Methods Study design and participants An exploratory cross-sectional study was conducted in 2017. In Germany, there is no national register for nurses. Data for this study were collected from a stratified 10% random sample of a database with outpatient facilities, hospitals and nursing homes in Germany from the Institution for Statutory Accident Insurance and Prevention in Health and Welfare Services in Germany. This institution is part of the German social security system. It is the statutory accident insurer for nonstate institutions in the health and welfare services in Germany and thus responsible for the health concerns of the target group investigated in the present study, namely nurses. Due to data protection rules, this institution was also responsible for the first contact with the health facilities. 126 of 3,278 (3.8%) health facilities agreed to participate in PLOS ONE Workload, burnout and resources among nurses PLOS ONE | https://doi.org/10.1371/journal.pone.0245798 January 22, 2021 3 / 17the survey. They informed the study team about how many nurses worked in their institution, and whether the nurses would prefer to answer a paper-and-pencil questionnaire (with a prefranked envelope) or an online survey (with an access code). 2,982 questionnaires/access codes were sent out to the participating health facilities (656 to outpatient care, 160 to hospitals and 2,166 to nursing homes), where they were distributed to the nurses (S1 Table). Participation was voluntary and anonymous. Informed consent was obtained written at the beginning of the questionnaire. Approval to perform the study was obtained by the ethics committee of the State Chamber of Medicine in Rhineland-Palatinate (Clearance number 837.326.16 (10645)). Questionnaire The questionnaire contained questions regarding i) nurse's sociodemographic information and information on current profession as well as ii) palliative care aspects. Furthermore, iii) parts of the German version of the Copenhagen Psychosocial Questionnaire (COPSOQ), iv) a resilience questionnaire [RS-13] and v) single questions relating to resources were added. i) Sociodemographic information and information on current profession. The nurse's sociodemographic information and information on current profession included the variables 'age', 'gender', 'marital status', 'education', 'professional qualification', 'working area', 'professional experience' and 'extent of employment'. ii) Palliative care aspects. Palliative care aspects included self-developed questions on 'additional qualification in palliative care', the 'number of patients' deaths within the last month (that the nurses cared for personally)' and the 'extent of palliative care'. The latter was evaluated by asking: how much of your working time (as a percentage) do you spend with care of palliative patients? The first two items were already used in the pilot study. The pilot study consisted of a qualitative part, where interviews with experts in general and specialised palliative care were performed [43]. These interviews were used to develop a standardized questionnaire which was used for a cross-sectional pilot survey [6, 44]. iii) Copenhagen Psychosocial Questionnaire (COPSOQ). The questionnaire included parts of the German standard version of the Copenhagen Psychosocial Questionnaire (COPSOQ) [45]. The COPSOQ is a valid and reliable questionnaire for the assessment of psychosocial work environmental factors and health in the workplace [46, 47]. The scales selected were 'quantitative demands' (four items, for example: "Do you have to work very fast?") measuring workload, 'burnout' (six items, for example: "How often do you feel emotionally exhausted?"), 'meaning of work' (three items, for example: "Do you feel that the work you do is important?") and 'workplace commitment' (four items, for example: "Do you enjoy telling others about your place of work?"). iv) Resilience questionnaire RS-13. The RS-13 questionnaire is the short German version of the RS-25 questionnaire developed by Wagnild & Young [48]. The questionnaire postulates a two-dimensional structure of resilience formed by the factors "personal competence" and "acceptance of self and life". The RS-13 questionnaire measures resilience with 13 items on a 7-point scale (1 = I do not agree, 7 = I totally agree with different statements) and has been validated in representative samples [49, 50]. The results of the questionnaire were grouped into persons with low, moderate or high resilience. v) Questions on resources. Single questions on personal, social and organizational resources assessed the nurses' views of these resources in being helpful in dealing with the demands of their work. Further, single questions collected the agreement to different statements such as 'Do you receive recognition for your work from the supervisor?' (see Table 4). These resources were frequently reported in the pilot study by nurses in specialised palliative care [6]. PLOS ONE Workload, burnout and resources among nurses PLOS ONE | https://doi.org/10.1371/journal.pone.0245798 January 22, 2021 4 / 17Data preparation and analysis The data from the paper-and-pencil and online questionnaires were merged, and data cleaning was done (e.g. questionnaires without specification to nursing homes, hospitals or outpatient care were excluded). The scales selected from the COPSOQ were prepared according to the COPSOQ guidelines. In general, COPSOQ items have a 5-point Likert format, which are then transformed into a 0 to 100 scale. The scale score is calculated as the mean of the items for each scale, if at least half of the single items had valid answers. Nurses who answered less than half of the items in a scale were recorded as missing. If at least half of the items