Accepted Article Physical activity and personal factors associated with nurse 1 resilience in intensive care units Fiona Yu1, Alana Cavadino2, Lisa Mackay3, Kim Ward4, Anna King5, Melody Smith6 Ph.D., Candidate, MN (HONS), School of Nursing, Faculty of Medical and Health Science, University of Auckland, RN in Intensive Care Unit, Waikato Hospital, Waikato, New Zealand. E-mail: syu037@aucklanduni.ac.nz 2 Ph. D., Biostatistician, School of Population Health, Faculty of Medical and Health Science, University of Auckland, Auckland, New Zealand. E-mail: a.cavadino@auckland.ac.nz 3 Ph.D., Lecturer, School of Sport and Recreation, Auckland University of Technology, Auckland, New Zealand. E-mail: lisa.mackay@aut.ac.nz 4 Ph.D., RN, Lecturer, School of Nursing, Faculty of Medical and Health Science, University of Auckland, Auckland, New Zealand. E- mail: k.ward@auckland.ac.nz 5 Ph.D., BNurs (HONS), RN, Lecturer, School of Nursing, Faculty of Medical and Health Science, University of Auckland, Auckland, New Zealand. E-mail: a.king@auckland.ac.nz 6 Ph.D., Associate Professor, Co-Associate Head (Research), School of Nursing, Faculty of Medical and Health Science, University of Auckland, Auckland, New Zealand. E-mail: melody.smith@auckland.ac.nz Correspondence should be addressed to Fiona Yu, 85 Park Road, Grafton, Auckland, New Zealand (E-mail: syu037@aucklanduni.ac.nz). Key words: Axivity AX3 accelerometer, dynamic standing, ICU, job demands and resources model, leisure-time physical activity, moderate to vigorous physical activity, MVPA, occupational physical activity, recovery, resilience A short running title: Physical activity and nurse resilience Conflict of Interest Statement: The authors declare no conflict of interest. Funding: Melody Smith was supported by the University of Auckland and Sir Charles Hercus Research Fellowship (grant number 17/013). Fiona Yu was supported by the University of Auckland Doctoral Scholarship. Authors contribution statement: Fiona Yu: Methodology, Formal analysis, Investigation, Resources, Data curation, Writing-Original draft preparation, Reviewing and Editing, Visualisation Alana Cavadino: Reviewing and Editing Lisa Mackay: Resources, Reviewing and Editing Kim Ward: Reviewing and Editing, Supervision Anna King: Reviewing and Editing, Supervision This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/JOCN.15338 This article is protected by copyright. All rights reserved Accepted Article Melody Smith: Conceptualisation, Methodology, Resources, Reviewing and Editing, Visualisation, Supervision, Project administration This article is protected by copyright. All rights reserved Accepted Article 1 2 MISS FIONA YU (Orcid ID : 0000-0001-8747-3545) 3 4 5 Article type : Original Article 6 7 8 Physical activity and personal factors associated with nurse 9 resilience in intensive care units 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Key words: Axivity AX3 accelerometer, dynamic standing, ICU, job demands and resources model, 28 leisure-time physical activity, moderate to vigorous physical activity, MVPA, occupational physical 29 activity, recovery, resilience 30 31 Abstract 32 Aim and objectives: The study aimed to assess intensive care nurses’ resilience, and identify 33 associated personal factors and physical activity behaviours using a job demands-recovery 34 framework. 35 Background: Currently, there is inconsistent evidence as to whether nurse resilience is associated 36 with personal factors, or with physical activity at work or during leisure-time. 37 Design: A cross-sectional study was conducted with nurses from four intensive care units in Auckland, 38 New Zealand. This article is protected by copyright. All rights reserved Methods: An online survey was conducted to collect nurses’ personal information and assess their 40 resilience levels using the Connor-Davidson Resilience Scale 25. Participants were nurses working at 41 least 32 hours fortnightly and providing direct patient care. Physical activity was objectively measured 42 using a pair of accelerometers worn on the back and thigh over four consecutive days (two workdays 43 followed by two non-workdays). Bivariable and multivariable regression were used to identify personal 44 factors and physical activity behaviours associated with resilience. (Followed the STROBE checklist) 45 Results: A total of 93 nurses were included in the study. The participants’ average resilience level was 46 low. Resilience was positively associated with the objectively measured physical job demand factors: 47 occupational physical activity, moderate to vigorous physical activity at work, and dynamic standing at 48 work. Resilience was negatively associated with one objectively measured recovery factor: sleep 49 during leisure-time. In multivariable modelling, being married and moderate to vigorous physical 50 activity at work were positively associated with resilience, while not having religious beliefs and sleep 51 during leisure-time were negatively associated with resilience. 52 Conclusions: Resilient nurses have a greater tolerance to high physical activity at work and lower 53 sleep duration during leisure-time. Strategies are needed to improve intensive care nurses’ resilience 54 levels. 55 Relevance to clinical practice: Results may help managers gain a better understanding of the ICU 56 nurses’ characteristics associated with resilience, leading them to develop strategies for improving 57 ICU nurse resilience. Accepted Article 39 58 59 What does this paper contribute to the wider global clinical 60 community? 61 Impact statement: 62 This study is the first to objectively measure nurses’ physical activity utilising a job demands- 63 recovery framework and to identify the associations between resilience and nurses’ physical 64 activity at work/leisure-time. Dynamic standing, as a novel concept and a new way of 65 measuring physical activity at work, is introduced in this study. 66 67 68 69 Married or religious nurses were found to have higher levels of resilience, compared to single or non-religious nurses. Resilient nurses had higher levels of physical activity at work and lower sleep duration during leisure-time. 70 71 Introduction 72 Health promotion is of great importance for organisations as well as for nurses to support physical 73 and psychological wellbeing, ultimately improving work performance and staff retention. Prolonged 74 nursing shortages result in heavy workloads that lead to increased work stress and turnover, and this 75 profoundly affects nurses physically and mentally (Chan, Tam, Lung, Wong, & Chau, 2013; da Costa 76 & Pinto, 2017). Evidence shows that nurses from intensive care units (ICU) experience a high risk of 77 developing psychological problems, and resilience helps attenuate the effects of adverse outcomes. This article is protected by copyright. All rights reserved Mealer et al. (2012) collated data from 744 American ICU nurses studied in 2010 and found that 80% 79 had one or more symptoms of burnout syndrome (emotional exhaustion, depersonalisation, or lack of 80 personal accomplishment) (Mealer et al., 2012). The study also reported that highly resilient nurses 81 (22% of the 744 participants) had a lower profile of burnout syndrome, posttraumatic stress disorder, 82 anxiety or depression (Mealer et al., 2012). Similarly, another study by Rushton, Batcheller, 83 Schroeder, and Donohue (2015) surveyed 114 nurses from six ICUs at four American hospitals and 84 identified that highly resilient nurses exhibited lower stress levels, and that resilience helped protect 85 them from burnout. Increased personal resilience may help nurses improve adaptation to stressful 86 work environments (Hart, Brannan, & De Chesnay, 2014). It may also reduce nurse vulnerability 87 (Jackson, Firtko, & Edenborough, 2007) and mitigate the effects of emotional dissonance (Delgado, 88 Upton, Ranse, Furness, & Foster, 2017). Therefore, understanding resilience is crucial for intensive 89 care nurses to improve their ability and capacity to deal with stressful and challenging situations. Accepted Article 78 90 91 Scholars have identified that resilience encompasses various characteristics that have been explored 92 in different ways. Resilience is described as an individual’s ability to recuperate and capacity for 93 proactive protection (Youssef & Luthans, 2007). Resilience is also defined as a personality profile of 94 self-control and conscientiousness (Fisk & Dionisi, 2010), and a dynamic process or an innate life 95 force of motivation (Grafton, Gillespie, & Henderson, 2010). A recent study has identified five 96 resilience 97 multidimensional concept, mental immunity, and recovery ability (Aburn, Gott, & Hoare, 2015). Within 98 the context of intensive care nursing speciality, resilience has been explained as nurses’ cognitive 99 flexibility, coping ability and adaptability (Mealer et al., 2012). characteristics in general: overcoming adversity process, adaptation capacity, 100 101 Background 102 Resilience and job demands-recovery 103 Using the Job Demands and Resources Model (Bakker & Demerouti, 2007), researchers identified 104 that resilience mitigated the effects of job demands, thus helping nurses recover quickly from work 105 adversity (Yu, Raphael, Mackay, Smith, & King, 2019). Increased job demands can negatively affect 106 nurses physically and psychologically, while sufficient recovery can enhance their wellbeing and work 107 performance (Bakker & Demerouti, 2007). Numerous studies identified the associations between 108 resilience and the negative psychological impact of job demands, such as stress, burnout, fatigue, 109 anxiety/depression, posttraumatic stress disorder and workplace bullying (Yu et al., 2019). There is 110 very little in the literature to connect how physical activity fits within the Job Demands-Resources 111 Model and nurses’ resilience. In particular, a systematic review has identified that only a few studies 112 have objectively measured physical activity in relation to nurses’ job demands (Chappel, Verswijveren, 113 Aisbett, Considine, & Ridgers, 2017). 