Managing Safety at Work - Research Paper Example

Running Head: PATIENT SAFETY Patient Safety Introduction Providing safe care is the number one priority of health-care professionals. Safe care is associated with procedural efficiency, implementation of evidence-based standards and the use of tools designed to reduce the likelihood of medical error (such as computerized medication orders and bar-coded patient identification (e.g. Patterson et al. 2006, Press). Overall, there is good agreement that stress at work contributes to impaired health and well-being (e.g. Semmer et al. 2005a, 2). The association between work stress and safety (e.g. patient safety) is less well understood. It does seem likely, however, that, under stress, hospital staff is more likely to make mistakes that can contribute to the emergence of accidents, because high stress levels can impair the level of concentration, cognitive information processing, decision-making and work behavior (Furney, Press). Research Questions The main research question that guided the present study was whether objective work stressors and resources predict important characteristics of stressful events that are related to safety among novice nurses. The characteristics in question were the similarity of the particular event to previously experienced events and the probability that the event would recur in the future. Thus, the risk potential contained in the event was assessed in terms of the event being a recurrent rather than a singular event - either because it had occurred in the past and/or because it was estimated to have a high chance of recurrence. Work stressors and job control both assessed by observer ratings were respectively expected to be positively and negatively associated with the two indicators of risk potential. Literature Review Work stress, especially work overload (e.g. by time pressure) has been shown to be associated with occupational accidents (e.g. Frone 1998, 565, Zohar 2000, 591) and medical malpractice (Jones et al. 1998, 730). Conversely, resources such as job autonomy or job control are positively related not only to health and productivity but also to safe working practices (Parker et al. 2001, 214). This study investigates the association between chronic job stressors and characteristics of safety-related stressful events encountered at work. Typically, in this kind of research, both predictors and dependent variables are assessed by means of self-report, and this presents a risk of overestimating the association because of correlated error (Semmer et al. 2004, 210). As such, many researchers recommend measuring independent and dependent variables with different methods (e.g. Kristensen 1995, 18, cf. also Rick et al. 2001). In the present study, trained observers rated the level of work stressors and resources on the job, while young nurses made a record of every stressful event during a working week. Regarding bias in self-report of safety-related events, such as accidents, near-accidents, errors and near-misses with regard to medication, etc., there is no reason to assume that people tend to exaggerate them. Quite the opposite - there is a tendency not to report safety-related events because they are a potential threat to the self (e.g. Davidoff 2002, 623). Therefore, the instructions emphasized that every stressful event should be reported and they emphasized anonymity of responses in order to encourage truthful reporting. Methodology Sample The nurses were a sub-sample of the participants in the project. The study took place in the first 18 months after the nurses' vocational training was complete. The first phase of data collection started in September 1997 (0-1 months of work experience) and was completed in spring 1998 (5-6 months of work experience). The second phase started in September 1998 (12-13 months of work experience) and was completed in spring 1999 (17-18 months of work experience). As such, 2 working weeks 10-12 months apart were studied. Out of the 40 nurses. Work stress and patient safety 459,35 completed both phases. Two nurses were excluded from the analyses because they changed employer and ten cases were lost because of missing values in the study variables, mainly because workplace observations were not possible. Thus, complete event diaries and observer data were available for 23 nurses (21 female, two male). Their mean age was 24 (SD 3.7) years. During the study period the nurses were employed in 19 hospitals in the German-speaking part of Switzerland. All nurses were full-time employed and there was no significant change of task and team during the observation period. Informed consent was obtained from all participants. Measures Situation-related measures: Stressful events were measured by event sampling, using a paper-and-pencil version (pocket diary) of the computer assisted self-observation system (see Perrez and Reicherts 1996, 55). This measure captures both the content and process of coping in a stressful episode. Participants were instructed to document every stressful situation they experienced, both minor and major, over a 7-d period (5 working days and 2 days off). The answering format referred to intensity on a 6-point Likert scale ranging from 1 (very low) to 6 (very high). Situational control was assessed with one item: 'The chances that I could change the situation for the better were 1 (very small) to 6 (very large)'. Person-related measures: Compliance with safety regulations was measured with a single questionnaire item that included a comment in parentheses: 'I pay attention to safety at work (compliance with safety regulations, consideration of recommended means to protect the back and the eyes, etc.)', with a 5-point Likert scale, 1 (seldom) to 5 (always). Chronic job characteristics were assessed by observation data on chronic job stressors and job control. Observation was complemented by an interview with the employee (30-45 min), questioning of the supervisor(s) and colleague(s) and consultation of organizational documents, if necessary. The internal consistencies for the scales were satisfactory (table 1). Data Collection On the morning of the first day, participants were met by trained data collectors at the workplace. They received several pocket diaries and were given individual instruction as to how stressful events during the day should be documented as soon as possible after they occurred. The data collectors reminded participants personally and repeatedly during the first day to fill in their diaries each time they had a stressful experience (data collectors were present on the first day for the whole work shift, as they were involved in observational tasks and interviews, cf. Grebner et al. 2004, 13, Elfering et al. 2005, 237). Data Analysis For the analyses reported here, a special, and safety related coding system was developed (for the more general coding system normally used in this project see Grebner et al. 2004,16). Categories referred to issues of safety, such as 'documentation error', 'medication error', etc. Examples are shown in table 2. Two judges independently categorized the open-ended verbal descriptions of the stressful situations; inter rater-reliability (Cohen's K; Cohen 1990, 38) was 0.74 for all situation categories. Where the judges had disagreed in their initial assessments, consensus ratings were made. Multilevel Regression Analyses Only situations that were safety-related were included in the analyses reported here. The data contain information at the person-level and the situation-level, with stressful situations nested within persons. This leads to a number of statistical problems. Aggregating information results in a loss of situational information (as well as power); analysing the data on the situational level results in the dependency among variables being ignored, leading to potentially spurious results. The variance components model does not include any explanatory variables but only intercept variances as random parameters at each level (cf. Hox 2002). For each situation-related predictor variable tested to know whether a random model fits the data better. These random effects were tested by allowing the slopes to vary across persons (slope variance as first random parameter) and by estimating the covariance between slopes and intercept (second random parameter). Significant estimates of slope variance indicate cross-level interactions. Multilevel models are based on 62 situations reported by 23 persons (see table 1). Results Table 2 shows the frequency and content of the safety-related stressful events. Overall, i.e. including non safety-related events, 439 situations were reported over the two 7-day periods. Of these, 125 were private events and 314 were work-related. This corresponds to an average of 13.65 work-related events per person (SD 7.1). Of the 314 work-related events, 62 (19.7%) were coded as relating to patient safety, corresponding to an average of 2.65 events per person (SD 1.5). Per individual, between 10 and 40% of events were safety-related, with a median percentage of 20. Most safety-related events addressed documentation (25 events or 40.3%) and near misses in medication (13 events or 21%). Forgotten or incomplete briefing of patients and delays in delivery of patient care constituted another 14.5%. Most other events addressed patient casualties (four events or 6.5%) and violence among patients or towards nurses (six events or 9.7%). Table 2 shows various examples for the categories. There were more safety-related events reported in the first phase compared with the second phase, although this decrease did not reach statistical significance. Furthermore, there was no relevant change in the frequency of specific events, with the exception of near misses in medication, which decreased significantly with increasing work experience (p < 0.05). Near misses in medication were among those safety-related events that appeared to be less familiar than others. Table 1 shows the mean values for all study variables and table 3 shows their inter correlations. The high level of work interruptions and low level of job control compares unfavorably with the data reported for individuals who are new to the job in other professions (Kalin et al. 2000, 282). Correlations among stressors and between stressors and job control were moderate. All other correlations were also modest, with the exception of the association between outcome variables, which was 0.85. This is not surprising because one is enquiring about the frequency of similar events in the past (familiarity) and this should constitute a basis for predicting the second outcome variable, i.e. the probability of future episodes of the same kind. Both dependent variables were positively associated with concentration demands and negatively related to job control (table 3). The intercept-only-model (without any explanatory variables) yielded an ICC of 0.35 for familiarity of safety-related events and an ICC of 0.42 for the probability of recurrence. Thus 35 and 42% of the total variance in safety-related risk potential was located at the person-level and 65 and 58% at the situation-level. In multilevel regression