Treffer: Using computerized ambulatory diaries for the assessment of job characteristics and work-related stress in nurses
Northamptonshire Healthcare NHS Trust, United Kingdom
University of Dundee, United Kingdom
CC BY 4.0
Sauf mention contraire ci-dessus, le contenu de cette notice bibliographique peut être utilisé dans le cadre d’une licence CC BY 4.0 Inist-CNRS / Unless otherwise stated above, the content of this bibliographic record may be used under a CC BY 4.0 licence by Inist-CNRS / A menos que se haya señalado antes, el contenido de este registro bibliográfico puede ser utilizado al amparo de una licencia CC BY 4.0 Inist-CNRS
FRANCIS
Weitere Informationen
Much research into work-related stress is based on retrospective self-reports, whereas records made at the time could be more valuable. In this study the primary components of two models of work stress, Karasek's demand-control (DC) model and Siegrist's effort-reward imbalance (ERI), were assessed in trained nurses using ambulatory diaries and traditional questionnaire methods. The diaries were entered on small hand-held computers and the method used has been termed ecological momentary assessment (EMA), in which recordings are made in real time in the working environment. The participants were 36 nurses who completed standardized questionnaires evaluating ERI, strain (from DC model), and, over three shifts, computerized behavioural diaries that measured effort-demand, control, reward, and stress every 90 minutes on average, enabling determination of strain and ERI repeatedly in the work situation. A total of 674 observations were recorded. Using multilevel linear modelling, it was found that the questionnaire and computerized diary derived measures of strain (DC) and ERI were reliably correlated. In addition, the ambulatory measures of both strain (DC) and ERI correlated with ratings of stress taken at the same time. From this study it would appear that ambulatory diaries could be a powerful and flexible way of assessing work related stress and its putative determinants in a real life work setting.
AN0022249359;be801apr.06;2019Mar28.12:54;v2.2.500
Using computerized ambulatory diaries for the assessment of job characteristics and work-related stress in nurses.
Much research into work-related stress is based on retrospective self-reports, whereas records made at the time could be more valuable. In this study the primary components of two models of work stress, Karasek's demand-control (DC) model and Siegrist's effort-reward imbalance (ERI), were assessed in trained nurses using ambulatory diaries and traditional questionnaire methods. The diaries were entered on small hand-held computers and the method used has been termed ecological momentary assessment (EMA), in which recordings are made in real time in the working environment. The participants were 36 nurses who completed standardized questionnaires evaluating ERI, strain (from DC model), and, over three shifts, computerized behavioural diaries that measured effort-demand, control, reward, and stress every 90 minutes on average, enabling determination of strain and ERI repeatedly in the work situation. A total of 674 observations were recorded. Using multilevel linear modelling, it was found that the questionnaire and computerized diary derived measures of strain (DC) and ERI were reliably correlated. In addition, the ambulatory measures of both strain (DC) and ERI correlated with ratings of stress taken at the same time. From this study it would appear that ambulatory diaries could be a powerful and flexible way of assessing work related stress and its putative determinants in a real life work setting.
Keywords: demand control model; diary study; ecological momentary assessment; nurses; work-related stress; Effort reward imbalance
Introduction
Much research on work-related stress has relied extensively on retrospective self-reports of job characteristics and questionnaire measures of distress: even the seminal work of Karasek ([8]) and Siegrist ([17]). It is widely recognized that such methods are open to many sources of error and bias. Retrospective summary accounts are affected by the many processes that affect autobiographical memory, such as the participant's current affective state (Kihlstrom, Eich, Sandbrand, & Tobias, [11]; Shiffman & Stone, [16]), the time that has elapsed since the event, and the format in which responses are given (Jobe, Tourangeau, & Smith, [5]). Stone and Shiffman ([19]) advocate what they have termed ecological momentary assessment (EMA), in which the data of interest is collected repeatedly in real time in the participant's natural environment. There are various methods for conducting EMAs but the method of choice is probably that based on small handheld computers, Personal Digital Assistants (PDAs), which prompt the participant, in some random or quasi-random manner, to complete a diary entry on the computer. The time of such entries can be recorded, hence avoiding the retrospective complications of entries that can occur with paper and pencil methods of EMAs (Litt, Cooney, & Morse, [12]).