114 115 Resilience, activity, and recovery 116 Occupational / leisure-time physical activity and recovery This article is protected by copyright. All rights reserved Regular exercise can help enhance levels of psychological resilience, thus improving well-being 118 (Ozkara, Kalkavan, Alemdag, & Alemdag, 2016; Silverman & Deuster, 2014). Indeed, studies with 119 nurse populations have shown that the promotion of leisure-time physical activity can attenuate the 120 negative outcomes of job demands, such as back pain, premature death, stress, and burnout 121 (McCarthy, Wills, & Crowley, 2018; Reed et al., 2018; Schluter, Turner, Huntington, Bain, & McClure, 122 2011). Further, research with undergraduate student populations has highlighted the link between 123 resilience and leisure-time physical activity (Hegberg & Tone, 2014; Lines et al., 2018). Accepted Article 117 124 125 However, it is unclear whether resilience is associated with nurses’ physical activity at work and/or 126 leisure-time. Nurses’ occupational physical activity primarily involves standing and slow walking 127 behaviours, and they engage in a substantially higher amount of light-intensity activity along with 128 moderate-intensity nursing tasks (Chappel et al., 2017). Furthermore, high job demands, long work 129 hours, and frequently rotating shift work patterns are the main barriers preventing nurses from actively 130 exercising during leisure-time (Caruso, 2014; Reed et al., 2018; Schluter et al., 2011). 131 132 It is also unclear whether sleep, as an important recovery factor, is related to nurse resilience. Sleep 133 quality, as an essential health indicator, has a significant impact on nurses’ work performance and 134 personal life (Pérez-Fuentes, Molero Jurado, Simón Márquez, & Gázquez Linares, 2019). Sleep 135 disturbance is related to increased age, family dependents, unhealthy eating, lack of physical activity, 136 and low emotional intelligence (Pérez-Fuentes et al., 2019). Resilience is a protective resource and 137 can help mitigate the negative effect of perceived stress on sleep disturbance, thus improving sleep 138 quality (Liu et al., 2016). 139 140 Physical activity measurement and concepts 141 One issue that has hindered progress in this area is how physical activity has been measured and 142 conceptualised. For example, there has been a general lack of objective measurement of physical 143 activity in relation to nurse resilience. Objective measures, such as using accelerometry, are preferred 144 over subjective measures (e.g. surveys), as they eliminate issues with self-report bias, recall or social 145 desirability, which may overestimate physical activity (Chappel et al., 2017). Accelerometry is a 146 technique for quantifying human movement patterns using accelerometer-based systems. 147 Accelerometers can be worn on the wrist, ankle, hip or on the lower back. The application of multiple 148 accelerometer units allows for accurate determination of physical activity behaviour (such as sitting, 149 standing, moving, or lying) and for estimating the intensity of movement (such as moderate to 150 vigorous physical activities). The time-stamped feature of accelerometry also enables the extraction of 151 specific time-periods for data analysis. 152 153 In the context of this research, comprehending work-related physical activity “demands” and leisure- 154 time “recovery” is important in order to understand the different pathways to resilience. Moderate to 155 vigorous intensity physical activity (such as brisk walking or jogging) is of particular interest. This is 156 because studies have identified the beneficial outcomes of moderate to vigorous physical activity on This article is protected by copyright. All rights reserved health, such as improving metabolic rate and decreasing mortality risk (Ekelund et al., 2019; Saint- 158 Maurice, Troiano, Berrigan, Kraus, & Matthews, 2018). Evidence has also highlighted the link 159 between resilience and moderate to vigorous physical activity in studies with different populations 160 (Hegberg & Tone, 2014; Lines et al., 2018; Wermelinger Ávila, Corrêa, Lucchetti, & Lucchetti, 2018). 161 However, all these studies linking moderate to vigorous physical activity to health outcomes utilised 162 subjective measurement. Accepted Article 157 163 164 The use of accelerometry to measure free living activity behaviours is progressing rapidly. One 165 advancement of interest is the use of multiple accelerometers to detect postures, using algorithms to 166 predict specific activity behaviours. This is opposed to generic measures of physical activity intensity. 167 For example, one recent development of relevance to nursing physical activity behaviours is the 168 conceptualisation and objective measurement of “dynamic standing”. It is defined as “standing with 169 slight movement” which could include normal daily activities such as cooking, washing, or vacuuming 170 (Narayanan, Stewart, & Mackay, 2019). This type of movement aligns well with ICU nurse physical 171 activities as described above. These could include providing hygiene care (oral care and bed baths), 172 turning patients to relieve pressure areas, removing chest drains or pacing wires, preparing 173 medication for infusion pumps, dialysing patients and so on. To date, the concept of dynamic standing 174 has not been explored in workplace health research on nurses. 175 176 There is a clear knowledge gap in understanding relationships between resilience and nurses’ 177 physical activity at work (job demands) or during leisure-time (recovery). Accordingly, a research 178 question has been raised for this study: Is resilience related to nurses’ physical activity at work (job 179 demands) and/or during leisure-time (recovery)? Hence, this study aimed to examine ICU nurses’ 180 resilience levels in relation to personal factors and physical activity behaviours using a job demands- 181 recovery framework. 182 183 Methods 184 This study, titled “Physical activity and personal factors associated with nurse resilience in intensive 185 care units” was reported following the STROBE checklist (Supplementary File 1). 186 187 Design 188 Setting and ethical approval 189 A study-specific framework of job demands-recovery was developed for contextualising the study aim 190 (Figure 1). The framework categorises nurses’ occupational physical activity, moderate to vigorous 191 physical activity, and dynamic standing as “job demands” during workdays. Conversely, it classifies 192 leisure-time physical activity, moderate to vigorous physical activity, sedentary time, and sleep as 193 “recovery” during non-workdays. Additionally, it considers how resilience is related to the physical job 194 demands in the health impairment process as well as recovery in the motivational process. 195 [Please insert Figure 1 about here] This article is protected by copyright. All rights reserved A cross-sectional multi-centre study was designed and conducted in three tertiary teaching hospitals 197 within the Auckland region. Auckland is New Zealand’s largest city with 1.6 million residents, which is 198 approximately one-third of the nation’s population. The three hospitals are serviced by three different 199 District Health Boards that are responsible for public health services in the greater Auckland area. 200 One of these hospitals has two ICUs, while the others have one each. These three hospitals have a 201 collective workforce of approximately 500 ICU nurses and nurse managers (with varying workloads 202 and schedules). Accepted Article 196 203 204 Four Intensive Care Units from these three hospitals were invited to participate in the study via email 205 between May and June 2019, and the data were then collected from July to October 2019. Ethics 206 approval for the study was granted by the Auckland Health Research Ethics Committee (Ref. 000070). 207 Approval was also obtained from each District Health Board and research site. Participation was 208 voluntary and informed, with each participant required to sign a consent form before answering an 209 online survey and wearing a pair of accelerometers. 210 211 Inclusion/exclusion criteria 212 The study targeted ICU nurses who worked at least 32 hours fortnightly and were involved in direct 213 patient care. It excluded anyone allergic to the Elastoplast medical dressings that were used to attach 214 the accelerometers. It also excluded nurse coordinators, as their workloads differ from ICU nurses 215 who provide direct patient care. The study further required the participants to wear a pair of 216 accelerometers continuously for four consecutive days (two workdays followed by two non-workdays). 217 218 Recruitment and study size 219 Several strategies were utilised for recruitment. Nominated local investigators (independent of the 220 study) assisted the researcher in advertising the research. A $200 prize draw was offered at the 221 completion of data collection at each unit, and each participant received a report of their four-day 222 physical activity. Approximately 500 ICU nurses and nurse managers work in the three selected public 223 hospitals. A targeted sample size of 100 participants was initially identified, based on a previous study 224 which recruited 103 participants from 576 office workers (participation rate 17.8%) in Perth, Western 225 Australia (Tobin, Leavy, & Jancey, 2016). In this earlier study, a total of 103 participants provided 226 sufficient power to identify significant differences in respondents’ physical activity behaviours using an 227 activPAL activity monitor (Tobin, Leavy, & Jancey, 2016). The targeted sample size of 100 228 participants for the current study was further checked using G*Power. At a significance level of 5%, a 229 sample of 100 participants provided 90% power to detect an effect size as small as f² = 0.107 in linear 230 regression analysis, or a “medium” effect size.” (Cohen, 1988; Faul, Erdfelder, Lang, & Buchner, 231 2007). 