were answered, the scale value was calculated as the average of the items answered [46]. High values for the scales 'quantitative demands' and 'burnout' were considered negative, while high values for the scales 'meaning of work' and 'workplace commitment' were considered positive. The proportion of missing values for single scale items was between 0.5% and 2.7%. Cronbach's Alpha was used to assess the internal consistency of the scales. A Cronbach's Alpha > 0.7 was regarded as acceptable [35]. The score of the RS-13 questionnaire ranges from 13 to 91. The answers were grouped according to the specifications in groups with low resilience (score 13- 66), moderate resilience (67-72) and high resilience (73-91) [49]. The categorical resource variables were dichotomised (example: not helpful/little helpful vs. quite helpful/very helpful). The study was conceptualised as an exploratory study. Consequently, no prior hypotheses were formulated, so the p-values merely enable the recognition of any statistically noteworthy findings [51]. Descriptive statistics (absolute and relative frequency, M = mean, SD = standard deviation) were used to depict the data. Bivariate analyses (Pearson correlation, t-tests, analysis of variance) were performed to infer important variables for the regression-based moderation analysis. Variables which did not fulfil all the conditions for linear regression analysis were recoded as categorical variables [35]. The variable 'extent of palliative care' was categorised as ' 20 percent of working time' vs. '> 20 percent of working time' due to the median of the variable (median = 20). The first step with regard to the moderation analysis was to determine the resource variables. Therefore all resource variables that reached a p-value < 0.05 in the bivariate analysis with the scale 'burnout' were further analysed (scale 'meaning of work', scale 'workplace commitment', variables presented in Table 4). The moderator analysis was conducted using the PROCESS program developed by Andrew F. Hayes. First, scales were mean-centered to reduce possible scaling problems and multicollinearity. Secondly, for all significant resource variables the following analysis were done: the 'quantitative demand', one resource (one per model) and the interaction term between the 'quantitative demand' and the resource, as well as the covariates 'age', 'gender', 'working area', 'extent of employment', the 'extent of palliative care' and the 'number of patient deaths within the last month' were added to the moderator analysis, in order to control for confounding influence. If the interaction term between the 'quantitative demand' and the resource accounted for significantly more variance than without interaction term (change in R2 denoted as R2 , p < 0.05), a moderator effect of the resource was present. The interaction of the variables ( 1 SD the mean or variable manifestation such as yes and no) was plotted. All the statistical calculations were performed using the Statistical Package for Social Science (SPSS, version 23.5) and the PROCESS macro for SPSS (version 3.5 by Hayes) for the moderator analysis. Results Of the 2,982 questionnaires/access codes sent out, 497 were eligible for the analysis. The response rate was 16.7% (response rate of outpatient care 14.6%, response rate of hospitals PLOS ONE Workload, burnout and resources among nurses PLOS ONE | https://doi.org/10.1371/journal.pone.0245798 January 22, 2021 5 / 1718.1% and response rate of nursing homes 16.0%). Since only n = 29 nurses from hospitals participated, these were excluded from data analysis. After data cleaning, the final number of participants was n = 437. Descriptive results The basic characteristics of the study population are presented in Table 2. The average age of the nurses was 42.8 years, and 388 (89.6%) were female. In total, 316 nurses answered the question how much working time they spend caring for palliative patients. Sixteen (5.1%) nurses reported spending no time caring for palliative patients, 124 (39.2%) nurses reported between 1% to 10%, 61 (19.30%) nurses reported between 11% to 20% and 115 (36.4%) nurses reported spending more than 20% of their working time for caring for palliative patients. Approximately one-third (n = 121, 27.7%) of the nurses in this study did not answer this question. One hundred seventeen (29.5%) nurses reported 4 or more patient deaths, 218 (54.9%) reported 1 to 3 patient deaths and 62 (15.6%) reported 0 patient deaths within the last month. Table 3 presents the mean values and standard deviations of the scales 'quantitative demands', 'burnout', and the resource scales 'meaning of work' and 'workplace commitment'. All scales achieved a satisfactory level of internal consistency. Bivariate analyses There was a strong positive correlation between the 'quantitative demands' and 'burnout' scales (r = 0.498, p 0.01), and a small negative correlation between 'burnout' and 'meaning of work' (r = -0.222, p 0.01) and 'workplace commitment' (r = -0.240, p 0.01). Regarding the basic and job-related characteristics of the sample shown in Table 2, 'burnout' was significantly related to 'extent of palliative care' ( 20% of working time: n = 199, M = 46.06, SD = 20.28; > 20% of working time: n = 115, M = 53.80, SD = 20.24, t(312) = -3.261, p = 0.001). Furthermore, there was a significant effect regarding the 'number of patient deaths during the last month' (F (2, 393) = 5.197, p = 0.006). The mean of the burnout score was lower for nurses reporting no patient deaths within the last month than for nurses reporting four or more deaths (n = 62, M = 42.47, SD = 21.66 versus n = 116, M = 52.71, SD = 20.03). There was no association between 'quantitative demands' and an 'additional qualification in