Partnering an implementation strategy with a multi-modal education to improve delirium practices Authors: Kaye Rolls, Mandana Mayahi-Neysi, Arjay Clamonte, Angela Williams, Nadine Oborn, Elle Waters, Susan Hair, Victoria Traynor, Josephine Sau Fan Chow Published: 2026 ABSTRACT AbstractAimsTo improve assessment of and communication for delirium to 85% of eligible patients and to improve nurses’ knowledge of and self-efficacy for delirium.BackgroundDetection and prevention of delirium have become a high priority safety and quality issue globally. However, inadequate screening and objective assessment by nurses remains ongoing problems, often attributed to an inadequate knowledge base.MethodsA repeated cross-sectional audit study, incorporating the 4Es implementation model and a multimodal education program for nurses, was undertaken across four wards of a rural hospital between March 2023 and March 2024. Bimonthly clinical practice audits using a point prevalence approach were undertaken with a pre-determined benchmark of 85%. The 3-module education program (1. Face to face; 2. Self-directed online; 3. Objective Structured Clinical Examination OSCE) was delivered in a new ward every 2 months. Nursing knowledge of (22 items) and self-efficacy (12 items) for delirium practices were collected before the education program (T0), after program (T1) and 6 weeks following program (T2).Results378 patient notes were audited across 7 audits. Practice outcomes for 6 key practices reached and remained at the benchmark over the study period and for three months after. 187 knowledge surveys were submitted by 90 participants across all three time points, with 30 participants submitting all three. 86% were RNs with a mean age of 41.03 and 15.86-years’ experience. Knowledge outcomes were satisfactory and remain unchanged over time [T0 16.62 [SD 2.96]; T1 17.07 [SD 2.01]; T2 17.67 [SD 1.62]). Participant self-efficacy for detection and management of delirium improved significantly between T0 and T2. Participant feedback indicated that the education program, especially module 3-OSCE, contributed significantly to this self-efficacyConclusionThis study demonstrated that pairing an evidence-based education program with an implementation strategy created synergies that led to sustainable improvements in key delirium assessment and communication practices. Nursing leaders, including nursing managers, nurse educators, and specialised expert nurses, play an integral role in ensuring successful translation of research into practice and quality improvement projects. For such projects to be successful, nurse managers need to be highly visible to legitimise the project and implementation team and to facilitate processes.Clinical trial numberNot applicableSupplementary InformationThe online version contains supplementary material available at 10.1186/s12912-026-04644-6.Keywords: Delirium, Nursing education, Implementation science, Self-efficacy, Quality improvement Introduction Delirium is characterised by acute changes in cognition and behaviour. Up to one third of individuals aged more than 65 may be admitted with delirium and one quarter of patients may develop delirium following surgery [1]. Detection and prevention practices for delirium have become a high priority safety and quality issue globally with World Delirium Awareness Day an annual event since 2017 [2]. Nurses are responsible for key practices however under-screening remains a significant problem [3]. The purpose of this paper is to describe the implementation of a delirium project to improve key delirium identification practices at a regional hospital in NSW, Australia. This study is reported in alignment with the Revised Standards for Quality Improvement Reporting Excellence (SQUIRE 2.0) [4] ‘Delirium is a severe neuropsychiatric syndrome characterized by the acute onset of deficits in attention ‘and altered cognition” [1] pp90 that occurs across the lifespan and has serious short- and long-term morbidity and mortality [1, 5]. Currently delirium is divided into one of three types (hypoactive, hyperactive and mixed) depending on the presenting signs and symptoms. However, these subtypes are now under scrutiny because they do not account for the complex pathophysiology leading to the cognitive changes, which limits development of targeted prevention and treatment strategies [6]. After recovering from delirium, patients report they experienced delusions and hallucinations that were extremely distressing and confronting. [7] Delirium prevention and management have evolved significantly in recent years, emphasizing a multifaceted approach that combines risk assessment, early detection, comprehensive assessment, and targeted interventions. The cornerstone of effective delirium care remains early screening and objective assessment, which are crucial for timely identification and prevention [8, 9]. Nursing assessment plays a pivotal role in delirium management, with