The utility of real-time measurement has been acknowledged for many years in physiological monitoring (Fahrenberg & Myrtek, [3]; Pickering, [14]), and computer-based methods of EMA have been used with considerable success in a number of research areas, such as smoking cessation (Stone, Turkkan, Bachrach, Jobe, Kurtzman, & Cain, [20]), the psychosocial correlates of fluctuations in blood pressure (Kamarck et al., [7]), and recently in occupational psychology (van Eerde, Holman, & Totterdell, [24]). Of course such methods have disadvantages including the heavy burden that they place on respondents, who may be required to complete diary assessment frequently throughout the day for many days. Perhaps even more problematically, complex constructs such as demand, control, and reward have to be reduced to a few easily completed ratings or questions that may not capture the construct well (see Bolger, Davis, & Rafaeli, [2], for a discussion of EMA and related methods).
In this study we tested the feasibility of using EMA methods to measure the key concepts from Karasek's widely used demand-control (DC) model of work-related strain (Karasek, [8]), Siegrist's ([17]) model of effort reward imbalance (ERI), and self-reported stress in a real-time framework using signal-contingent recording (Stone et al., [20]). The participants in the study were nurses, an occupational group in which many experience high levels of stress (Jones & Johnston, [6]; Williams, Michie, & Pattani, [26]) and in whom both DC and ERI models have been applied with some success. For example, Nolting, Berger, Schiffhorst, Genz, and Kordt ([13]) found high job strain was associated with twice as many accidents at work, such as accidental cuts and needle punctures. ERI in nurses has been shown to relate to components of burnout such as depersonalization and emotional exhaustion (Bakker, Killmer, Siegrist, & Schaufeli, [1]) and intention to leave nursing (Hasselhorn, Tackenberg, & Peter, [4]).
The aims of the study were to determine:
Method
Participants
A total of 293 nurses (271 female, 22 male) from 14 general medical and surgical wards in a general hospital in the North East of Scotland were invited to take part in the study. Forty-eight volunteered and were sent questionnaire packs, 43 (35 females, 8 males, mean age = 34.3 years,
Table I. Demographic information on the sample of 36 nurses.
Apparatus
The diary phase entailed the use of four PDAs (Palm III
The PDA (henceforth called "diary") itself was user-friendly, with all entries made via tapping the stylus on the screen (see <bold>Materials</bold> and
Materials
The initial phase of the study involved participants completing a battery of questionnaires, with the order of questionnaires counterbalanced across participants. Demand, control, and hence "strain" were assessed using Karasek's ([9]) Job Content Questionnaire (JCQ). Research has found the JCQ to have a good psychometric profile (for example, a major cross cultural study [Karasek, Brisson, Nawakami, Houtman, Bongers, & Amick, [10]] reported average Cronbach's alphas for the various subscales averaging over.7, a stable and predicted factor structure and the expected relationships to various occupations).
Effort-reward imbalance (ERI) was measured using the ERI questionnaire (Siegrist, [17]). A major study in different European countries of almost 20,000 participants (Siegrist et al., [18]) confirmed the ERI questionnaire factor structure and showed Cronbach's alphas to average over.7.
The format for the diary questions was based on the Diary of Ambulatory Behavioural States (DABS; Kamarck et al., [7]). An example of the questions taken from DABS is shown in Figure 1, which enquired about how the individual was feeling, and occurred for every "standard" entry. Further questions were added to the diary that mapped the constructs assessed by the questionnaires, i.e., demand/effort, control, reward, and self-reported stress (see Figures 1 and 2). Demand was assessed by asking how hard and how fast participants had worked over the previous 10 minutes, which would allow for calculation of strain and ERI. These items were extracted from the original questionnaire measures, and were modified in an attempt to briefly assess each construct. As there is considerable overlap between the concepts of "demand" and "effort," they were treated as a unitary construct in the diary phase in order to reduce the measurement burden on the participants.
