale definitive factorial randomised controlled trial (RCT) of NSI versus no NSI, and RM versus no RM, using pre-defined feasibility criteria. Two systematic literature reviews and a critical appraisal of current ETS clinical guidelines were also undertaken to determine the current strength of evidence supporting ETS interventions and current practice recommendations. This preliminary work was necessary to identify the gap in the existing evidence base and to clearly identify the scope and aim of the PhD program of research. Design: The research was underpinned by the Medical Research Council’s framework for the evaluation of complex interventions and consisted of three phases: a prospective clinical audit of current ETS practice; a qualitative exploration of nurses’ experiences using NSI and RMs with ETS; and a pilot factorial RCT comparing NSI versus no NSI, and RM versus no RM for paediatric ETS. Phase 1 Research questions: 1. What is current practice for NSI and RM use with ETS?; 2. How frequently do AEs occur with ETS?; 3. What risk factors are associated with ETS AEs in the PICU population? Setting: PICU of a Queensland tertiary hospital. Sample: 100 children, aged less than 18 years, requiring ETS during an episode of invasive ventilation. Measurements: The main outcome was a composite measure of any ETS related AE. Data on patient and suction variables (indication for ETT suction, number of suction episodes per invasive ventilation episode, indication for NSI and NSI dose), including potential predictive variables (age, Paediatric Index of Mortality 3 severity of illness [PIM3], NSI, positive end-expiratory pressure [PEEP], hyperoxygenation), were collected. Main result: A total of 955 suction episodes were recorded in 100 children. AEs occurred in 211 (22%) ETT suctions. Suction related AEs were not associated with age, diagnostic category or index of mortality score. Desaturation was the most common AE (180 suctions; 19%), with 69% of desaturation events requiring clinician intervention. Univariate logistic regression showed the odds of desaturation decreased as the internal diameter of the ETT increased (odds ratio [OR] 0.59; 95% confidence interval [CI] 0.37-0.95; p = 0.02). Multivariable modelling revealed NSI was significantly associated with an increased risk of desaturation (adjusted OR [aOR] 3.23; 95% CI 1.99-5.40; p<0.001) and the occurrence of an AE (aOR 2.76; 95% CI 1.74 -4.37; p<0.001). Presuction increases in the fraction of inspired oxygen (FiO2) were significantly associated with an increased risk of experiencing an AE (aOR 2.0; 95% CI 1.27 - 3.15; p = 0.003). Phase 2 Research questions: 1. What are nurses’ experiences with using NSI and RM with ETS in their practice?; 2. What are the clinical indicators that influence nurses’ use of NSI or RMs with ETS? Setting: PICU of a Queensland tertiary hospital. Sample: 12 registered nurses. Study design: A descriptive, exploratory study was conducted using semi-structured interviews. Interview data were analysed using inductive thematic analysis. Main findings: Variability in nurses’ ETS practice was evident. Thematic analysis revealed three themes: patients’ clinical presentation, clinician judgement and unit practice norms. In the absence of evidence-based clinical guidelines, nurses relied on knowledge derived from clinical experience and the local setting to guide NSI and RM intervention decisions. Participants reported uncertainty regarding ETS best practice and perceived the lack of research evidence as a barrier to making informed clinical decisions at the bedside. Phase 3 Research questions: 1) Is it feasible to conduct a factorial RCT to test the effectiveness and safety of NSI and RMs with ETS in mechanically ventilated children?; 2) In mechanically ventilated children requiring ETS, is (i) NSI superior to (ii) no NSI to prevent ventilator associated pneumonia (VAP) and improve measures of gas exchange, lung function and impedance measures?; 3) In mechanically ventilated children requiring ETS, is (i) RM superior to (ii) no RM to prevent VAP and improve measures of gas exchange, lung function and impedance measures? Setting: PICU of a Queensland tertiary hospital. Sample: 60 children who were less than 16 years of age and required ETS during an episode of invasive mechanical ventilation. Study Design: Single-centre, pilot factorial RCT. Interventions: Participants were first randomised to receive, either: - NSI 0.1 ml/kg (maximum 2 ml); or - no NSI with ETS. In a second randomisation, participants were allocated to receive either: - RM, an increase in PEEP by a factor of two, for two minutes (maximum PEEP 18 mm Hg); or - No RM following ETS. Primary and secondary outcomes: As per the trial protocol (Schults et al., 2018), the primary outcome was the feasibility of a definitive factorial RCT. Feasibility was determined through composite analysis of eligibility, recruitment, retention, protocol adherence and missing data, and sample size