an emphasis on frequent, systematic evaluations for delirium risk and presence. These assessments should be integrated into routine care using validated tools, such as the Confusion Assessment Method (CAM) or the 4AT, which have both demonstrated high sensitivity and specificity in diverse clinical settings [10, 11]. Early communication with relatives and caregivers is crucial, not only for understanding the patient’s baseline cognition and behaviour [12] but also for providing support and education to allay anxiety [13]. Escalation to the medical team for comprehensive evaluation to identify the possible cause/s of the delirium is essential as delirium represents a significant deterioration in clinical status. This could include blood tests and radiology as well as a medication review to identify and manage delirogenic medications, such as opiates and benzodiazepines and consider complex drug interactions. These practices are part of international standards or guidelines including Australia [14], United kingdom [15] and United States [16] and are critical elements to establish a safe environment for vulnerable patients. A lack of early screening and objective delirium assessment by nurses is an ongoing problem [3], generally attributed to an inadequate knowledge base contributing to limited self-efficacy for delirium practices [17–19]. Self-efficacy refers to the ability of an individual to have personal agency in each situation; it is a personal judgement that they can act appropriately [20], and has been identified as a predictor of nursing performance when caring for patients in critical care [21, 22]. A recent systematic review by Lee and colleagues [23] found that a constructivist‑informed, multi‑component education program can improve clinicians’ delirium practices, knowledge and confidence by using active‑learning strategies such as case studies that require learners to integrate prior knowledge, deliberate with peers, and apply concepts to real‑world clinical scenarios. The authors, however, caution a lack of data supporting a sustained improvement in knowledge and/practice in current studies. This suggests that pairing an implementation strategy with an education intervention may have longer term impact, especially where teams are attempting to improve the delivery of care across a whole hospital. Implementation science is a complex and evolving field and has generated many different approaches for improving or translating evidence into practice [24, 25]. When choosing an implementation framework or model, and developing their implementation strategy, nursing leaders need to be cognisant of local conditions impacting on the desired clinical practice and the resources required. A 2024 systematic review by Fontaine and colleagues found that multifaceted and tailored interventions that address known barriers and facilitate desired practices are more likely to have a positive effect on nursing practices and patient outcomes, with reminders, education (both group and individual), and opinion leaders identified as important components [26]. Methods The aims of the study were to improve assessment of and communication for delirium to 85% of eligible patients and to improve nurses’ knowledge of and self-efficacy for delirium. Study design A repeated cross-sectional audit study incorporating the 4Es implementation model [27] and a multimodal education program for nurses was undertaken across four wards of a rural hospital between March 2023 and March 2024. The primary outcomes were improvement in delirium assessment, knowledge of delirium and nurse self-efficacy (competence and confidence) in relation to management and assessment. Setting and participants. The study was undertaken within current resources. The study was undertaken at a 96-bed district hospital in NSW, Australia. The case mix of the hospital includes General Medicine, Aged care, General Surgery, Orthopaedic surgery, Paediatrics, Obstetrics & Gynaecology, and dialysis. Undergraduates from the University of Wollongong undertake clinical placement. The wards included in the study were inpatient units on ground floor (18 beds, 32 staff) and level 1(20 beds, 30 staff), the emergency department (ED, 15 beds, 45) and intensive care unit (ICU, 8 beds, 19 staff). Wards are predominantly staffed by registered (RN, Bachelors degree) and endorsed enrolled nurses (EEN, Diploma of nursing) with small numbers of assistant in nursing (AIN, undergraduate nurses or 6-week training program). All nursing staff were eligible for the knowledge intervention unless they were on leave. All inpatients from these wards were included in the bi-monthly clinical audits. Intervention The 4Es implementation model underpinned the project and the primary intervention was a three-module education program based on active learning principles (see Table 1). The 4Es model emphasises a ‘bottom-up top-down’ approach to ensure involvement of key stakeholders and