Objectives/Aims: To describe, from the clinicians’ perspective, an ‘ideal’ ICU environment, defined as one which enabled efficient, safe delivery of best care and supporting optimal patient healing. Objectives were to identify features of the physical and sensory environments that hindered or enabled best practice, and opportunities for improvement. Methods: A qualitative descriptive study with data collected in multiple stream-specific focus groups analysed using a framework approach. Results: Thirty allied health, medical and nursing staff participated. We found that views about an ‘ideal’ environment and the composition of the care team were shaped by professional roles, with nurses typically considering themselves ‘responsible’ for patients and care coordination. Staff were concerned about visibility of, and access to patients, and whether the bed-space facilitated the provision of best care. Issues impacting on patient recovery included noise, suboptimal lighting, and an inability to personalise the environment. These mostly corresponded with issues identified by patients in separate interviews. However, there were examples of staff and patients having conflicting views, such as whether visiting hours should be restricted. Conclusion: We interviewed ICU staff to help develop a detailed under- standing of the physical and psychosocial environment, and staff’s perspective of what constitutes an ideal ICU. The information collected was used to help co-design a new and innovative ICU bed-space built around the needs and requirements of both staff and patients. https://doi.org/10.1016/j.aucc.2020.04.147 THE EFFECT OF EXERCISE DOSAGE ON RECOVERY FOLLOWING CRITICAL ILLNESS: A RETROSPECTIVE MEDICAL NOTE AUDIT Michelle Paton 1,2 , Carol Hodgson 2, 3 , Rebecca Lane 4 1 Monash Health, Clayton, Australia 2 Monash University, Melbourne, Australia 3 ANZIC-RC, Mebourne, Australia 4 Victorian University, Footscray Park, Australia Introduction: Despite improved survival, ICU patients may suffer from decreased functional recovery that can have long term effects on quality of life. Early mobilisation is an important and feasible intervention to improve physical function, but there is minimal data on exercise pre- scription for this cohort. Objectives/Aims: To determine the effect of exercise dosage on functional recovery measured using patient reported outcome measures. Methods: A secondary analysis of patient data from two Victorian Tertiary Hospitals part of the “Impact of Disability” study were included. Pre- and post-admission EQ-5D scores were obtained via phone call 6 months following ICU admission. The intensity (ranked using the Intensive Care Mobility Scale (IMS)) and amount (measured as number of sessions ach- ieved per day) of active exercise performed during their ICU admission was obtained through medical records. Results: One hundred and seventy-seven patients that performed active exercise (IMS >1) during their ICU stay were included. Improved EQ-5D mobility and total health scores at 6 months following ICU admission were significantly associated with an increasing IMS level achieved in ICU (p ¼ 0.044, p ¼ <0.0001 respectively) but not with the amount of exercise sessions performed (p ¼ 0.473, p ¼ 0.473). On multivariate analysis adjusting for age, gender, severity of illness (APACHE III), mechanical ventilation duration, length of stay and admission diagnosis, higher IMS levels showed no independent association with EQ5D total health score (p ¼ 0.179), but a trend toward significance (p ¼ 0.095) was seen for improved EQ-5D mobility scores. Conclusion: Mobilising patients to their highest functional level during their ICU stay may provide better functional outcomes than increasing the amount of exercise sessions performed. Further research is required to confirm these results. https://doi.org/10.1016/j.aucc.2020.04.148 CREATING HIGH PERFORMANCE RESUSCITATION TEAMS: AN EXPLORATORY STUDY ON RESUSCITATION EDUCATION IN INTENSIVE CARE Kylie Moon 1, 2 , Peter Morley 1, 2 , Rebecca Szabo 2, 3 , Sonia Baldock 1,2 1 Melbourne Health, Melbourne, Australia 2 University of Melbourne, Melbourne, Australia 3 Royal Women's Hospital, Melbourne, Australia Introduction: How good is your ICU at resuscitation? And how well aligned is your resus education program to your unit’s learning needs and actual clinical demand? While there is considerable data on the care and outcomes of patients in ICUs, the incidence of cardiac arrest (CA) in ICU is poorly understood, and there is currently no benchmark for ICUs to measure themselves against. And there is not enough research on whether resuscitation education programs in ICU are meeting the high level of performance required. What is known is that patients in ICU are critically unwell and if they suffer a CA, they have a very high mortality (between 50-100%). There is also concern from the AHA that while there have been advances in resuscita- tion science, there hasn’t been a corresponding improvement in the sur- vival and outcomes for CA patients, implying there is a problem knowledge delivery and retention. Aims: This exploratory study aimed to improve the alignment of resusci- tation education to the learning needs and clinical demands of ICU. It also aimed to identify what high performance in resus looks like in ICU, how this level of performance can be achieved and whether a hidden curricu- lum exists to develop this high performance. Methods: This mixed methods research included a 3-year retrospective review of CA in a large, tertiary level ICU, where the incidence/outcomes of CA in ICU were not well known previously. It also included a 6-month prospective audit of CA in ICU, and interviews with 20 front-line nurses and doctors involved in these CAs. Results: The incidence of CAs in ICU and a hidden curriculum was iden- tified in this study, used to develop high performance in resus. These re- sults will be used to improve the constructive alignment of resus education to the clinical demands of ICU and learning needs. https://doi.org/10.1016/j.aucc.2020.04.149 VERY LOW INTERNAL CONSISTENCY IN MODIFIED EARLY WARNING SCORE PARAMETERS: AUDIT OF 873,182 PHYSIOLOGICAL OBSERVATION SETS OF 15,103 INPATIENTS Ekavi Georgousopoulou, Jennifer Haskell, Imogen Mitchell Medical School, Australian National University, Canberra, Australia Introduction: Recognising inpatient deterioration is paramount for pa- tient safety and Early Warning Systems and Scores were introduced worldwide to signal patients at risk. However, the performance of these scores and systems remains a challenge as sensitivity and specificity are quite low, with an impact on both patients and clinicians. Objectives/Aims: The aim of the analysis was to explore the internal consistency of the Modified Early Warning Score (MEWS) for inpatients’ deterioration recognition in acute hospital general wards. Methods: De-identified hospital records from Tertiary Canberra Hospital (TCH) were retrospectively accessed. Patients not admitted in general wards were excluded from the analysis. Analysis was performed on non- missing data. Score for systolic blood pressure was excluded due to lack of information for usual reference levels. Cronbach’s alpha was used to assess internal consistency of the MEWS parameters. Statistical software SPSS IBM 25.0 was used. Descriptive statistics are presented as mean±standard deviation or relative frequencies. Results: In total, 873,182 physiological observation sets were recorded for 15,103 adult patients admitted to general wards of Canberra Hospital for the Abstracts / Australian Critical Care 33 (2020) S1eS48 S47