232 233 Data collection 234 Protocol and measures 235 Survey This article is protected by copyright. All rights reserved An online survey was used to measure ICU nurse resilience (25 questions) and personal factors 237 (sociodemographic information, work factors, and subjectively assessed health behaviours; 20 238 questions). Sociodemographic information included age, sex, marital status, religious beliefs, ethnicity, 239 family dependents, highest qualification attained. It also collected information regarding shift patterns, 240 fortnightly work hours, frequency of working night shifts, years of nursing experience, and years of 241 ICU nursing experience. The health behaviours encompassed frequency of physical activity per week, 242 cigarettes smoked per day, cups of coffee consumed per day, frequency of alcohol consumption, 243 usual sleep duration per 24-hour period, sleep medication use, sleep quality over the last 30 days, 244 and general health status. These questions were mainly adapted from the New Zealand Health 245 Survey (Ministry of Health, 2018). Accepted Article 236 246 247 Resilience was measured using the Connor-Davidson Resilience Scale 25 (Connor & Davidson, 248 2003). This scale assesses resilience from five aspects: personal competence and persistence, 249 negative outcome tolerance, adaptation to change, self-control, and spiritual influences. The answers 250 employ a five-point Likert Scale with 0 representing “not true at all”, 1 “rarely true”, 2 “sometimes true”, 251 3 “often true”, and 4 “true nearly all the time” (Connor & Davidson, 2003). The total scores range from 252 0 to 100, with higher scores indicating greater resilience (Connor & Davidson, 2003). Accordingly, the 253 resilience levels are categorised as “lowest (0-73)”, “low-medium (74-82)”, “high-medium (83-90)” and 254 “high (91-100)” (Connor & Davidson, 2003). The Cronbach’s alpha was 0.93 in the study of Connor 255 and Davidson (2003), indicating that the reliability of this tool makes it pertinent to the current 256 research. 257 258 Objective assessment of physical activity 259 Each participant was issued with a pair of Axivity AX3, triaxial accelerometers, incorporating a 260 temperature sensor and a real-time clock, to wear continuously for four consecutive days (two 261 workdays followed by two non-workdays). Accelerometers were distributed to participants with a 262 unique numerical code used to match the recorded data. One unit was affixed to each participant’s 263 lower back (offset from the spine) and the other onto the anterior aspect of their thigh, using 264 hypoallergenic medical dressings. Figure 2 shows the image of an Axivity AX3 accelerometer, and the 265 locations where a pair of accelerometers were affixed to a participant. OMGUI software (Version 266 1.0.0.30, Open Movement, Newcastle University, UK) was used to configure the accelerometers 267 before wearing, and to download the physical activity data after removal. 268 [Please insert Figure 2 about here] 269 Machine-learning has been utilised to develop algorithms using the data collected from the Axivity 270 AX3 accelerometers. This process enables detection and classification of postures (such as sitting, 271 standing, lying, or moving) and movement intensity (such as sedentary, light, moderate or vigorous 272 levels of physical activity) (Stewart et al., 2018). Movement intensity can also be estimated as being 273 sedentary, or being of a light, moderate or vigorous nature. The protocol for accelerometer wear 274 (dual-placement) and data processing has been validated by previous studies (Duncan et al., 2018; 275 Schneller et al., 2017; Stewart et al., 2018). This article is protected by copyright. All rights reserved Accepted Article 276 277 Variables 278 Physical activity behavioural variables were classified following the job demands-recovery framework 279 and are described below: 280 Job demands variables: 281 282 283 284 285 286 287 290 291 292 293 behaviours over two 12-hour shifts. Moderate to vigorous physical activity comprises moderate intensity and vigorous intensities of movement over two 12-hour shifts. Dynamic standing refers to standing with slight movements over two 12-hour shifts. Recovery variables 288 289 Occupational physical activity consists of standing, dynamic standing, walking, and running Leisure-time physical activity encompasses dynamic standing, walking and running behaviours over two non-workdays. Moderate to vigorous physical activity includes moderate intensity and vigorous intensities of movement over two non-workdays. Sedentary time comprises the sum of sitting and lying (excluding sleeping) behaviours over two non-workdays. Sleep refers to sleeping over two non-workdays. 294 295 Statistical analyses 296 MATLAB (release 2017b, The MathWorks, Inc., MA, USA) was used to convert the Axivity AX3 297 accelerometer raw data into daily 6-part movement behaviours (lying, sitting, standing, dynamic 298 standing, walking and running) over 24-hours. SPSS version 25 was utilised to analyse the 299 accelerometer data and the online survey information. Descriptive statistics were used to identify 300 participants’ demographic characteristics (frequency and percentage), resilience levels (mean and 301 standard deviation), and physical activity (mean and standard deviation). Shapiro-Wilk test, 302 independent samples t-test, and Chi-square were also undertaken to determine participants’ inclusion 303 and exclusion from analyses. Any participant who failed to wear their accelerometers for time-periods 304 exceeding 104.4 minutes in total were excluded from the study, as missing data can affect the 305 accuracy of overall physical activity measurement. 306 307 Bivariable analysis (linear regression) was performed to identify potential associations between 308 resilience and each independent variable. Variables associated with resilience at p < 0.1 were then 309 included in a multivariable linear regression analysis. Stepwise multivariable regression was 310 employed to remove variables with the largest non-significant p-value from the group step by step in 311 order to achieve the best fit model in predicting resilience. Models were repeated with additional 312 adjustment for ICU unit, to account for potential differences in average levels of resilience between 313 the different ICU settings. Variance inflation factor, tolerance, and coefficient correlations were 314 assessed for collinearity diagnostics. A p-value < 0.05 was set as the level of significance for this 315 study. This article is protected by copyright. All rights reserved Accepted Article 316 317 Results 318 Participation rate and compliance rate 319 A total of 374 ICU nurses met the inclusion criteria and were invited to participate in the study. Of 320 these, 132 (35.3%) agreed to participate and signed the consent forms. Of these 132 nurses, 107 321 completed the online survey and wore the accelerometers for four consecutive days. The remaining 322 25 participants were excluded due to unavailable shift schedules (15), injuries (3), or personal 323 reasons (7). Of those 107 participants, one person’s data was lost due to a faulty accelerometer, 324 while 13 failed to continuously wear their accelerometers over the required four consecutive days. 325 The time periods that their accelerometers were unattached varied between 188 minutes and 1,368 326 minutes. The average time-period the accelerometers were not worn was 104.4 minutes (Figure S1 in 327 supporting information). Therefore, the 13 participants were excluded from the study, as the total time 328 their accelerometers were not worn for exceeded 104.4 minutes. 329 330 As a result, 93 of the 106 participants wore the accelerometers continuously over the four consecutive 331 days, giving a compliance rate of 87.7%. A flow diagram (Figure 3) explains the process of the 332 participants’ inclusion and exclusion. There were no statistically significant differences between 333 participants who were included and excluded in terms of resilience, age, ethnicity, self-reported 334 physical activity, and objectively measured moderate to vigorous physical activity at workdays/leisure- 335 time (Table S1 in supporting information). Using the Shapiro-Wilk test, the normality of the distribution 336 of the 93 participants’ resilience scores distribution was confirmed (p = 0.608, Figure S2 in supporting 337 information). 338 [Please insert Figure 3 about here] 339 Sample characteristics 340 Resilience 341 A total of 93 participants were recruited from four intensive care units (ICU1, ICU2, ICU3, ICU4). Of 342 the 93 respondents, 31 (33.3%) and 27 (29.0%) were recruited from ICU1 and ICU2 respectively, 343 while 23 (24.7%) were from ICU3, and the remaining 12 (13.0%) were from ICU4. Table 1 344 summarises the participants’ resilience scores and levels, socio-demographic information, work 345 factors, self-reported health behaviours, and objectively assessed physical activity. The mean 346 resilience score for the total sample was 73.0 ±9.6, indicating the average resilience level was in the 347 lowest category (resilience score 0-73) (Connor & Davidson, 2003). Of the four units, ICU2 had the 348 highest average resilience score at 75.7 ±9.2, while ICU4 had the lowest score of 70.0 ±6.9. Of the 93 349 participants, 55.9% had the lowest levels (0-73), and 29.0% demonstrated low-medium levels (74-82), 350 while only 5.4% exhibited high resilience levels (91-100), and 9.7% scored at high-medium levels (83- 351 90) (Connor & Davidson, 2003). The mean resilience level in each subgroup for each unit and the 352 total sample is shown in Table S2 (in supporting information): Demographic and work-related 353 characteristics by resilience. 