opinion leaders, is highly adaptable and builds the capacity of clinicians so they can sustain practice change [27]. The model comprises four elements: Engagement, Educate, Execute and Evaluate and has been recommended to reduce catheter associated urinary tract infection [28] and has been successfully used to reduce delirium in surgical ICU patients [29]. The Engagement component began in mid-2022 when JC connected KR and VT with Bowral Nursing executives, who then facilitated introductions to clinical leaders (clinical nurse consultant, nursing unit managers (NUM) and clinical nurse educators) in late 2022. The hospital Director of Nursing provided executive sponsorship for the project with the NUMs providing local project support. These nursing leaders were key members of the steering committee along with the Aged care clinical nurse consultant (CNC), Deputy Director of Research and the academic research team. The Aged Care CNC coordinated the project at the hospital, while author 1 was responsible for research planning and data analysis. A draft project plan was customised for local conditions during meetings with working party (CNC, ward educators, research team, and authors 1 and 8) and approved by the steering committee. This customisation included decisions regarding data collection, how to deliver the education program, and how to achieve and maintain the engagement of nursing staff on each ward. This engagement included evidence-based processes including audit and feedback [26], ward meetings and discussion in daily safety huddles [30]. Table 1Multimodal education programModuleDeliveryPurpose1. Interactive delirium flip chart activity45 minutesSmall group discussion session using an A3 flip chartAcademic detailing using an A5 pamphletFor participants to develop and consolidate their delirium knowledge2. Online delirium vignettes & Self-reflective case study activity30 + 30 minutesPart 1 – view videos and complete formative assessmentsPart 2 – guided reflection on a caseFor participants to build their knowledge of assessment through guided activities3. Delirium care OSCEs45 minutesSimulation of a delirium caseFor participants to demonstrate and receive feedback on their ability to identify, manage and escalate the care of a person with deliriumOSCE – objective structured clinical examination Multimodal education program 45 minutes Small group discussion session using an A3 flip chart Academic detailing using an A5 pamphlet 30 + 30 minutes Part 1 – view videos and complete formative assessments Part 2 – guided reflection on a case 45 minutes Simulation of a delirium case OSCE – objective structured clinical examination The second element Education was an education program, based on Constructivist learning theory [31, 32], which was designed to improve the assessment of delirium risk and presence and communication with families and the medical team. The multimodal education intervention was adapted from an education program [33, 34] developed by nurses and doctors who were experts in gerontology, delirium, and education. The program consisted of three modules that are described in Table 1. The small‑group flip‑chart work and academic detailing in Module 1 activates the clinicians’ prior knowledge of delirium and support co‑construction and consolidation of understanding through discussion and shared problem‑solving. The authentic online vignettes and guided self‑reflection in Module 2 enables clinicians to build their delirium schemas through iterative engagement with realistic cases and feedback. Module 3 operationalised constructivism by using an OSCE‑based simulation of delirium care, where clinicians enacted their knowledge in a realistic scenario and reconstructed their clinical reasoning through structured feedback. The Execution component involved delivery of the education program, which took approximately 6 weeks and was delivered in a different ward by the ward educator and supported by the CNC (ward 1- March, ward 2 June, ward 3 August; and ward 4 October). Evaluation (the final element) of project outcomes included assessment of practice and nursing knowledge and self-efficacy. A bi-monthly point prevalence audit [35] was completed by the respective nurse educator and CNC on each ward to assess changes in delirium practices using a 17-item audit tool based on the Australian National Delirium standard [14]. Outcome data from the bi-monthly audit were provided as combined and individual wards reports. This audit and feedback [36] allowed the ward educator to celebrate with their respective wards and re-enforce delirium practices, keeping the project as a live topic. Nursing knowledge and self-efficacy were assessed using an existing validated knowledge tool [37]. These outcomes were analysed after the last ward had implemented the education program. The working party met each fortnight to discuss current progress and achievements, and to discuss and develop strategies to improve