calculations based on effect size estimates (Lancaster, Dodd, & Williamson, 2004; Thabane et al., 2010). Secondary outcomes were ratio of oxygen saturation (SpO2) to FiO2, dynamiccompliance (Cdyn, ml/cmH2O), end-expiratory lung impedance (EELI), tidal impedance variation (VARt) and VAP (clinically suspected, not confirmed microbiologically, which was defined in accordance with best practice literature; (Centers for Disease Control and Prevention, 2004a; Foglia, Meier, & Elward, 2007). Data analysis: Comparability of groups at the baseline was assessed using clinical parameters and reported using descriptive statistics. Means and standard deviations were used to report normally distributed continuous data; medians and interquartile ranges were used for interval data that could not be approximated with a normal distribution. Feasibility was reported descriptively against predefined criteria. Incidence rates of VAP per 1000 ventilator days and 95% CIs were calculated using Poisson regression. Interaction effects between NSI and RMs were investigated in the regression models. In the absence of significant interaction, Poisson regression models were constructed with treatment as the main effect and the pre-suction measurement of the outcome included as a covariable. Secondary outcomes measured using interval data (Sp02/Fi02, Cdyn) were analysed, adjusting for baseline measurement using linear regressions in a pairwise sequential manner to compare NSI and RMs (Bland & Altman, 2011). To assess EELI and VARt, we used a mixed effects linear regression model with time and intervention (NSI or RM) included as main effects and a time by intervention interaction. Patient was included as a random effect to account for the repeated measures nature of the data. Analyses were undertaken on an intention-to-treat basis. Data were analysed using StataSE v14.1 (StataCorp Pty Ltd, College Station, Texas). The Type I error was set at 0.05. Main findings: Recruitment, retention and missing data feasibility criteria were achieved, with 90% of patients approached agreeing to enrol and no patient lost to follow up. Eligibility and protocol adherence criteria were not achieved, with 3881 screened (2521 non-ventilated), 818 (21%) patients eligible and 58 enrolled. Cardiac surgery was the primary reason for exclusion (479/818; 59%), followed by readmission (123/818; 15%). Approximately 30% of patients in the RM and no NSI groups had at least one episode of non-adherence; good protocol adherence was achieved in the NSI and no RM groups. Participants were, on average, 11 months old (interquartile range [IQR] 2–43), admitted for a respiratory diagnosis (27/58; 47%), with a median PIM3 of 1.0 (IQR 0.45–3.38). Participants in the NSI group had a reduced incidence rate (IR) (3%; IR 13 per 1,000 ventilator days, 95% CI 1.89–95.60) and were eight times less likely to acquire VAP when compared with the no saline group (14%; IR 109 per 1,000 ventilator days, 95% CI 40.88–290.23). However, this did not reach statistical significance (Incidence Rate Ratio 0.12, 95% CI 0.01–1.10; p = 0.06). When compared to no RM (IR 26 per 1,000 ventilator days, 95% CI 6.63–106.03), the application of an RM resulted in a decreased risk of developing VAP; however, this was not statistically significant (IR 84 per 1,000 ventilator days, 95% CI 27.21–261.59; incident rate ratio [IRR] 0.31, 95% CI 0.05–1.88, p = 0.20). RMs were found to result in a significantly improved SpO2/FiO2 ratio at 2 (10.11 mm Hg, 95% CI 1.02–19.37; linear regression, p = 0.02) and 10 minutes (16.62 mm Hg, 95% CI 6.94–26.24, linear regression, p = 0.01) post ETS. When compared with no NSI, NSI led to a significantly reduced SpO2/FiO2 ratio at 2 (-12.58, 95% CI - 21.83–3.45; p = <0.01) and 10 minutes (-10.63, 95% CI -20.51–0.87; linear regression, p = 0.03) post ETS. When compared to no RM, RM application increased the mean Cdyn by 0.20 ml/cmH2O/kg at 10 minutes post ETS (95% CI 0.14–0.34; linear regression, p = 0.001). RMs applied post-ETS significantly increased EELI at 2 and 5 minutes postsuction (p = <0.001). No significant difference in tidal volume impedance measures was found in either factorial group following ETS. Conclusion: This PhD research revealed a high rate of AEs associated with current ETS practices and has extended current evidence related to modifiable risk factors for suction related AEs. Evidence practice gaps were identified in suction guidelines, which were found to impact on clinician decision-making in the context of NSI and RM use in the PICU. The pilot trial confirmed it is feasible to conduct a definitive factorial RCT of NSI and RMs with protocol modifications to widen participant eligibility to include cardiac surgical patients and PICU readmissions. It also confirmed the need for further research to identify effective methods to prevent ETS AEs by definitively testing the safety and efficacy of NSI and RMs to reduce retained respiratory secretions and alveolar collapse.