354 [Please insert Table 1 about here] This article is protected by copyright. All rights reserved Participants’ mean age was 33.9 ±9.6 years old. ICU1 had the youngest group (30.7 ±6.4), while 356 ICU2 contained the oldest (38.1 ±11.6). Of the 93 nurses, 72.0% were between 20 and 34 years old, 357 three quarters (73.1%) were female, 59.1% were unmarried, and 52.7% were religious. Overall, 46.2% 358 of the 93 participants identified as being of European ethnicity, 37.6% had one or more family 359 dependents, and 62.4% attained a postgraduate qualification. It is noted that ICU1 had the highest 360 percentage of non-Europeans and non-family dependents (71.0% of 31). A majority (86.0%) worked 361 fulltime, 8.6% worked night shifts permanently, 73.1% did night shifts every two weeks, and 18.3% 362 had monthly rotated night shifts or only worked days. Two-thirds (66.7%) had worked more than five 363 years in nursing, while 39.8% had experience in the ICU specialty longer than five years. ICU2 364 employed more experienced nurses than other units, as 51.9% of 27 ICU2 participants had 11 years 365 or more of nursing experience, and 40.7% had worked in ICU over 11 years or more. Accepted Article 355 366 367 Health behaviours 368 Eight-five percent of nurses reported exercising one to five times per week, while only 6.5% never 369 exercised, and 8.5% did not know or preferred not to answer (Table 1). Almost all (98.9%) participants 370 were non-smokers, 22.6% did not drink coffee, and 26.9% did not consume alcohol. Almost two-thirds 371 (72.0%) of respondents had seven or more hours sleep, 86.0% did not use any sleep medication, 372 76.3% had “very good” or “fairly good” sleep quality over the last 30 days, and 63.4% stated that their 373 general health status was “good or fair”. Of the four units, ICU1 had the highest percentage of 374 reporting “fairly good” sleep quality over the last 30 days (71.0% of 31 participants) and “good or fair” 375 general health status (74.2% of 31 participants). 376 377 Physical activity 378 Table 1 shows the participants’ mean physical activity levels per 12-hour shift and per non-workday 379 for each unit and the total sample. The mean level of moderate to vigorous physical activity for the 93 380 participants was 0.8 ±0.6 hours at work (per 12-hour shift) and 0.7 ±0.4 hours in leisure-time (per non- 381 workday). Nurses from ICU1 had the highest mean level of moderate to vigorous physical activity at 382 work (1.0 ±0.7 hours), while ICU3 showed the highest mean level of moderate to vigorous physical 383 activity during leisure time (0.8 ±0.5 hours). The mean level of dynamic standing for the total sample 384 was 2.8 ±0.8 hours per 12-hour shift. 385 386 Table 1 also highlights that the average occupational physical activity level (i.e. non-sedentary time) 387 was 8.9 ±1.2 hours for the total sample per 12-hour shift, and the mean value was similar at each unit. 388 The mean level of leisure-time physical activity (per non-workday) was 3.2 ±1.3 hours from the total 389 sample. There was no vigorous-intensity physical activity measured at any of the units during work 390 hours and during leisure-time. Mean levels for sedentary time and sleep during leisure-time (per non- 391 workday) were 9.9 ±2.2 hours and 8.6 ±1.8 hours respectively, while 3.2 ±1.1 hours was the mean 392 value for sedentary time and 0.3 ±0.5 hours for sleep per 12-hour shift. 393 394 Main results This article is protected by copyright. All rights reserved Bivariable analysis 396 Table 2 summaries the results of linear regressions between resilience and the independent variables. 397 The table shows that resilience was significantly associated with sex (male) (ß = -0.3, 95% confidence 398 interval -9.8 to -1.1, p = 0.015), marital status (being married) (ß = 0.3, 95% confidence interval 2.6 to 399 10.2, p = 0.001), religious beliefs (non-religiousness) (ß = -0.4, 95% confidence interval -11.7 to -4.5, 400 p < 0.001), and ethnicity (European) (ß = 0.2, 95% confidence interval 0.6 to 8.4, p = 0.024). It also 401 shows that resilience had an association with the job demand behavioural variables as follows: 402 moderate to vigorous physical activity (ß = 0.2, 95% confidence interval 0.5% to 6.4%, p = 0.021), 403 dynamic standing (ß = 0.3, 95% confidence interval 0.7% to 4.8%, p = 0.010), and occupational 404 physical activity (ß = 0.2, 95% confidence interval 0 to 2.7%, p = 0.047). In addition, the table 405 illustrates that sleep (ß = -0.2, 95% confidence interval -1.9% to -0.1%, p = 0.026), as only one of the 406 recovery variables, was related to resilience. Accepted Article 395 407 408 Table 2 shows the change in the outcome (resilience score) was associated with a one unit or 409 category change in each predictor. Using “marital status” as an example with “single” as the reference 410 group, those classified as “being married” had, on average, resilience scores that were 6.4 points 411 higher than those who were “single”. Similarly, for religiosity, compared with those classified in the 412 “non-religious group”, those classified as having religious beliefs had, on average, resilience scores 413 that were 8.1 points higher. Their correlation coefficients with resilience for these variables were 0.3 414 (being married) and -0.4 (non-religiousness), indicating that these two predictors have a medium or 415 large significant effect on predicting resilience (±0.1, ±0.3, ±0.5 represent a small, medium, or large 416 effect, respectively (Field, 2014)). Similarly, moderate to vigorous physical activity (ß = 0.2), 417 occupational physical activity (ß= 0.2), dynamic standing (ß = 0.3) and sleep (ß = -0.2) all had a 418 medium significant effect on predicting resilience.” 419 [Please insert Table 2 about here] 420 Multivariable analysis 421 Four assumptions were made before performing the multivariable regression. Firstly, resilience had a 422 linear relationship with each independent variable. Secondly, it was assumed that residuals of the 423 regression were normally distributed. Thirdly, the independent variables were not highly correlated. 424 Lastly, the data should follow a homoscedastic pattern, or the variance is equal. 425 426 Multivariable regression was performed between resilience and the selected variables (if p < 0.1) 427 according to the linear regression results shown in Table 2. These multivariable results were 428 observed for the individual variables when all others were held at their respective means. These 429 selected variables were sex, marital status, religious beliefs, ethnicity, family dependents, frequency 430 of physical activity per week, job demand variables (moderate to vigorous physical activity, dynamic 431 standing, and occupational physical activity), and a recovery variable (sleep). Multicollinearity 432 between these variables was assessed before performing the analysis. Variance inflation factors (VIF) 433 ranged from 1.1 to 2.5, tolerance (T) ranged from 0.4 to 0.9, and correlation coefficients (r) ranged This article is protected by copyright. All rights reserved from 0.0 to 0.7. The results indicated that there was no multicollinearity between these variables (VIF 435 < 10, T > 0.2, and r < 0.8). Accepted Article 434 436 437 A seven-step multivariable linear regression (with and without ICU as a fixed factor) was carried out 438 by removing the variable (with the largest non-significant p value) from the group at each step. The 439 final model is summarised in Table 3. Table 3 shows that the significant predictors for resilience were 440 marital status (being married) (ß = 0.2, 95% confidence interval 1.1 to 8.0, p = 0.011), religious beliefs 441 (non-religiousness) (ß = -0.3, 95% confidence interval -10.0 to -3.1, p < 0.001), moderate to vigorous 442 physical activity over two 12-hour shifts (job demands factor) (ß = 0.2, 95% confidence interval 0.0 to 443 0.1, p = 0.021), and sleep during two non-workdays (recovery factor) (ß = -0.223, 95% confidence 444 interval 0.0 to -0.0, p = 0.013). 445 [Please insert Table 3 about here] 446 The variance inflation factor (VIF), collinearity tolerance (T), and coefficient correlations (R) ranged 447 from 1.0 to 1.1 (VIF <10), 0.9 to 1.0 (T>0.2), and 0.0 to 0.4 (r <0.8) respectively. The range of the 448 absolute standardised ß values was from 0.2 to 0.3, suggesting that religious belief (ß = -0.3) was the 449 most important predictor in the model. The value of R was 32.5%, indicating these four significant 450 predictors accounted for 32.5% of the variation in resilience. There was no association between ICU 451 site and resilience, and adding ICU as a fixed factor to the model did not improve the model fit or have 452 an impact on the other four predictors. Additionally, independent samples t-tests were performed to 453 compare the differences in the mean resilience levels between ICUs, and Table S3 (in supporting 454 information) shows the results. No significant differences in resilience were observed between ICUs; 455 thus no further analysis by units was undertaken. 2 456 457 Table S2 (in supporting information) shows the mean resilience levels for the subgroups of the 458 identified significant demographic factors: marital status and religious beliefs. The table highlights that 459 mean resilience level for single nurses was 70.4 ±8.3, lower than those of the married (76.8 ±10.1). 460 Mean resilience score for nurses who reported having religious beliefs was 76.8 ±9.4, compared with 461 a mean score of 68.7 ±7.9 for nurses who reported not having religious beliefs. 462 463 Discussion 464 This study measured ICU nurses’ resilience levels and the associated personal and physical activity 465 behavioural factors across four main ICU sites in Auckland, in New Zealand. This study is the first to 466 objectively measure nurses’ physical activity utilising a job demands-recovery framework and to 467 introduce a new concept of “dynamic standing” as a nursing activity behaviour for analysis. Data 468 describe a low resilience level in this study cohort. Marital status (being married vs. being single) and 469 religious beliefs (religiousness or non-religiousness) were found to be related to resilience. Within the 470 job demands-recovery framework, resilient nurses had higher levels of physical workloads and lower 471 duration of sleep in leisure time. Marital status, religious beliefs, moderate to vigorous physical activity 472 at work, and sleep during leisure-time were associated with resilience in the final multivariable model. 