intervention delivery and address issues that were impacting on the delivery of the project. This facilitated timely resolution of these issues and escalation to nurse executives where necessary. Data collection Practice outcomes were collected electronically in each ward every other month for 15 months between March 2023 and June 2024 and included three months after the Project was completed. Each clinical leader would review the electronic medical record of all inpatients during a pre-designated week. The audit tool was a revised version of the District’s standard delirium audit (created in 2019) and included 17 items: screening and assessment (4 items), communication with relatives (1 item), escalation to medical officer (1 item), and management of individual after delirium was identified (11 items). The clinical leaders were highly familiar with the Audit tool having used it since 2020, so no inter-rater reliability was undertaken. This audit was embedded within the District and state quality systems - Quality Audit Reporting System (QARS). NSW Health QARS [38] is a web-based audit hosting platform that is used to support continuous improvement in healthcare delivery, compliance with national standards and implementation of best practice. Nursing knowledge and self-efficacy were assessed using a survey previously used to assess delirium knowledge in post-operative settings [37, 39, 40] and intensive care [41, 42]. Survey development [37] followed a rigorous process including: 1. development and review by clinical experts (content validity); 2. review by registered nurses in a focus group (face validity); and 3. pilot testing demonstrating an acceptable Cronbach alpha of 0.72 (internal consistency). The knowledge survey included 63 items: Demographic (8 items), self-efficacy (12 items, confidence and competence in delirium practices [1–4 Likert]), knowledge (22 items, multiple choice) and risk factor awareness (10 items, binary). In addition, post OSCE participants were asked six questions regarding their opinions of the education program (4 open-ended and 2 scale questions). Knowledge and self-efficacy were assessed before module 1 was delivered (T0), after module 3 (T1) and six-eight weeks following module 3 (T2). The knowledge survey was hosted on Qualtrics with QR codes providing participants with access. Ethics The study was undertaken in accordance with the Australian National Statement on Ethical conduct in Human Research 2023 [43] in compliance with the principles of the Helsinki Declaration [44] for the conduct of research involving humans or human data. A site-specific approval was obtained from the Human Research Ethics Committee of Southwestern Sydney Local Health District (2022/STE02275) following on from an Ethics approval from University of Wollongong HREC (2019_ETH037). Nurses were made aware of the study via posters and ward meetings. The education intervention was provided for all nurses. Completion of the knowledge survey was voluntary and a QR code provided for potential participants at the three time points. A participant information sheet was included on the landing page with participants specifically asked to agree or disagree to participation and completion of knowledge surveys. Data analysis Data was exported from Qualtrics XM (Provo UT) into Excel (Windows 365) and cleaned before exporting to SPSS (IBM Corp. Released 2023. IBM SPSS Statistics for Windows, Version 29.0.2.0 Armonk, NY: IBM Corp). Pooled data for the whole hospital were analysed. Quantitative data were analysed using descriptive or inferential statist6ics where applicable. The District benchmark of 85% was maintained for practice outcomes. The clinical team set 70% (16/22) as satisfactory for the knowledge test. Open-ended questions (from the post-OSCE knowledge survey) were analysed using inductive content analysis [45, 46]. After exporting the data from SPSS, Author 1 pasted the data, grouped according to question, into Excel and then sent this to the research team to read independently. Open coding was used by Author 1 to develop preliminary categories and these results were presented to research team. The categories were refined during meetings until consensus was reached regarding the final categories. The Excel workbook provided the audit trail and group meetings address credibility, thus addressing trustworthiness tenets of qualitative research [46] Results Practice outcomes 378 patient notes were audited across 7 audits over 15 months (includes 12 months of project plus 3 month follow up). In this manuscript we are reporting the six practice outcomes related to screening or assessment for delirium and communication and Fig. 1 outlines this data (Ward level data is provided in the additional file). Please note the specific wards are not identified to provide anonymity. ‘Screening for delirium on admission to ED or ward presentation’ achieved 85% by Audit 3 from a baseline of 65% and remained