Graph: Figure 1. Examples of a Palm Pilot screen for the assessment of current emotion. "Stressed" was the dependent measure used in this study
Graph: Figure 2. The Palm Pilot screen used to assess Demand and Effort. Work hard and work fast were averaged.
Study design
This study was a within-subjects design incorporating both cross-sectional and longitudinal elements, involving the questionnaire and diary phases, respectively. The independent variables in the questionnaire phase were strain (DC, calculated by dividing demand by control, as illustrated by Theorell et al., [21]) and ERI (calculated by dividing effort by [reward×0.5454]; Siegrist et al., [18]). To emphasise the underlying concept and to avoid confusion, strain will be referred to as DC. In the diary phase, independent variables were DC (calculated as above) and ERI (effort divided by reward) while the dependent variable was the score from the diary stress scale.
Selection and experimental procedures
At the outset of the study, posters were put up on participating wards to advertise the study and aid recruitment. Approximately a week later, all nursing staff were sent an information pack explaining the basis for the study and a consent form. Two weeks later reminders were sent out to individuals who had not yet replied, to give them a final opportunity to sign up for the study. Once individuals had given their consent they were allocated a reference number, which was then used to code both the questionnaire pack and computerized diary to ensure confidentiality. Waves of questionnaire packs were handed out on the wards on return of the consent forms.
Once individuals had returned their completed questionnaire, a block of three or four days was identified for the diary phase of the study dependent on the nurses' shift pattern. However since few nurses provided data for a fourth shift we restricted analysis to the first three shifts. Diaries were delivered to the ward before the start of the first shift, and participants were then instructed on the use of the diary and given a user guide which reiterated this information. The first diary entry, or a practice entry, was always completed with the participant to ensure that they were entirely confident about the use of the diary. Once the alarm sounded, participants had to remove the stylus from the back of the diary and follow the on-screen instructions. Displays on the screens that followed for the standard diary entries were answered by indicating to what extent the individual agreed or disagreed with a statement by selecting a point on the analogue scale (see Figures 1 and 2). Once they had become accustomed to the workings of the diary, most participants found they could complete a set of entries in 1–2 minutes. Participants then kept the diary for at least three shifts. It was programmed to be active only for work hours, and therefore would not disturb them after work, or on days off if they fell during the block of shifts. At the end of the last shift the diary was collected from the ward and the data was downloaded directly into a Microsoft Access 97 database. During the running of the diary phase, a newsletter was sent to all wards to inform all participants about progress of the study and also to encourage the last few volunteers to take part.
Statistical analysis
DC and ERI were calculated as described in <bold>Study design</bold> for both the questionnaire and diary phases of this study. The ipsative approach to missing values in the questionnaires was adopted, such that values were calculated by averaging across all available responses for each individual. Missing diary entries, which were rare at 3.7% of total diary entries requested, were not estimated. The computer software converted the analogue values to scores between 1 and 100. To increase comparability with the questionnaires, which are assessed on 4- or 5-point scales, the diary scores for demand, control, effort, and reward were rescaled into a 1–5 format (0–20 = 1; 21–40 = 2; 41–60 = 3; 61–80 = 4; 81–100 = 5), thereby aiding calculation of the DC and ERI figures. DC and to a lesser extent ERI were heavily skewed and kurtotic and were log transformed. To ease interpretation of the multilevel analysis, a constant was added to the scores after transformation so that minimal DC or ERI scored 0. Stress was scored on a scale of 1–100.