analyses concentration demands significantly predicted higher risk potential in terms of both dependent variables, after individual safety compliance, other work factors and situational control were controlled for. Table 4 shows that uncertainty was a significant predictor of probability of recurrence, but not of familiarity, a difference justifying a separate analysis for each outcome variable. The third significant work factor was job control, with low control predicting higher risk, i.e. similarity and recurrence of events. Safety compliance was positively and situational control was negatively related to both outcome variables, but the partial regression coefficients for these were not statistically significant. Generally, the variation of the intercept across participants was not significant for both dependent variables. This indicates that significant differences between individuals in mean event-related familiarity and probability of recurrence did not remain when all predictors were in the model. Thus, the inclusion of predictors led to a reduction of individual variability. There were no indications of individual differences in the impact of situational characteristics on outcome variables (significant variation in slopes), as no situational predictor variable showed a significant random effect. Discussion and Conclusion This study investigated the relationship between work characteristics and the no singularity of safety-related events, using observer ratings of stressors and control as indicators of work characteristics and the diary-reported characteristics of safety-related Predictions of the non-singularity of safety-related events by work characteristics and situational characteristics in multilevel analyses. Events as outcomes the first important result is a descriptive one: Roughly 20% of all events reported were coded as being safety-related. One can reasonably assume that the number of events reported altogether represents, if anything, an underreporting, since reporting may be associated with anxiety and shame (DavidofT 2002, Firth-Cozens 2002, 115). In general, taking out the diary and reporting events is perceived to be tedious, especially when under stress (cf. Semmer et al. 2004, 212). This implies that the average 2.69 safety-related events, reported over 10 working days, likely represent a somewhat conservative estimate. The second important finding is that the non-singularity of safety-related events is related to stressors, most notably concentration demands, and to lack of control. Put differently, when people work under high demands and low control, events that endanger safety are experienced as more familiar, and more likely to recur. This association cannot be explained in terms of reporting bias, e.g. people claiming unfavorable working conditions also overestimating the work hazard, because in this study work characteristics were assessed by observer rating rather than self-report; in this sense, it is difficult to attribute the results to common method variance (Semmer et al. 2004, 215). Recommendations This study, adds to the growing literature that suggests a relationship between (medical) error and stress (e.g. Aiken et al. 2002, 288, Barling and Zacharatos 1999, 9). How can such a relationship be conceptualized? The most likely explanation lies in changes in the regulation of actions: 'secondary tasks' (such as double-checking, as well as documenting) may be reduced, and actions may be carried out less attentively, resulting in reduced monitoring (cf. Schonpflug and Battmann 1988, 700, Hockey 1997, 75, Matthews et al. 2000, Press). Self-regulation may be impaired (Baumeister et al. 1999, 66), leading to, for instance, less competent social behaviour, which may affect the behaviour of patients. Training of novice nurses should address the association between workload and patient safety, and the processes behind these, and educate nurses in self-management strategies for stressful situations. Limitations A weakness of this study was certainly the small sample size, which reduced the power of the analyses. While the researchers are quite certain about the effects that were found, the possibility cannot be ruled out that there may have been additional effects (especially interaction effects) that could not be detected with the given sample. One further methodological point should be mentioned. Otie may speculate that the greater the stress experienced as a result of an event, the greater the time that elapses until that event is recorded. In the present study, both points of time were recorded and there was a positive but non-significant relationship (r = 0.21) between the two. Ethical or Political Considerations Hence, it is unlikely that the reliability of recordings showed any systematic dependence on the stressfulness of events. The sample in this study consisted of youtig nurses. In general, young workers have an especially high risk of non-fatal workplace accidents and injuries (Castillo 1999, 160). Frone (1998, 566) showed that work overload and boredom were each related to young workers' job injuries. Novice and experienced registered nurses differ in experience (Benner et al. 1996, 1, Ebright et al. 2004, 533). The expert nurse practice is focused on informed action, based on knowing both the patient and skilled involvement, whereas novice nurse practice centers on organization, priority setting and task completion. Novice nurses develop and refine their clinical and care-related knowledge over time. High task load and low control at work can inhibit the development of safety