palliative care' (no qualification: n = 328, M = 55.77, SD = 21.10; additional qualification: n = 103, M = 54.39, SD = 20.44, p = 0.559). The association between 'burnout' and the evaluated (categorical) resource variables is presented in Table 4. Nurses mostly had a lower value on the 'burnout' scale when reporting various resources. Only the resources 'family', 'religiosity/spirituality', 'gratitude of patients', 'recognition through patients/relatives' and an 'additional qualification in palliative care' were not associated with 'burnout'. Moderator analyses In total, 16 moderation analyses were conducted. Table 5 presents the results of the moderation analyses where a significant moderation was found. For 'workplace commitment', there was a positive and significant association between 'quantitative demands' and 'burnout' (b = 0.47, SE = 0.051, p < 0.001). An increase of one value on the scale 'quantitative demands' increased the scale 'burnout' by 0.47. 'Workplace commitment' was negatively related to 'burnout', meaning that a higher degree of 'workplace commitment' was related to a lower level of 'burnout' (b = -0.11, SE = 0.048, p = 0.030). A model with the interaction term of 'quantitative demands' and the resource 'workplace commitment' accounted for significantly more variance in 'burnout' than a model without interaction term (R2 = 0.021, p = 0.004). The impact of 'quantitative demands' on 'burnout' was dependent on 'workplace commitment' (b = -0.01, SE = 0.002 p = 0.004). The variables explained 31.9% of the variance in 'burnout'. Regarding the 'good working team' resource, the variables 'quantitative demands' and 'burnout' were positively and significantly associated (b = 0.76, SE = 0.154, p < 0.001), and the Table 2. Basic and job-related characteristics of the sample (n = 437). Variable Age in years, mean (SD) 42.8 (11.8) Age grouped, no. (%) < 35 118 (27.7) 35-49 154 (36.2) 50 154 (36.2) Gender, no. (%) male 45 (10.4) female 388 (89.6) Marital status, no. (%) single 140 (32.6) married 210 (49.0) divorces/widowed 79 (18.4) Education, no. (%) without a school-leaving qualification/ secondary school leaving certificate/ other qualification 69 (16.0) intermediate school-leaving certificate 239 (55.3) qualification for university entrance 124 (28.7) Professional qualification, no. (%) nursing assistant 79 (18.6) nurse 75 (17.7) geriatric nurse 196 (46.2) others (in training, other education) 74 (17.5) Working area, no. (%) nursing home 344 (78.7) outpatient care 93 (21.3) Professional experience in years, mean (SD) 14 (10.6) Extent of employment, no. (%) part-time job 175 (40.4) full-time job 258 (59.6) Additional qualification in palliative care, no. (%) no 329 (76.2) yes/ currently absolving furhter qualification 103 (23.8) Extent of palliative care (as percentage), no. (%) 20 of working time 201 (63.6) > 20 of working time 115 (36.4) Number of patient deaths (in the last month), no. (%) 0 62 (15.6) 1-3 218 (54.9) 4 117 (29.5) Note. Shown are valid percentages; Missing values: age (n = 11), sex (n = 4), marital status (n = 8), education (n = 5), professional qualification (n = 13), professional experience (n = 16), extent of employment (n = 4), additional qualification in palliative care (n = 5), extent of palliative care (n = 121), number of patient deaths (n = 40) https://doi.org/10.1371/journal.pone.0245798.t002 PLOS ONE Workload, burnout and resources among nurses PLOS ONE | https://doi.org/10.1371/journal.pone.0245798 January 22, 2021 7 / 17variables 'good working team' and 'burnout' were not associated (b = -3.15, SE = 3.52, p = 0.372). A model with the interaction term of 'quantitative demands' and the 'good working team' resource accounted for significantly more variance in 'burnout' than a model without interaction term (R2 = 0.011, p = 0.040). The 'good working team' resource moderated the impact of 'quantitative demands' on 'burnout' (b = -0.34, SE = 0.165, p = 0.004). The variables explained 29.7% of the variance in 'burnout'. The associations between 'quantitative demands' and 'burnout' (b = 0.63, SE = 0.085, p < 0.001), between 'recognition supervisor' and 'burnout' (b = -7.29, SE = 2.27, p = 0.001), and the interaction term of 'quantitative demands' and the resource 'recognition supervisor' (b = -0.34, SE = 0.108, p = 0.002) were significant. Again, a model with the interaction term accounted for significantly more variance in 'burnout' than a model without interaction term (R2 = 0.024, p = 0.002). 'Recognition from supervisor' influenced the impact of 'quantitative demands' on burnout for -0.34 on the 0 to 100 scale. The variables explained 33.7% of the variance in 'burnout'. Figs 1-3 demonstrates simple slopes of the interaction effects of 'workplace commitment' predicting 'burnout' at high, average and low levels (Fig 1) respectively with and without the resource 'good working team' (Fig 2) and 'recognition from supervisor' (Fig 3). Higher 'quantitative demands' were associated with higher levels of 'burnout'. At low 'quantitative demands', the 'burnout' level was quite similar for all nurses. However, when 'quantitative demands' increased, nurses who confirmed that they had the resources stated a lower 'burnout' level than nurses who denied having them. This trend is repeated by the resources 'workplace commitment', 'good working team' and 'recognition from supervisor'. The palliative care aspect 'extent of palliative care' showed that spending more than 20 percent of working time in care for palliative patients increased burnout significantly by a value of approximately 5 on a 0 to 100 scale (Table 5).

The article presented is quantitative, correct?

e. Did the researchers use any statistical tests? If yes, were any of the results statistically significant?