473 Sex and ethnicity were no longer associated with resilience after adjustment in this final model. This article is protected by copyright. All rights reserved Additionally, a test for multiplicity was not performed due to the exploratory nature of this study 475 (StatsImprove, 2019). Accepted Article 474 476 477 Resilience and personal factors 478 The study found that nurses, who were married or religious, exhibited higher levels of resilience, 479 respectively compared to the single or non-religious nurses. This may be because married or religious 480 nurses have more social connectedness, as resilience is positively related to social support (Yu et al., 481 2019). The finding is in support of a previous study which found that the married nurses reported 482 higher levels of life satisfaction and less emotional turmoil (Ang et al., 2018). In contrast, another 483 study concluded that religious beliefs did not play a key role in improving nurse resilience (Hsieh, 484 Hung, Wang, Ma, & Chang, 2016). 485 486 Our data support marital status (being married) as a protective factor for resilience. A settled 487 partnership may enable a couple to support each other to cope with stress from work and daily life, 488 thus improving resilience (Melvin, Gross, Hayat, Jennings, & Campbell, 2012). Improved resilience in 489 the context of high-quality marital relationships are reflected in studies of highly resilient US soldiers 490 (Melvin et al., 2012) and of Chinese HIV patients (Huang, Zhang, & Yu, 2018). Indeed, a high-quality 491 marital relationship was an essential factor in tolerating stress and increasing personal resilience 492 resources (Huang et al., 2018). 493 494 Our findings align with previous research where religiosity was positively associated with resilience 495 (Choi & Hastings, 2019; Fradelos et al., 2018). Religiousness is often regarded as a source of 496 strength for buffering the effects of adversity (Pargament & Cummings, 2010). Nurses believing in a 497 God are convinced that their God can provide them with shelter or a miracle to deal with stressful 498 work environments (Perera, Pandey, & Srivastava, 2018). This religious coping strategy may be 499 accompanied by greater self-efficacy (Pargament & Cummings, 2010), thus resulting in a higher level 500 of resilience (Yu et al., 2019). The religious coping strategy is of particular importance to ICU nurses, 501 as the highly stressful ICU work environment may lead them to have irrational beliefs, such as low 502 frustration tolerance, or condemnation, that negatively contribute to resilience (Pargament & 503 Cummings, 2010). 504 505 Resilience and job demands-recovery 506 A nursing workload refers to the number of tasks that a nurse performs (Alghamdi, 2016). Accordingly, 507 in this study, occupational physical activity, moderate to vigorous physical activity and dynamic 508 standing during workdays can all be regarded as a nurse’s physical workload (or physical demands). 509 Workload has been shown to directly influence the job outcome (Bogaert, Clarke, Willems, & 510 Mondelaers, 2012). Therefore, understanding the relationships between resilience and physical 511 demands can help improve a nurse’s ability to achieve better work outcomes. 512 This article is protected by copyright. All rights reserved The study found that increased resilience was associated with higher levels of occupational physical 514 activity and with moderate to vigorous physical activity at work. The findings align with previous 515 research where there was a link between resilience and participants’ self-reported physical activity 516 (Hegberg & Tone, 2014; Lines et al., 2018; Wermelinger Ávila et al., 2018). The findings indicate that 517 nurses who are able to accommodate a high physical workload may be resilient and corroborates 518 other studies (Lanz & Bruk-Lee, 2017; Mealer et al., 2012). In previous research, resilience appeared 519 to have a mediating role on the relationship between stress and job outcomes, and highly resilient 520 nurses had a higher ability to control themselves when faced with high workloads (Lanz & Bruk-Lee, 521 2017). Similarly, negative outcomes from the stressful ICU environment had less impact on highly 522 resilient nurses (Mealer et al., 2012). Nevertheless, future research should be conducted to further 523 identify the associations between resilience and physical work activity, and the results would give 524 more credence to the current study findings. Accepted Article 513 525 526 Additionally, the positive relationship between resilience and dynamic standing indicates that resilient 527 nurses spend a greater proportion of their working time in dynamic standing. As dynamic standing is 528 representative of typical nursing tasks, this may be indicative of greater productivity at work. However, 529 further research is necessary to explore dynamic standing in relation to resilience under ICU clinical 530 settings. 531 532 Unexpectedly, the study identified that sleep was the only recovery factor negatively associated with 533 resilience. The finding suggests that the nurses who sleep longer may have lower resilience levels, or 534 resilient nurses may have a lower sleep duration. The finding is in line with the conclusion from a 535 previous longitudinal study that surveyed 903 Chinese students from 16 universities and colleges in 536 Hong Kong and Macau (Wong et al., 2013). The study found that sleep duration and quality had a 537 direct impact on the satisfaction of physical wellbeing and indirect (via mood) effect of psychological 538 health (Wong et al., 2013). However, this finding is not supported by the current study which found 539 that sleep quality was not related to nurses’ resilience according to the online survey results. There 540 may be more underlying factors related to this current finding that sleep during leisure-time was 541 associated with resilience. Therefore, further research is required to understand the relationship 542 between resilience and sleep recovery for ICU nurses. 543 544 Strengths and limitations 545 This study has a number of strengths, which include the objective measurement of physical activity, 546 use of job demands-resources theory to drive analyses, comparatively high participation rate, and use 547 of established survey measures. It also takes a comprehensive approach to understanding resilience 548 in a unique workforce. Moreover, a high compliance rate (87.7%) of accelerometer wear highly 549 strengthens this study. 550 551 However, this study is not without its limitations; its cross-sectional nature means causality cannot be 552 determined. Larger multicentre studies are warranted to corroborate findings presented here. This article is protected by copyright. All rights reserved Additionally, the resilience scale used is from a self-reported questionnaire, so the results may have 554 been affected by the participants’ interests and time-period available in which to answer the questions. 555 However, objectively measured physical activity data minimises bias in reporting physical job 556 demands. Accepted Article 553 557 558 The participants’ recollection, personal interests, and social desirability were the primary potential 559 sources of bias in this study. Some participants may have overreported their health behaviours when 560 answering the online survey. For example, they may have overestimated or underestimated their 561 frequency of physical activity per week, sleep hours, sleep quality during the past month, or health 562 status. Some participants may not have answered the questions carefully during the resilience 563 assessment. Additionally, some participants may have filed inaccurate reports while wearing 564 accelerometers. To reduce the potential sources of bias, the researcher arranged a generous time- 565 period for participants to answer the online survey and reiterated that the success of the study 566 depended on the accuracy of the information from them. 567 568 The study offered a $200 prize draw to each unit as well as a four-day activity graph to each 569 participant, and these may have stimulated the ICU nurses’ interest in participation. It is possible 570 some reactivity may have occurred in response to the research; however, participants were 571 encouraged to undertake their usual physical activities. Nonetheless, changes to usual behaviours 572 may have occurred during the testing period. 573 574 The results may also have been affected by the number of days the participants worked or had off 575 prior to the day when their accelerometers were fitted. Most nurses from the research sites are shift 576 workers with randomly assigned rosters. Some, working full-time, may have three or four 12-hour 577 shifts within a week, but may only be required to do two 12-hour shifts the following week. The study 578 was thus designed to require the participants to wear a pair of accelerometer over four consecutive 579 days (two workdays followed two non-workdays) to maximise recording of their work and leisure-time 580 physical activity. During a testing week, some participants had three or four 12-hour consecutive shifts, 581 while others only worked two continuous 12-hour shifts. Therefore, those who worked one or two 12- 582 hour shift(s) already before their accelerometers were fitted were not as refreshed as someone who 583 had their accelerometers fitted on their first day back at work after a few days off. 584 585 Furthermore, the results may have been affected by the participants’ workload, as some of them 586 might not have had an ICU patient to care for while wearing the accelerometers. Extending the time 587 that the accelerometers are worn to seven days instead of four may have provided more reliable 588 assessments of job demands. However, the protocol for two workdays followed by two non-workdays 589 was chosen to minimise participant variability of job demands and recovery periods. 