expertise. For example, overwork, understaffing and low time control may reduce the contact time with patients to the point that it becomes inadequate for providing the attentive care that is necessary for the timely recognition of subtle signs of safety-related changes in the patient's condition. This suggests that these results may be especially valid for novice nurses. References Aiken, L H., Clarke, S.P., Sloane, D.M., Sochalski, J. and Silber, J.H., 2002, Hospital nurse staffing and patient mortality, nurse burnout, and job dissatisfaction. Journal of the American Medical Association, 288. Barling, J. and Zacharatos, A., 1999, High performance safety systems: ten management practices for creating safe organizations. In Managing Safety at Work: Beyond Blaming the Individual, S.K. Parker and M.A. Griffin (Chairs), Symposium conducted at the Academy of Management Conference, 9-11, Chicago, IL. Baumeister, R.F., Faber, J.E. and Wallace, H.M., 1999, Coping and ego depletion. In Coping: The Psychology of What Works, C.R. Snyder (Ed.), pp. 50-69 (New York, NY: Oxford University Press). Benner, P., Tanner, C. and Chesla, C , 1996, Expertise in Nursing Practice: Caring, Clinical Judgment, and Ethics (New York: Springer), pp. 1-2. Castillo, D.N, 1999, Occupational safety and health in young people. In Young Workers: Varieties of Experience, J. Barling and E.K. Kelloway (Eds.), pp. 159-200 (Washington, DC: APA). Cohen, J., 1990, A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37-46. Davidoff, F., 2002, Shame: The elephant in the room. British Medical Journat, 324, 623-624. Ebright, P.R., Urden, L., Patterson, E. and Chalko, B., 2004, Themes surrounding novice nurse near-miss and adverse-event situations. Journal of Nursing Administration, 34, 531-538. Elfering, A., Grebner, S., Semmer, N.K., Kaiser-Freiburghaus, D., Lauper-Del Ponte, S. and Witschi, 1., 2005, Chronic job stressors and job control: Effects on event-related coping success and well-being. Journal of Occupational and Organizational Psychology, 78, 237-252. Firth-Cozens, J., 2002, Anxiety as a barrier to risk management. Quality and Safety in Health Care, 11, 115. Frone, M.R., 1998, Predictors of work injuries among employed adolescents. Journal of Applied Psychology, 83, 565-576. Furney, S.R., 1986, Stress Influences on Accident Incidence. Human Stress: Current Selected Research (New York: AMS Press). Grebner, S., Elfering, A., Semmer, N.K., Kaiser-Probst, C. and Schlapbach, M.L., 2004, Stressful situations at work and in private life among young workers: An event sampling approach. Social Indicators Research, 67, 11-49. Hockey, G.R.J., 1997, Compensatory control in the regulation of human performance under stress and high workload: A cognitive-energetical framework. Biological Psychology, 45, 73-93. Hox, J.J., 2002, Multilevel Analysis (Mahwah, NJ: Lawrence Erlbaum). Jones, J.W., Barge, B.N., Steffy, B.D., Fay, L.M., Kunz, L.K. and Wuebker, L.J., 1998, Stress and medical malpractice: Organizational risk assessment and intervention. Journal of Applied Psychology, 73, 727-735. Kalin, W., Semmer, N.K., Elfering, A., Dauwalder, J.-P., Tschan, F., Heunert, S. and Crettaz Von Roten, F., 2000, Work characteristics and well-being of Swiss apprentices entering the labor market. Swiss Journal of Psychology, 59, 272-290. Kristensen, T.S., 1995, The Demand-Control-Support Model: Methodological challenges for future research. Stress Medicine, 11, 17-26. Matthews, G., Davies, D.R., Westerman, S.J. and Stammers, R.B., 2000, Human Performance: Cognition, Stress and Individual Differences (Hove, UK: Psychology Press). Parker, S.K., Axtell, CM. and Turner, N., 2001, Designing a safer workplace: Importance of job autonomy, communication quality, and supportive supervisors. Journal of Occupational Health Psychology, 6, 211-228. Patterson, E.S., Rogers, M.L., Chapman, R.J. and Render, M.L., 2006, Compliance with intended use of bar code medication administration in acute and long-term care: an observational study. Human Factors, in press. Perrez, M. and Reicherts, M., 1996, A computer-assisted self-monitoring procedure for assessing stress-related behaviour under real life conditions. In Ambulatory Assessment: Computer-Assisted Psychological and Psychophysiological Methods in Monitoring and Field Studies, J. Fahrenberg and M. Myrtek (Eds.), pp. 51-67 (Kirkland, WA: Hogrefe & Huber). Rick, J., Briner, R.B., Daniels, K., Perryman, S. and Guppy, A., 2001, A critical review of psychosocial hazard measures. Heath and Safety Executive contract research report 356/2001. Available online at: http://www.hse.gov.uk/research/crr01356.pdf Schonpflug, W. and Battmann, W., 1988, The costs and benefits of coping. In Handbook of Life Stress, Cognition and Health, S. Fisher and J. Reason (Eds.), pp. 699-713 (New York, NY: Wiley). Work stress and patient safety 469 Semmer, N.K., Grebner, S. and Elfering, A., 2004, Beyond self-report: Using observational, physiological, and event-based measures in research on occupational stress. In Emotional and Physiological Processes and Positive Intervention Strategies, Research in Occupational Stress and Well-being, P.L. Perrewe and D.C. Ganster (Eds.), Vol. 3, pp. 205-263 (Amsterdam: JAI). Semmer, N.K., Mcgrath, J.E. and Beehr, T.A., 2005a, Conceptual issues in research on stress and health. In Handbook of Stress and Health, C.L. Cooper (Ed.), 2nd ed., pp. 1^3 (New York: CRC Press). Zohar, D., 2000, A group-level model of safety climate: Testing the effect of group climate on microaccidents in manufacturing jobs. Journal of Applied Psychology, 85, 587-596. Read More