near or above the benchmark for the following four audits. ‘Screening within 24 hours of admission to ward’ was 85% for five of seven) audits and was consistently above benchmark for most wards across the project timeline (see additional file). ‘Asking patient or carer about recent changes in behaviour or cognition’ improved from 57% to 85% in the second audit and remained above benchmark for the hospital and most wards the rest of the study. ‘Objective assessment for delirium for at risk patients’ improved from 40% (audit 1) to 88% (audit 2) and remain above the benchmark for all but one of the following audits. ‘Discussion with patient or carer regarding delirium risk’ was at 21% at Audit 1 but did not achieve benchmark until after all wards had received their education. ‘Notification of a medical officer regarding positive delirium assessment’ was at 82% at the first audit and remained above benchmark for all the following audits (NB there were limited numbers of patients identified with delirium). A steady improvement in ward level practices was observed following delivery of the education program in that ward (see Additional file) with all wards reaching and maintaining the benchmark for most subsequent audits. An exception was ward one where the benchmark was not maintained for two practices including objective delirium assessment and communication with patient or carer. Practice outcomes Knowledge and self-efficacy 187 knowledge surveys were submitted by 90 participants (response rate 71.42%) across all three time points however only 30 participants (response rate 23.81%) submitted all three surveys. A large majority were RNs (86%) with a mean age of 41.03 (SD 10.66) and 15.86 (SD 13.86) years’ experience in nursing (See Table 2). The pre-education survey showed there was a good level of knowledge already existing with 74% scoring 16 or greater (n = 90; mean 16.93 SD 2.00). There was a small but significant improvement in knowledge scores between Pre education and at 6 weeks post OSCE (mean 17.07; SD 1.77) with 86% achieving 16 or more. Thirty participants submitted all three knowledge surveys however there was not a significant improvement in knowledge score [T0 16.62 [SD 2.96]; T1 17.07 [SD 2.01]; T2 17.67 [SD 1.62]). Awareness of risk factors associated with delirium was above 80% for six of ten factors (see Fig. 2). Participant Self-efficacy (competence and confidence) to detect and manage delirium and use delirium assessment instruments improved significantly between knowledge surveys T0 and T2 (See Table 2). Table 2Participant demographics and outcomesAll participantsRNEENAINN=907893Age mean (SD)39.39 (11.34)41.03 (10.66)31.67 (10.15)20 (0.00)Total years nursing mean (SD)14.96 (13.52)15.86 (13.86)11.83 (9.75)1 (0.00)KnowledgePre education survey - T0Mean (SD)Post education - T1Mean (SD)6–8 weeks post education - T2Mean (SD)All participants16.93 (2.00) [n = 90]16.62 (3.69)[n = 53]17.70 (1.77)[n = 44]Participants who submitted all three n = 3016.62 (2.96)17.24 (1.86)17.76 (1.70)T0–T1 t − 1.030, p = 0.312T0–T2 t − 1.933, p = 0.196Self-efficacy – competence (participants who submitted all surveys)#Before EducationMean (SD)6–8 weeks after education mean (SD)significanceTo detect hyperactive delirium2.80 (0.61)3.43 (0.50)<0.001To detect hypoactive delirium2.77 (0.63)3.30 (0.54)<0.001To manage a hyperactive delirium2.73 (0.51)3.33 (0.55)<0.001To manage a hypoactive delirium2.70 (0.54)3.30 (0.47)<0.001To detect delirium using CAM3.23 (0.68)3.57 (0.57)<0.05To detect delirium using 4AT3.23 (0.59)3.63 (0.56)<0.05Self-efficacy – Confidence (participants who submitted all surveys)#To detect hyperactive delirium3.07 (0.45)3.47 (0.51)=0.001To detect hypoactive delirium2.93 (0.64)3.30 (0.64)=0.10To manage a hyperactive delirium2.73 (0.64)3.30 (0.53)<0.001To manage a hypoactive delirium2.73 (0.64)3.30 (0.53)<0.001To detect delirium using CAM3.27 (0.58(3.63 (0.49)=0.011To detect delirium using 4AT3.30 (0.57)3.60 (0.59)=0.05# 4 point LikertSignificance level p ≤ 0.05; RN registered nurse; EEN Endorsed enrolled nurse; AIN assistant in nursing Participant demographics and outcomes Pre education survey - T0 Mean (SD) Post education - T1 Mean (SD) 6–8 weeks post education - T2 Mean (SD) 16.62 (3.69) [n = 53] 17.70 (1.77) [n = 44] T0–T1 t − 1.030, p = 0.312 T0–T2 t − 1.933, p = 0.196 Before Education Mean (SD) Significance level p ≤ 0.05; RN registered nurse; EEN Endorsed enrolled nurse; AIN assistant in nursing Delirium risk factor Awareness Following their OSCE, 25 participants provided their views of the education program and changes in clinical practice moving forward. All participants indicated that the education program made a high to medium contribution to the development of their self-efficacy for delirium care (See Fig. 3). Participants had varying opinions on which modules worked best for them. Module 3 (OSCE) was identified most often (n = 8) as the best component of the program because