The descriptive analyses, examination of distributions, and transformation of the data were conducted using SPSS V12. Hypothesis testing was based on hierarchical linear modelling (Raudenbush & Byrk, [15]) using the program HLM5 (V5.04). The multivariate sub-program HMLM was used to test 2-level models in which the EMA measures (Level 1) were nested within participants (Level 2). The Level 1 (within participant) variables were diary captured DC (i.e., demand-control), ERI (effort-appreciation), and stress. The Level 2 (between participants) variables were DC and ERI from the standard questionnaires. Two sets of analyses were conducted. First, the relationship between the standard and EMA derived measures of DC and ERI were determined by relating the equivalent Level 1 and Level 2 measures. Second, the within-participant relationships between stress and either ERI or DC were established using only the Level 1 variables. In preliminary analyses, two models of the repeated measures (within participant) data were tested. These were that the variances at each measurement occasion were homogeneous or that there was a first-order autoregressive relationship between successive within-participant data points. Measures taken repeatedly over time with participants are likely to be autocorrelated, and in all cases the best fitting model was obtained with the autoregressive assumption and this was used for all subsequent analyses. All variables were analysed without centring. Intercept and slope were treated as random effects since it is probable that individuals differ in the within subject relationship between DC, ERI, and stress.
Results
The sample characteristics are shown in Table I. DC (the ratio of demand to control) was 0.55 (
Associations between DC and ERI assessed by EMA and questionnaire methods
The primary association of interest in this case is the slope of the EMA measures on the questionnaire measures. The analysis of EMA-derived DC on questionnaire DC showed that the measures were reliably related; participants with higher DC on the questionnaire also had higher EMA-derived DC scores when taking into account all the observations. The multilevel analysis is shown in Table II. The relationship between the two measures of ERI was similar and highly reliable.
Table II. The regression of EMA-derived measures of demand control and effort reward imbalance on questionnaire measures of DC and ERI.
Within-participant (EMA) relationship between stress and strain and ERI
Self-reported stress covaried reliably with both DC and ERI (see Table III). When participants reported that demand exceeded control, or greater ERI, they also reported experiencing greater stress. The DC and ERI scales were adjusted so that what might be considered the most favourable DC or ERI ratios (minimal demand or effort and maximal control or reward) score 0. This means that, for example, when ERI is 0, predicted stress is 6.9 on the 100-point stress scale. The log ERI and DC scales run from 0 to 3.22, so the regression coefficient of 15.1 for the stress on ERI slope indicates that the predicted stress level of a participant at the least advantageous point on effort reward imbalance is 55.5.
Table III. The regression of EMA-derived self-reported stress on EMA measures of DC and ERI.
Discussion
This study aimed to examine the use of computerized behavioural diaries in the well-established field of work-related stress. Significant relationships between the questionnaire-derived measures of DC and ERI and their EMA analogues are a necessary step in validating the EMA measures. The magnitude of the relationship between the two ERI measures was sufficient to provide some evidence of the concurrent validity of the EMA measures but also suggests that it is not simply measuring the same perceptions as are assessed by questionnaire. This is, of course, what one hopes for when using EMA methods. The two DC measures also related reliably although not as strongly as did the two measures of ERI.
Both EMA-derived ERI and DC were associated with self-reported stress across three nursing shifts. This indicates that EMA measures of these hypothesized determinants of job stress related as predicted to stress, and provides further evidence for the validity of the methods. This further suggests that EMA methods can be used to examine short-term perceptions of the job situation and that the obvious simplification of the key concepts did not undermine the measures, at least as applied to the study population. It should be acknowledged that the measures of demand, control, and reward were determined at the same time and with the same instrument, and the relationships obtained may in part reflect common method variance. While the use of ratio scores rather than the raw rating scales for measure DC and ERI may reduce this possibility somewhat, resolving this issue is outside the scope of this study.