590 591 Some participants reported itchiness on the skin areas where the accelerometers were fitted, while 592 some found it difficult to sleep while wearing these devices. Therefore, they removed the This article is protected by copyright. All rights reserved accelerometers and reaffixed them when they felt comfortable. However, some of these participants 594 were excluded when their non-wear times exceeded 104.4 minutes, and this may have also affected 595 the participation rate in this study (although noting no significant differences in key variables between 596 those included and excluded were found). Finally, the generalisability of this study is limited due to the 597 ICU-based participants, as the ICU clinical setting is different from other nursing roles. Further 598 research with nurses working in non-ICU areas, or with the ICU nurses from the same research sites 599 at a later date, may improve the external validity of the study. Accepted Article 593 600 601 Conclusions 602 The low level of nurse resilience in this study indicates that it is necessary to develop resilience 603 training programs. Married or religious nurses were found to have higher levels of resilience, 604 compared to singles or non-religious nurses. Increased resilience was found to be related to higher 605 levels of nurses’ occupational physical activity, moderate to vigorous physical activity, and dynamic 606 standing for job demands, but it was associated with less sleep duration in leisure-time for recovery. 607 Marital status (being married), religiosity (having religious beliefs), and moderate to vigorous physical 608 activity at work were protective factors in predicting resilience, while sleep on non-workdays was 609 negatively associated with resilience. The findings also suggest that highly resilient nurses have a 610 greater ability to accommodate a substantial amount of physical activity at work and require less sleep 611 duration during leisure-time for recovery. Future research with a larger sample size is needed to 612 further analyse the associations between nurse resilience and physical activity at work and during 613 leisure-time. 614 615 Relevant to clinical practice 616 The findings may help managers gain a better understanding of the ICU nurses’ characteristics 617 associated with resilience, leading them to develop strategies to improve ICU nurse resilience. For 618 example, setting up a network, such as a social club or website, for ICU nurses may increase their 619 social connectedness preventing them developing mental health problems like depression (Santini, 620 Koyanagi, Tyrovolas, Mason, & Haro, 2015). Mindfulness and self-compassion training accompanied 621 with positive religious coping strategies, such as spiritual connectedness, meditation or praying may 622 also help increase nurse resilience (Perera et al., 2018). 623 624 The findings imply that enhancing resilience is vital for ICU nurses to tolerate their physical job 625 demands. Therefore, it is imperative to develop resilience training programs and workplace support 626 for ICU nurses. Previous studies have shown that mindfulness-based stress reduction, event- 627 triggered counselling sessions, and exercise are effective for improving resilience (Mealer et al., 628 2014). Mindfulness-based stress reduction techniques such as yoga or meditation, can be taught to 629 maintain or improve resilience levels. Combining experienced nurses with new staff members in 630 event-triggered counselling sessions can be utilised to promote resilience. Exercise programs, 631 tailored to meet personal interests, can be delivered by mobile phone, email or website. For example, 632 a self-care poster campaign along with personalised physical activity feedback could be used to This article is protected by copyright. All rights reserved encourage nurses to interact with colleagues, set personal goals, and self-monitor progress (Raney & 634 Zanten, 2019). Additionally, several sessions of progressive muscle relaxation combined with music 635 could be offered to ICU nurses to help reduce their stress and fatigue levels, thus improving resilience 636 (Ozgundondu & Metin, 2019). Overall, all these strategies may help ICU nurses increase their 637 resilience levels, thus accommodating high physical workload and enhancing wellbeing. Developed 638 strategies may assist nurses to maintain a healthy profile and improve patient quality of care. Accepted Article 633 639 640 Funding 641 MS was supported by the University of Auckland and Sir Charles Hercus Research Fellowship (grant 642 number 17/013). FY was supported by the University of Auckland Doctoral Scholarship. 643 This article is protected by copyright. All rights reserved References 645 Aburn, G., Gott, M., & Hoare, K. (2015). What is resilience? An integrative review of the empirical Accepted Article 644 646 647 648 649 literature. Journal of Advanced Nursing, 72(5), 980-1000. doi:10.1111/jan.12888 Alghamdi, M. G. (2016). Nursing workload: A concept analysis. Journal of Nursing Management, 24, 449-457. doi:10.1111/jonm.12354 Ang, S. Y., Uthaman, T., Ayre, T. C., Mordiffi, S. Z., Ang, E., & Lopez, V. (2018). Association between 650 demographics and resilience: A cross-sectional study among nurses in Singapore. International 651 Nursing Review, 65(3), 459-466. doi:10.1111/inr.12441 652 653 Bakker, A. B., & Demerouti, E. (2007). The job demands-resources model: State of the art. Journal of Managerial Psychology, 22(3), 309-328. doi:10.1108/02683940710733115 654 Bogaert, P. V., Clarke, S., Willems, R., & Mondelaers, M. (2012). Nurse practice environment, 655 workload, burnout, job outcomes, and quality of care in psychiatric hospitals: A structural 656 equation model approach. Journal of Advanced Nursing, 69(7), 1515–1524. 657 doi:10.1111/jan.12010 658 659 660 Caruso, C. C. (2014). Negative impacts of shiftwork and long work hours. Rehabilitation Nursing, 39(1), 16-25. doi:10.1002/rnj.107 Chan, Z. C. Y., Tam, W. S., Lung, M. K. Y., Wong, W. Y., & Chau, C. W. (2013). A systematic 661 literature review of nurse shortage and the intention to leave. Journal of Nursing Management, 662 21, 605-613. doi:10.1111/j.1365-2834.2012.01437.x 663 Chappel, S. E., Verswijveren, S. J. J. M., Aisbett, B., Considine, J., & Ridgers, N. D. (2017). Nurses’ 664 occupational physical activity levels: A systematic review. International Journal of Nursing 665 Studies, 73, 52-62. doi:10.1016/j.ijnurstu.2017.05.006 666 Choi, S. A., & Hastings, J. F. (2019). Religion, spirituality, coping, and resilience among African 667 Americans with diabetes. Journal of Religion & Spirituality in Social Work: Social Thought, 38(1) 668 doi:10.1080/15426432.2018.1524735 669 670 Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum Associates. 671 Connor, K. M., & Davidson, J. R. T. (2003). Development of a new resilience scale: The Connor- 672 Davidson resilience scale (CD-RISC). Depression and Anxiety 18:76–82 (2003), 18, 76-82. 673 doi:10.1002/da.10113 This article is protected by copyright. All rights reserved da Costa, B., & Pinto, I. (2017). Stress, burnout and coping in health professionals: A literature review. Accepted Article 674 675 676 Journal of Psychology and Brain Studies, 1(1:4), 1-8. Delgado, C., Upton, D., Ranse, K., Furness, T., & Foster, K. (2017). Nurses' resilience and the 677 emotional labour of nursing work: An integrative review of empirical literature. International 678 Journal of Nursing Studies, 70, 71-88. doi:10.1016/j.ijnurstu.2017.02.008 679 Duncan, S., Stewart, T., Mackay, L., Neville, J., Narayanan, A., Walker, C., . . . Morton, S. (2018). 680 Wear-time compliance with a dual-accelerometer system for capturing 24-h behavioural profiles 681 in children and adults. International Journal of Environmental Research and Public Health, 15(7), 682 1296. doi:10.3390/ijerph15071296 683 Ekelund, U., Tarp, J., Steene-Johannessen, J., Hansen, B. H., Jeeris, B., Fagerland, M. W., . . . Lee, I. 684 (2019). Dose-response associations between accelerometry measured physical activity and 685 sedentary time and all cause mortality: Systematic review and harmonised meta-analysis. Bmj, 686 366(I4570) doi:10.1136/bmj.l4570 687 Faul, F., Erdfelder, E., & Lang, A., & Buchner, A. (2007). G*Power 3: A flexible statistical power 688 analysis program for the social, behavioral, and biomedical sciences. Behavior Research 689 Methods, 39, 175–191. doi:10.3758/BF03193146 690 691 Field, A. (2014). Discovering statistics using IBM SPSS statistics (4th ed.). London: SAGE. 692 693 Fisk, G. M., & Dionisi, A. M. (2010). Chapter 7: Building and sustaining resilience in organizational 694 settings: The critical role of emotion regulation. Research on Emotion in Organizations, 6, 167- 695 188. doi:10.1108/S1746-9791(2010)0000006011 696 Fradelos, E. C., Latsou, D., Mitsi, D., Tsaras, K., Lekka, D., Lavdaniti, M., . . . Papathanasiou, I. V. 697 (2018). Assessment of the relation between religiosity, mental health, and psychological 698 resilience in breast cancer patients. Contemporary Oncology, 22(3), 172–177. 699 doi:10.5114/wo.2018.78947 700 701 702 703 704 Grafton, E., Gillespie, B., & Henderson, S. (2010). Resilience: The power within. Oncology Nursing Forum, 37(6), 698-705. doi:10.1188/10.ONF.698-705 Hart, P. L., Brannan, J. D., & De Chesnay, M. (2014). Resilience in nurses: An integrative review. Journal of Nursing Management, 22(6), 720-734. doi:10.1111/j.1365-2834.2012.01485.x Hegberg, N. J., & Tone, E. B. (2014). Physical activity and stress resilience: Considering those at-risk 705 for developing mental health problems. Mental Health and Physical Activity, 706 doi:10.1016/j.mhpa.2014.10.001 This article is protected by copyright. All rights reserved Hsieh, H. F., Hung, Y. T., Wang, H. H., Ma, S. C., & Chang, S. C. (2016). Factors of resilience in Accepted Article 707 708 emergency department nurses who have experienced workplace violence in Taiwan. Journal of 709 Nursing Scholarship, 48(1), 23-30. doi:10.1111/jnu.12177 710 Huang, J., Zhang, J., & Yu, N. X. (2018). Close relationships, individual resilience resources, and 711 well-being among people living with HIV/AIDS in rural China. AIDs Care, 30(55), S49–S57. 