it was a practical application of what they had learnt: ‘we can apply what we have learnt theoretically’. Module 2 (video and reflection) (n = 4) provided participants with ‘a scenario/demonstrations on how to manage (d) a patient with delirium’. Overall, the program increased their knowledge (n = 6) ‘help(ing) to identify it (delirium) earlier and provides treatment as needed’ and the structure (n = 3) provided opportunities for participants learn about delirium and the people they were care for. These respondents plan to change their practice in several ways. Most participants indicated they would be more systematic in their assessment and management practices (n = 12: ‘I plan to be proactive in identifying patients at increased risk of delirium and identifying a patient experiencing delirium’. Participants (n = 4) also indicated they would: ‘more aware now and will use the online assessments in highlighting risks of delirium’, as well as prioritise (n = 1) delirium. One participant indicated they are going to enrol in a dementia course. Contribution of education package for development of self-efficacy in delirium care Discussion The aims of this repeated cross-sectional audit study were to improve assessment of and communication for delirium to 85% of eligible patients and nurses’ knowledge of and practice for delirium. There was early and sustained improvement in clinical practice with the pre-existing benchmark of 85% achieved and sustained for three months after the project ended. Self-efficacy for and delivery of delirium practices significantly improved over time despite delirium knowledge remaining stable. Respondents identified that the Education program, especially the OSCE element, contributed to their confidence and competence for Delirium practice. This study demonstrates that partnering a multi-modal education program with an evidence-based implementation strategy creates synergies, leading to sustained practice improvement. Multimodal education interventions that include group and individual education, practical components and authentic case studies, such as the one used in this study, have been shown to have a greater impact on clinical practice and knowledge of nurses [23, 26, 47]. In this present study knowledge levels remained high while both self-efficacy and clinical practice improved significantly. This contrasts with other studies where knowledge was unsatisfactory and improved in response to an intervention [23]. Group education increases the perceptions of expected clinical practices among a group of nurses while individual education improves attitudes towards and self -efficacy in relation to a clinical practice [26]. The improvement in self-efficacy may also be due to the inclusion of an OSCE and the increased practice while caring for patients. While there is mixed evidence on the positive effect of OSCEs on self-efficacy for delirium [18, 33] there is evidence supporting their role as a crucial component of improving practice [34, 48]. The OSCEs were completed using in-situ simulation [49], thus participants were learning in an authentic setting which promotes the transfer of learning into clinical practice. An OSCE provides the learner with personalised feedback which is vital for the development of clinical competence [50] and were seen by participants in this study as a critical component. Additionally, because screening and assessment were becoming standard clinical practice, participant self-efficacy may also have improved because they are practising these skills more often. The decision to use the 4Es model and incorporating evidence-based processes, including audit and feedback, ward meetings and discussion in daily safety huddles were crucial in achieving the sustained practice change. Early and sustained engagement with senior nurse executives and managers legitimised the project, ensuring that essential resources were made available and facilitated solutions when cross-departmental problems arose. Managerial and executive level promotion and support are important when seeking to improve practice because clinicians have identified a lack of support contributing to poor delirium recognition due to organisational barriers [51] and this has been critical components of other successful delirium projects [23]. However, nurse leaders in hospitals with larger staff establishments have an increased span of control and workload which can limit their interaction with frontline staff and subsequently these staff satisfaction [52]. Therefore, achieving this visibility may only have been possible because the size of the facility facilitates manager- and executive- level nursing staff visibility. Delirium practices were kept at the forefront using proven engagement processes including audit and feedback [26] and safety huddles [30], adding to the effect of group education. The visible partnership between nursing executive and senior nurses demonstrated that improving