The EMA methods we have described could be of value in studying specific aspects of a work situation that was associated with stress or dissatisfaction or in examining the effects of specific changes in the work situation. One of the arguments in favour of EMA is that it provides better measures of the underlying constructs by reducing sources of bias and distortion relating to mood, memory, etc. We cannot comment on that since we have no external criteria against which we can validate EMA and questionnaire measures. Further work is needed to do this and to relate EMA-derived measures to important outcomes, such as health, which do relate to questionnaire measures of DC and ERI (van der Doef & Maes, [22][23]; van Vegchel, de Jonge, Bosma, & Schaufeli, [25]). While it may be that EMA measures are more sensitive, they are also more specific and may not capture the enduring chronic conditions or states that may relate to long-term health outcomes.
A significant weakness in this study was the very poor response rate (less than 15%), although diary completion rate was very high in those who actually took part. This has at least two implications for this study: it may be that computerized EMA was unacceptable with this population and this may have biased the sample. However, it was our impression that the EMA methods used were acceptable to those who actually volunteered, with a very low rate of missed diary entries. No participant complained about the equipment or the burden that it imposed, and all completed that phase of the study. It may be that the low response rate reflected in part the way participants had to be recruited for this study. Local regulations required that the recruitment letter described the study in rather daunting detail, and this may have deterred some people, particularly those who were less confident or inexperienced computer users. In this respect, it may be relevant that men were over represented among the volunteers. As well as gender, there may have been a selection bias, such that the most stressed/distressed individuals were not represented in the sample as they felt they could not undertake such an obligation, and this may have affected the generalizability of the results.
In this study we evaluated the acceptability, utility, and validity of computerized measures of the key components of Karasek's demand-control model and Siegrist's effort reward imbalance model in nurses using both computerized ambulatory diaries and traditional questionnaire methods. We found that the two methods were reliably correlated, and that the ambulatory measures of both strain (DC) and ERI correlated with the nurses' ratings of stress taken at the same time. The diaries were accepted by the participating nurses (who made a large proportion of the required entries), and provided coherent information on ERI, DC, and the relationships between DC, ERI, and stress. We conclude from this study that the use of computerized ambulatory diaries could be a powerful and flexible way of assessing work-related stress and its determinants in real time, and in the work setting of the participants.
Acknowledgments
We would like to thank Ken Munro for his assistance and programming of the PDAs, K. Montgomery, V. Trott, and C. E. Young for assistance with data collection and data entry, and all of the nurses who took part in the study.
References
1 Bakker, A. B., Killmer, C. H., Siegrist, J. and Schaufeli, W. B.2000. Effort-reward imbalance model and burnout among nurses. Journal of Advanced Nursing, 31: 884–891.
2 Bolger, N., Davis, A. and Rafaeli, E.2003. Diary methods: Capturing life as it is lived. Annual Review of Psychology, 54: 579–616.
3 Fahrenberg, J. and Myrtek, M.2001. Progress in ambulatory assessment: Computer assisted psychological and psychopathological methods in monitoring and field studies, Gottingen: Hogrefe & Huber.
4 Hasselhorn, H., Tackenberg, P. and Peter, R.2004. Effort-reward imbalance among nurses in stable countries and countries in transition. International Journal of Occupational and Environmental Health, 10: 401–408.
5 Jobe, J. B., Tourangeau, R. and Smith, A. F.1993. Contributions of survey research to the understanding of memory. Applied Cognitive Psychology, 7: 567–584.
6 Jones, M. C. and Johnston, D. W.1997. Distress, stress and coping in first year student nurses. Journal of Advanced Nursing, 26: 475–482.
7 Kamarck, T. W., Shiffman, S. M., Smithline, L., Goodie, J. L., Thompson, H. S.Ituarte, P. H. G.1998. "The diary of ambulatory behavioural states: A new approach to the assessment of psychosocial influences on ambulatory cardiovascular activity". In Technology and methodology in behavioural medicine, Edited by: Krantz, D. and Baum, A.163–193. Hillsdale, NJ: Laurence Erlbaum.