712 doi:10.1080/09540121.2018.1496222 713 Jackson, D., Firtko, A., & Edenborough, M. (2007). Personal resilience as a strategy for surviving and 714 thriving in the face of workplace adversity: A literature review. Journal of Advanced Nursing, 715 60(1), 1-9. doi:10.1111/j.1365-2648.2007.04412.x 716 Lanz, J. J., & Bruk-Lee, V. (2017). Resilience as a moderator of the indirect effects of conflict and 717 workload on job outcomes among nurses. Journal of Advanced Nursing, 73(12), 2973-2986. 718 doi:10.1111/jan.13383 719 Lines, R. L. J., Ducker, K. J., Ntoumanis, N., Thøgersen-Ntoumani, C., Fletcher, D., McGarry, S., & 720 Gucciardi, D. F. (2018). Stress, physical activity, and resilience resources: Tests of direct and 721 moderation effects in young adults. Sport, Exercise, and Performance Psychology, 722 doi:10.1037/spy0000152 723 Liu, X., Liu, C., Tian, X., Zou, G., Li, G., Kong, L., & Li, P. (2016). Associations of perceived stress, 724 resilience and social support with sleep disturbance among community-dwelling adults. Stress 725 and Health, 32(5), 578-586. doi:10.1002/smi.2664 726 McCarthy, V. J. C., Wills, T., & Crowley, S. (2018). Nurses, age, job demands and physical activity at 727 work and at leisure: A cross-sectional study. Applied Nursing Research, 40, 116-121. 728 doi:10.1016/j.apnr.2018.01.010 729 Mealer, M., Conrad, D., Evans, J., Jooste, K., Solyntjes, J., Rothbaum, B., & Moss, M. (2014). 730 Feasibility and acceptability of a resilience training program for intensive care unit nurses. 731 American Journal of Critical Care, 23(6), 97. doi:10.4037/ajcc2014747 732 Mealer, M., Jones, J., Newman, J., McFann, K. K., Rothbaum, B., & Moss, M. (2012). The presence 733 of resilience is associated with a healthier psychological profile in intensive care unit (ICU) 734 nurses: Results of a national survey. International Journal of Nursing Studies, 49(3), 292-299. 735 doi:10.1016/j.ijnurstu.2011.09.015 736 Melvin, K. C., Gross, D., Hayat, M. J., Jennings, B. M., & Campbell, J. C. (2012). Couple functioning 737 and post-traumatic stress symptoms in US army couples: The role of resilience. Research in 738 Nursing & Health, 35, 164-177. doi:10.1002/nur.21459 This article is protected by copyright. All rights reserved Ministry of Health. (2018). New Zealand health survey. Retrieved from https://www.health.govt.nz/nz- Accepted Article 739 740 741 health-statistics/national-collections-and-surveys/surveys/new-zealand-health-survey Narayanan, A., Stewart, T., & Mackay, L. (2019). A dual-accelerometer system for detecting human 742 movement in a free-living environment. Medicine and Science in Sports and Exercise, 743 doi:10.1249/MSS.0000000000002107 744 Ozgundondu, B., & Metin, Z. G. (2019). Effects of progressive muscle relaxation combined with music 745 on stress, fatigue, and coping styles among intensive care nurses. Intensive & Critical Care 746 Nursing, 54, 54-63. doi:10.1016/j.iccn.2019.07.007 747 748 749 Ozkara, A. B., Kalkavan, A., Alemdag, S., & Alemdag, C. (2016). The role of physical activity in psychological resilience. Baltic Journal of Sport & Health Sciences, 3(102), 24-29. Pargament, K. I., & Cummings, J. (2010). Anchored by faith: Religion as a resilience factor. In J. W. 750 Reich, A. J. Zautra & J. S. Hall (Eds.), Handbook of adult resilience (pp. 193-210). New York: 751 The Guilford Press. 752 Perera, C. K., Pandey, R., & Srivastava, A. K. (2018). Role of religion and spirituality in stress 753 management among nurses. Psychological Studies, 63(2), 187-199. doi:10.1007/s12646-018- 754 0454-x 755 Pérez-Fuentes, M. C., Molero Jurado, M. M., Simón Márquez, M. M., & Gázquez Linares, J. J. (2019). 756 Analysis of sociodemographic and psychological variables involved in sleep quality in nurses. 757 International Journal of Environmental Research and Public Health, 16(20), 3846. 758 doi:10.3390/ijerph16203846 759 Raney, M., & Zanten, E. V. (2019). Self-care posters serve as a low-cost option for physical activity 760 promotion of hospital nurses. Health Promotion Practice, 20(3), 354-362. 761 doi:10.1177/1524839918763585 762 Reed, J. L., Prince, S. A., Pipe, A. L., Attallah, S., Adamo, K. B., Tulloch, H. E., . . . Reid, R. D. (2018). 763 Influence of the workplace on physical activity and cardiometabolic health: Results of the multi- 764 centre cross-sectional Champlain nurses’ study. International Journal of Nursing Studies, 81, 49- 765 60. doi:10.1016/j.ijnurstu.2018.02.001 766 Rushton, C. H., Batcheller, J., Schroeder, K., & Donohue, P. (2015). Burnout and resilience among 767 nurses practicing in high-intensity settings. American Journal of Critical Care, 24(5), 412-420. 768 doi:10.4037/ajcc2015291 This article is protected by copyright. All rights reserved Saint-Maurice, P. F., Troiano, R. P., Berrigan, D., Kraus, W. E., & Matthews, C. E. (2018). Volume of 770 light versus moderate-to-vigorous physical activity: Similar benefits for all-cause mortality? 771 Journal of the American Heart Association, 7(7), e008815. doi:10.1161/JAHA.118.008815 772 Santini, Z. I., Koyanagi, A., Tyrovolas, S., Mason, C., & Haro, J. M. (2015). The association between Accepted Article 769 773 social relationships and depression: A systematic review. Journal of Affective Disorders, 175, 53- 774 65. doi:10.1016/j.jad.2014.12.049 775 Schluter, P. J., Turner, C., Huntington, A. D., Bain, C. J., & McClure, R. J. (2011). Work/life balance 776 and health: The nurses and midwives e-cohort study. International Nursing Review, 58(1), 28-36. 777 doi:10.1111/j.1466-7657.2010.00849.x 778 Schneller, M. B., Bentsen, P., Nielsen, G., Brond, J. C., Ried-Larsen, M., Mygind, E., & Schipperijn, J. 779 (2017). Measuring children’s physical activity: Compliance using skin-taped accelerometers. 780 Medicine & Science in Sports & Exercise, 1261-1269. doi:10.1249/MSS.0000000000001222 781 782 Silverman, M. N., & Deuster, P. A. (2014). Biological mechanisms underlying the role of physical fitness in health and resilience. Interface Focus, 4(5) doi:10.1098/rsfs.2014.0040 783 Statslmprove. (2019). Multiple testing: When should we adjust for multiplicity? Retrieved from 784 https://www.statsimprove.com/en/multiple-testing-when-should-we-adjust-for-multiplicity/ 785 Stewart, T., Narayanan, A., Hedayatrad, L., Neville, J., Mackay, L. M., & Duncan, S. (2018). A dual- 786 accelerometer system for classifying physical activity in children and adults. Medicine & Science 787 in Sports & Exercise, doi:10.1249/MSS.0000000000001717 788 Tobin, R., Leavy, J., & Jancey, J. (2016). Uprising: An examination of sit-stand workstations, mental 789 health and work ability in sedentary office workers, in Western Australia.55, 359-371. 790 doi:10.3233/WOR-162410 791 Wermelinger Ávila, M. P., Corrêa, J. C., Lucchetti, A. L. G., & Lucchetti, G. (2018). The role of 792 physical activity in the association between resilience and mental health in older adults. Journal 793 of Aging and Physical Activity, 26, 248-253. doi:10.1123/japa.2016-0332 794 Wong, M. L., Lau, E. Y. Y., Wan, J. H. Y., Cheung, S. F., Hui, C. H., & Mok, D. S. Y. (2013). The 795 interplay between sleep and mood in predicting academic functioning, physical health and 796 psychological health: A longitudinal study. Journal of Psychosomatic Research, 74, 271-277. 797 doi:10.1016/j.jpsychores.2012.08.014 798 Youssef, C. M., & Luthans, F. (2007). Positive organizational behavior in the workplace: 799 The impact of hope, optimism, and resilience. Journal of Management, 33(5), 774-800. 800 doi:10.1177/0149206307305562 This article is protected by copyright. All rights reserved Yu, F., Raphael, D., Mackay, L., Smith, M., & King, A. (2019). Personal and work-related factors Accepted Article 801 802 associated with nurse resilience: A systematic review. International Journal of Nursing Studies, 803 93, 129-140. doi:10.1016/j.ijnurstu.2019.02.014 804 805 806 807 808 809 810 811 812 813 814 815 816 This article is protected by copyright. All rights reserved Accepted Article Tables and figures Figure 1: The job demands-recovery framework (adapted from Bakker & Demerouti, 2007) Thigh Dimensions: 23mm (width) x 32.5mm (length) x 7.6mm (height); Weight: 11g. Further details can be found at https://axivity.com Figure 2: Accelerometer (Axivity AX3) image and locations when affixed to participants This article is protected by copyright. All rights reserved Lower Back Accepted Article Figure 3: A flow diagram for participant recruitment, participation, eligibility, and inclusion This article is protected by copyright. All rights reserved Accepted Article Table 1: Resilience, personal factors, and physical activity for ICU1, ICU2, ICU3, ICU4 and the total sample ICU ICU1 ICU2 ICU3 ICU4 Total sample Variables 1 Resilience score (Mean (SD)) 71.7 (9.5) 75.7 (9.2) 73.1 (11.0) 70.0 (6.9) 73.0 (9.6) Average resilience level 2 Lowest Low- Lowest Lowest Lowest medium Mean age (SD) 30.7 (6.4) 38.1 (11.6) 33.2 (9.0) 34.2 (10.8) 33.9 (9.6) Number of participants N (%) 31 (33.3) 27 (29.0) 23 (24.7) 12 (13.0) 93 (100) Age group 20-34 27 (87.1) 14 (51.9) 17 (73.9) 9 (75.0) 67 (72.0) 35+ 4 (12.9) 13 (48.1) 6 (26.1) 3 (25.0) 26 (28.0) Male 10 (32.3) 7 (25.9) 6 (26.1) 2 (16.7) 25 (26.9) Female 21 (67.7) 20 (74.1) 17 (73.9) 10 (83.3) 68 (73.1) Single 22 (71.0) 14 (51.9) 12 (52.2) 7 (58.3) 55 (59.1) Married 9 (29.0) 13 (48.1) 11 (47.8) 5 (41.7) 38 (40.9) Yes 19 (61.3) 15 (55.6) 9 (39.1) 6 (50.0) 49 (52.7) No 12 (38.7) 12 (44.4) 14 (60.9) 6 (50.0) 44 (47.3) European 9 (29.0) 15 (55.6) 12 (52.2) 7 (58.3) 43 (46.2) Other 22 (71.0) 12 (44.4) 11 (47.8) 5 (41.7) 50 (53.8) Family 0 22 (71.0) 16 (59.3) 13 (56.5) 7 (58.3) 58 (62.4) dependents 1+ 9 (29.0) 11 (40.7) 10 (43.5) 5 (41.7) 35 (37.6) Highest Undergraduate 16 (51.6) 7 (25.9) 7 (30.4) 5 (41.7) 35 (37.6) qualification Postgraduate 15 (48.4) 20 (74.1) 16 (69.6) 7 (58.3) 58 (62.4) Work part-time or Part-time 1 (3.2) 8 (29.6) 3 (13.0) 1 (8.3) 13 (14.0) fulltime Fulltime 30 (96.8) 19 (70.4) 20 (87.0) 11 (91.7) 80 (86.0) Work night shifts Permanently 3 (9.7) 4 (14.8) 1 (4.3) 0 8 (8.6) Every two weeks 25 (80.6) 17 (63.0) 14 (60.9) 12 (100) 68 (73.1) Sex Marital status Religious beliefs Ethnicity attained Every month/ Don’t work night shifts 3 (9.7) 6 (22.2) 8 (34.8) 0 17 (18.3) Years of nursing Under 2 2 (6.5) 2 (7.4) 2 (8.7) 2 (16.7) 8 (8.6) experience 2 to 5 10 (32.3) 5 (18.5) 8 (34.8) 0 23 (24.7) 6 to 10 14 (45.2) 6 (22.2) 6 (26.1) 6 (50.0) 32 (34.4) 11+ 5 (16.1) 14 (51.9) 7 (30.4) 4 (33.3) 30 (32.3) Years of ICU Under 2 6 (19.4) 5 (18.5) 6 (26.1) 4 (33.3) 21 (22.6) nursing 2 to 5 16 (51.6) 7 (25.9) 11 (47.8) 1 (8.3) 35 (37.6) experience 6 to 10 8 (25.8) 4 (14.8) 3 (13.0) 4 (33.3) 19 (20.4) 11+ 1 (3.2) 11 (40.7) 3 (13.0) 3 (25.0) 18 (19.4) Frequency of 1-2 9 (29.0) 11 (40.7) 5 (21.7) 0 25 (26.9) physical activities 3-4 12 (38.7) 8 (29.6) 11 (47.8) 5 (41.7) 36 (38.7) per week 5+ 4 (12.9) 5 (18.5) 3 (13.0) 6 (50.0) 18 (19.4) Never 3 (9.7) 1 (3.7) 2 (8.7) 0 6 (6.5) Don’t know/I prefer to not answer 3 (9.7) 2 (7.4) 2 (8.7) 1 (8.3) 8 (8.5) Cigarettes None 30 (96.8) 27 (100.0) 23 (100) 12 (100) 92 (98.9) smoked per day Less than 5 1 (3.2) 0 0 0 1 (1.1) Cups of coffee None 8 (25.8) 5 (18.5) 3 (13.0) 5 (41.7) 21 (22.6) consumed per 1-2 17 (54.8) 13 (48.1) 17 (73.9) 6 (50.0) 53 (57.0) day 3+ 6 (19.4) 9 (33.3) 3 (13.0) 1 (8.3) 19 (20.4) Frequency of None 11 (35.5) 4 (14.8) 5 (21.7) 5 (41.7) 25 (26.9) alcohol Monthly or less 13 (41.9) 11 (40.7) 11 (47.8) 4 (33.3) 39 (41.9) consumption Up to 4 times a month 5 (16.1) 7 (25.9) 6 (26.1) 2 (16.7) 20 (21.5) 4 or more times a week 2 (6.5) 5 (18.5) 1 (4.3) 1 (8.3) 9 (9.7) This article is protected by copyright. All rights reserved 5-6 12 (38.7) 5 (18.5) 4 (17.4) 1 (8.3) 22 (23.7) duration per 24- 7+ 18 (58.1) 21 (77.8) 17 (73.9) 11 (91.7) 67 (72.0) hour period Don’t know 1 (3.2) 1 (3.7) 2 (8.7) 0 4 (4.3) Sleep medication No 29 (93.5) 25 (92.6) 17 (73.9) 9 (75.0) 80 (86.0) use Yes 2 (6.5) 2 (7.4) 6 (26.1) 3 (25.0) 13 (14.0) Sleep quality Very good 2 (6.5) 3 (11.1) 3 (13.0) 3 (25.0) 11 (11.8) over the last 30 Fairly good 22 (71.0) 18 (66.7) 13 (56.5) 7 (58.3) 60 (64.5) days Fairly bad/Very bad 7 (22.6) 6 (22.2) 7 (30.4) 2 (16.7) 22 (23.7) General health Excellent / Very good 8 (25.8) 14 (51.9) 8 (34.8) 4 (33.3) 34 (36.6) status Good / Fair 23 (74.2) 13 (48.1) 15 (65.2) 8 (66.7) 59 (63.4) Job demands7 Sedentary 7.2 (0.9) 7.3 (1.0) 7.7 (0.9) 7.8 (0.6) 7.4 (0.9) Physical work Light 4.1 (0.6) 4.3 (0.9) 3.9 (0.5) 3.4 (0.5) 4.1 (0.7) activity, intensity, Moderate 1.0 (0.7) 0.8 (0.4) 0.7 (0.6) 0.7 (0.2) 0.8 (0.5) sedentary time, Vigorous 0 0 0 0 0 and sleep MVPA3 1.0 (0.7) 0.8 (0.4) 0.7 (0.6) 0.7 (0.2) 0.8 (0.6) duration per 12- Sitting 2.7 (0.8) 3.3 (1.5) 2.8 (0.8) 3.6 (0.8) 3.0 (1.1) hour shift Standing 4.5 (0.9) 4.4 (1.0) 4.8 (0.8) 4.4 (0.7) 4.5 (0.9) Dynamic standing 3.4 (0.6) 3.6 (0.9) 3.2 (0.8) 2.9 (4.4) 2.8 (0.8) Lying (excluding sleeping) 0.2 (0.4) 0.1 (0.1) 0.2 (0.4) 0.2 (0.2) 0.1 (0.3) Walking 1.1 (0.4) 1.1 (0.3) 1.0 (0.3) 1.1 (0.4) 1.1 (0.4) Running 0 0 0 0 0 Accepted Article Usual sleep Mean (SD) Unit: hours Occupational PA 4 9.0 (1.1) 8.3 (1.6) 9.0 (1.1) 8.4 (0.8) 8.9 (1.2) Sedentary time5 2.9 (1.0) 3.3 (1.4) 3.0 (0.9) 3.9 (0.8) 3.2 (1.1) Sleep 0.5 (0.6) 0.1 (0.2) 0.4 (0.7) 0.2 (0.2) 0.3 (0.5) Recovery Sedentary 19.6 (2.6) 19.3 (2.8) 19.6 (1.8) 19.2 (2.0) 19.5 (2.4) Leisure- time Light 3.0 (1.1) 3.1 (1.2) 3.4 (1.3) 3.4 (0.9) 3.2 (1.2) physical activity, Moderate 0.7 (0.5) 0.7 (0.5) 0.8 (0.5) 0.7 (0.3) 0.7 (0.4) intensity, Vigorous 0 0 0 0 0 sedentary time, MVPA3 0.7 (0.5) 0.7 (0.5) 0.8 (0.5) 0.7 (0.3) 0.7 (0.4) and sleep Sitting 7.3 (2.2) 8.6 (2.1) 8.0 (2.2) 6.9 (1.4) 7.8 (2.2) duration per non- Standing 1.8 (0.8) 2.0 (0.8) 2.3 (1.1) 2.4 (1.0) 2.0 (0.9) Dynamic standing 2.2 (1.1) 2.2 (1.1) 2.6 (1.3) 2.6 (0.9) 2.3 (1.1) Lying (excluding sleeping) 0 0 0 0 0 Walking 0.7 (0.4) 0.8 (0.5) 0.8 (0.4) 0.9 (0.5) 0.8 (0.4) Running 0 0.0 (0.1) 0.0 (0.1) 0.1 (0.2) 0.0 (0.1) Leisure time PA5 2.9 (1.3) 3.0 (1.3) 3.4 (1.5) 3.5 (1.1) 3.2 (1.3) Sedentary time6 9.9 (2.2) 10.0 (2.1) 10.0 (2.6) 9.8 (2.1) 9.9 (2.2) Sleep 8.9 (2.1) 8.5 (1.8) 8.4 (1.6) 8.3 (1.8) 8.6 (1.8) 7 workday Mean (SD) Unit: hours 1 Resilience score: Ranging from 0 to 100, with higher scores reflecting greater resilience. 2 Resilience level: Lowest (0-73), low-medium (74-82), high-medium (83-90), and high (91-100) (Connor & Davidson, 2003). Of the 93 participants, 55.9% had the lowest levels (0-73), and 29.0% demonstrated low-medium levels (74-82), while only 5.4% exhibited high resilience levels (91-100), and 9.7% scored at high-medium levels (83-90). 3 MVPA stands for moderate to vigorous physical activity, consisting of moderate-intensity and vigorous-intensity physical activity. 4 Occupational PA stands for occupational physical activity, consisting of standing, dynamic standing, walking, and running behaviours over two 12-hour shifts. 5 Leisure time PA stands for leisure time physical activity, consisting of dynamic standing, walking and running behaviours over two non- workdays. 6 Sedentary time comprises the sum of sitting and lying (excluding sleeping) behaviours over two 12-hour shifts or two non-workdays. 7 The unit for each variable in the job demands and recovery is “hours”. This article is protected by copyright. All rights reserved Accepted Article Table 2: Associations between resilience and personal /physical activity behavioural factors Factors Unstandardised Coefficients Standardised Coefficients B Std. Error Coefficient ß t Age group 20-34 1.5 35+ Reference Male Reference Female -5.4 2.2 0.1 0.7 2.2 -0.3 -2.5 Married 6.4 Single Reference 1.9 0.3 3.4 No -8.1 1.8 -0.4 -4.5 -2.9 to 5.9 0.015 -9.8 to -1.1 Yes Reference European Reference Other 4.5 1.9 0.2 2.3 0 3.4 2.0 0.2 1.7 1+ Reference Undergraduate Reference Postgraduate -1.3 0.001 2.6 to 10.2 0.000 Ethnicity 0.000 -11.7 to -4.5 0.024 Family dependents 0.024 0.6 to 8.4 0.097 Highest qualification 0.097 -0.6 to 7.4 0.544 2.1 -0.1 -0.6 Work part-time or 0.544 -5.3 to 2.8 0.623 Part-time Reference Fulltime 1.4 2.9 0.1 0.5 Work night shifts 0.623 -4.2 to 7.1 0.388 Permanently Reference Every two weeks 3.1 3.6 0.1 0.9 0.393 -4.0 to 10.2 Every month / Don’t -4.4% 4.1 -0.2% -1.1% 0.991 -8.2 to 8.1 work night shifts Years of nursing 0.166 Less than 2 Reference 2 to 5 -2.0 3.9 -0.1 -0.5 0.612 -9.7 to 5.7 6 to 10 2.2 3.7 0.1 0.6 0.555 -5.2 to 9.7 11+ 3.8 3.8 0.2 1.0 0.320 -3.7 to 11.3 Years of ICU nursing experience 0.498 0.001 Religious beliefs experience Interval 0.015 Marital status fulltime 95% Confidence p value 0.498 Sex attained Significant 0.457 Under 2 Reference 2 to 5 -3.2 2.7 -0.2 -1.2 0.229 -8.5 to 2.1 6 to 10 -0.9 3.0 -3.9% -0.3 0.762 -6.9 to 5.1 11+ 0.8 3.1 3.2% 0.2 0.806 -5.4 to 6.7 physical activities 1-2 12.3 4.2 0.6 2.9 0.005 3.9 to 20.7 per week 3-4 8.1 4.1 0.4 2.0 0.052 -0.1 to 16.3 5+ 8.1 4.4 0.3 1.8 0.070 -0.7 to 16.8 Never Reference Don’t know/I prefer 7.5 5.0 0.2 1.5 0.138 -2.5 to 17.5 Frequency of 0.062 to not answer Cups of coffee consumed per day 0.526 None Reference 1-2 -2.8 2.5 -0.1 -1.1 0.264 -7.7 to 2.1 3+ -1.6 3.0 -0.1 -0.5 0.603 -7.6 to 4.5 Frequency of alcohol consumption 0.754 None Reference Monthly or less 0.9 2.5 4.5% 0.3 0.728 -4.1 to 5.8 Up to 4 times a -1.8 2.9 -0.1 -0.6 0.535 -7.6 to 3.9 month This article is protected by copyright. All rights reserved Accepted Article 4 or more times a -1.5 3.8 -4.6% -0.4 Usual sleep duration per 24-hour period 0.612 Reference 7+ -2.1 2.4 -0.1 -0.9 0.371 -6.8 to 2.6 Don’t know -3.7 5.2 -0.1 -0.7 0.484 -14.1 to 6.3 0.784 No -0.8 Yes Reference 2.9 -2.9% -0.3 Very good Reference Fairly good 4.1 3.1 0.2 1.3 0.193 -2.1 to 10.4 Fairly bad/Very bad 2.5 3.5 0.1 0.7 0.474 -4.5 to 9.6 Sleep quality over the last 30 days 0.784 -6.5 to 4.9 0.395 General health status -9.0 to 6.0 5-6 Sleep medication use 0.692 week 0.280 Excellent /Very good Reference Good / Fair -2.2 2.1 -0.1 -1.1 0.280 -6.3 to 1.8 1 Job demands MVPA 3.5% 1.5% 0.2 2.3 0.021 0.5% to 6.4% during two 12-h Dynamic standing 2.8% 1.1% 0.3 2.6 0.010 0.7% to 4.8% shifts Occupational PA2 1.4% 0.7% 0.2 2.0 0.047 0.0 to 2.7% Recovery during two MVPA 2.9% 1.9% 0.2 1.6 0.117 -0.8% to 6.6% non-workdays Leisure time PA3 0.8% 0.6% 0.1 1.3 0.193 -0.4% to 0.9% (leisure-time) Sedentary time4 0.1% 0.4% 3.8% 0.4 0.717 -0.6% to 0.9% Sleep -1.0% 0.4% -0.2 -2.3 0.026 -1.9% to -0.1% Note: The low values from this table (which become zero if rounded to one digit) have been presented as percentages. 1 MVPA stands for moderate to vigorous physical activity. PA stands for physical activity. 2 Occupational PA consists of standing, dynamic standing, walking, and running behaviours during two 12-h shifts. 3 Leisure-time PA encompasses dynamic standing, walking, and running behaviours during two non-workdays. 4 Sedentary time includes sitting and lying (excluding sleeping)behaviours over two non-workdays. This article is protected by copyright. All rights reserved Accepted Article Table 3: Summary of the final model in resilience for multivariable analysis B Variables 2 Standard Standard- Error ised Beta t p 95% Collinearity Variance Confidence Tolerance Inflation Interval Factor R Marital status 4.6 1.8 0.2 2.6 0.011 1.1 to 8.04 0.9 1.1 =32.5% Religious -6.5 1.7 -0.3 -3.8 0.000 -10.0 to -3.1 0.9 1.1 0.0 0.0 0.2 2.4 0.021 0.0 to 0.1 0.9 1.0 -0.0 0.0 -0.2 -2.5 0.013 -0.0 to -0.0 0.9 1.0 beliefs MVPA* over two 12-hour shifts Sleep during two nonworkdays (leisure-time) *MVPA stands for moderate to vigorous physical activity. This article is protected by copyright. All rights reserved