delirium practice was a shared priority on across all levels of nursing [53] and legitimized the role of CNCs and ward nurse educators as change agents. The CNC and clinical nurse educators worked together to deliver the education and ward engagement components of the 4Es project, overcoming the problem of healthcare silos which have a negative impact on whole of hospital practice improvement [54, 55]. Internal educators are ideal change agents because they have established relationships with colleagues, as well as intimate knowledge of local policy, organisational culture and current challenges. This context-specific knowledge, experience and high credibility help to effectively bridge the gap between theory and practice [56]. Lastly, a simple multiplier effect [57, 58] may have been created because an individual’s self-efficacy for delirium practice increased at the same time as these practices were being normalised as standard practice, leading to an improvement in the collective ward nursing team’s self-efficacy and social acceptance of these practices as standard practice. Strengths and limitations While this study had a sustained positive effect on clinical practices and improved nurses’ self-efficacy for delirium practices a causal link cannot be claimed because of the non-randomised design and lack of a sample size calculation [59]. Randomisation of wards or participants was not possible as the risk of contamination was high in a facility of this size. The single audit pre-study does not rule out the possibly that practice was already improving (maturation). Furthermore, our claims of sustainability are tempered due to the short follow-up [35, 54, 55] time. Sample bias was minimised because a point-prevalence audit [35] method was used to collect practice data. A key strength of the study was the repeated measures approach using a validated tool was used to collect knowledge and self-efficacy outcomes, however attrition bias may have impacted these data because only one-third completed all three knowledge surveys and a sample size calculation was not done. The core strength of the study was the pragmatic execution that capitalised on current assets including strong local collegial relationships and the existing audit system. The latter meant that the ward educators were already familiar with staff needs and the delirium audit processes questions, facilitating delivery of the education program and enhancing the reliability of the data. The size of the facility was a strength because the research team were able to leverage collegial relationships and there was high visibility of nurse leaders. However, the transferability to different size facilities is unknown. Conclusion This project demonstrated that combining a structured implementation strategy with a multimodal, evidencebased education program can meaningfully accelerate and sustain improvements in delirium care practices by nurses, specifically assessment and communication practices. This study highlights the importance of using a deliberate, systemoriented approach that embeds new behaviours into everyday clinical work. Crucially, the study reinforces the central role of nursing leadership at all levels. Nurse managers, clinical nurse educators, and specialised expert nurses were pivotal in legitimising the initiative, driving engagement, and creating the conditions for consistent implementation. This model offers a practical, scalable approach for healthcare organisations seeking to translate evidence into everyday practice and improve outcomes for patients at risk of delirium. Electronic supplementary material Below is the link to the electronic supplementary material. Supplementary material 1 Author contributions KR and VT developed the research protocol. All authors reviewed the manuscript and provided feedback. MNN, AC, AW, NO and EW ran the study on site, and collected audit data. SH and JC provided high level oversight. KR collected the knowledge survey data, KR undertook data analysis. KR developed the draft manuscript, and all other authors provided edits and approved final manuscript. Funding This study was undertaken with no additional funding from grants or development funds. Data availability Data subject to third party restrictions – Bowral Hospital. Declarations Ethical approval and consent to participate The study was undertaken in accordance with the National Statement on Ethical conduct in Human Research 2023 published by the National Health and Medical Research Council and Australian Research Council. The study was given site-specific approval by South Western Sydney Local Health District Human Research Committee 2022/STE02275 following on from full ethics approval from University of Wollongong HREC (2019_ETH037). Participants provided informed consent. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary material 1 Data Availability Statement Data subject to third party restrictions – Bowral Hospital.