8 Karasek, R. A.1979. Job demands, job decision latitude, and mental strain: Implications for job redesign. Administration Science Quarterly, 24: 285–307.
9 Karasek, R. A.1985. Job content questionnaire and user's guide (revision 1.1), Lowell: University of Massachusetts Lowell, Department of Work Environment.
Karasek, R., Brisson, C., Nawakami, N., Houtman, I., Bongers, P. and Amick, B.1998. The Job Content Questionnaire (JCQ): An instrument for internationally comparative assessments of psychosocial job characteristics. Journal of Occupational Health Psychology, 3: 322–355.
Kihlstrom, J. F., Eich, E., Sandbrand, D. and Tobias, B. A.2000. "Emotion and memory: Implications for self-report". In The science of self-report: Implications for research and practice, Edited by: Stone, A. A., Turkkan, J. S., Bachrach, C. A., Jobe, J. B., Kurtzman, H. S. and Cain, V. S.81–99. London: Lawrence Erlbaum Associates.
Litt, M. D., Cooney, N. L. and Morse, P.1998. Ecological momentary assessment (EMA) with treated alcoholics: Methodological problems and potential solutions. Health Psychology, 17: 48–52.
Nolting, H. D., Berger, J., Schiffhorst, G., Genz, H. O. and Kordt, M.2002. Job strain as a risk factor for occupational accidents among hospital nursing staff. Gesundheitswesen, 64: 25–32.
Pickering, T. C.(1991). Ambulatory monitoring and blood pressure variability. London: Science Press.
Raudenbush, S. W. and Bryk, A. S.2002. Hierarchical linear models: Applications and data analysis methods, London: Sage.
Shiffman, S. and Stone, A. A.1998. Assessment of health-relevant variables in natural environments. Health Psychology, 17: 3–52.
Siegrist, J.1996. Adverse health effects of high-effort/low-reward conditions. Journal of Occupational Health Psychology, 1: 27–41.
Siegrist, J., Starke, D., Chandola, T., Godin, I., Marmot, M.Niedhammer, I.2004. The measurement of effort-reward imbalance at work: European comparisons. Social Science & Medicine, 58: 1483–1499.
Stone, A. A. and Shiffman, S.1994. Ecological momentary assessment (EMA) in behavioural medicine. Annals of Behavioral Medicine, 16: 199–202.
Stone, A. A., Turkkan, J. S., Bachrach, C. A., Jobe, J. B., Kurtzman, H. S., & Cain, V. S. (Eds.). (2000). The science of self-report: Implications for research and practice. London: Lawrence Erlbaum Associates.
Theorell, T., Perski, A., Akerstedt, T., Sigala, F., Ahlberg-Hulten, G.Svensson, J.1988. Changes in job strain in relation to changes in physiological state: A longitudinal study. Scandinavian Journal of Work, Environment and Health, 14: 189–196.
van der Doef, M. and Maes, S.1998. The job demand-control(-support) model and physical health outcomes: A review of the strain and buffer hypotheses. Psychology and Health, 13: 909–936.
van der Doef, M. and Maes, S.1999. The job demand-control(-support) model and psychological well-being: A review of 20 years of empirical research. Work & Stress, 13: 87–114.
van Eerde, D., Holman, D. and Totterdell, P.2005. Diary studies in work psychology. Journal of Occupational and Organizational Psychology, 78: 151–154.
van Vegchel, N., de Jonge, J., Bosma, H. and Schaufeli, W.2005. Reviewing the effort-reward imbalance model: Drawing up the balance of 45 empirical studies. Social Science & Medicine, 60: 1117–1131.
Williams, S., Michie, S. and Pattani, S.1998. Improving the health of the NHS workforces: Report of the partnership on the health of the NHS workforce, London: The Nuffield Trust.
By DerekW. Johnston; Alexis Beedie and MartynC. Jones
Reported by Author; Author; Author