Introduction
Aggressive behaviour exhibited by patients with mental illness is a major occupational challenge and hazard faced by mental health nurses. (Beech & Leather 2006). On the one hand, such behaviour puts nurses under extreme stress; on the other hand, it grossly undermines the quality of care in many instances as a result of a breakdown in carer-caregiver relationships. A British Audit Office survey in 2005 reported that violence and aggression could account for up to 40% of health and safety problems for workers in healthcare, but the actual figure might be considerably higher due to the high prevalence of unreported cases. (Beech & Leather, 2006). The high prevalence of this condition has prompted a huge number of training programs aimed at reducing the incidence and impact of such problems on healthcare staff. However, proper evaluation studies of the effectiveness of such programs are relatively rare. The paucity of systematically evaluated training programs was pointed out by Farrell & Cubit (2005), as the process poses multiple challenges and problems with design and execution. Other relevant factors in this regard include high costs of such studies, fears that negative results might harm commercial interests and, in some cases, a lack of expertise in carrying out such evaluation studies. (Holt, Boehm-Davis, & Beaubien, 2001). The current prospective study addresses this issue through a prospective, multi-centre randomized and controlled trial and attempts to provide a much-needed answer to a central question faced by many nursing educators: does training in aggression management alter the perception of aggressiveness, and thereby potentially lead to improvements in the quality of nursing?
Analysis of design factors
Being essentially a randomized controlled study, the design of the study was to compare two groups those who did and those who did not receive aggression management training. Their attitudes were assessed before and after the intervention in the study group at an interval of three months. The problem with this element of design is the fact that three months can be a long time for forgetting to occur about the elements learned in the training course unless efforts were made to help them recall the material at various stages or incorporate it into their work routine. No mention is made in the study methodology whether this was done. As a result, it is possible that a proportion of ‘trained’ nurses might have forgotten the instructional elements to the extent that it did not cause any change in their attitudes compared to the control group. If the re-assessment was done sooner, the scores might have been significantly different. Moreover, different people have different capacities for recall and learning, and therefore it is quite possible for numerous biases to creep into the study design.
There is yet another important factor that needs consideration. In the study, the control and study groups were six different acute psychiatric wards across German-speaking Switzerland. There are wide differences in the nature and severity of illnesses seen in the different patient populations, as well as restraints and safety practices. In addition, some nurses might already have had significant training and exposure to aggression management practices, which might mean that the program is administered would have had relatively little impact on the post-test scores.
Despite randomization and control, therefore, it is impossible to be absolutely sure that like is being compared with like, and therefore conclusions being drawn must be interpreted with caution due to the various confounding factors discussed above.
The Intervention itself
The training program designed by Oud (1997) combines various theoretical aspects of managing aggression from institutionalized patients and consists of 10 modules of 50 minutes duration each spread over 5 days. Although it appears to be comprehensive in its scope, it appears to concentrate more on the theoretical aspects of the problem of aggressiveness. The structure of the course itself allows little time for participants to reflect on the contents. In addition, rapid presentation of new material in succession can lead to forgetting of the old material, and often courses spread over a longer duration have much more retention.
In this regard, evaluation of training models need to be closely considered, and the best-known model is that of Kirkpatrick (1994). According to Kirkpatrick, there are four measures of the effectiveness of training programs. These are:
- Reaction: This is a measure of the participants’ immediate reaction to the learning material, after completion of course material;
- Learning: This refers to the improvement in awareness and knowledge directly as a result of the training program;
- Behaviour: This refers to the consolidation of knowledge following the program which leads to a measurable change in one’s work pattern or skills;
- Results: This refers to the effect of the learning on the organization’s functioning as a whole and often requires a long period of time to fully evaluate.
It has been observed ( Alliger et al. 1997; Arthur et al. 2003) that most studies tend to concentrate on the reaction aspect because it is relatively simple and easy to do so. However, the interpretation of the results of reactions need not necessarily reflect the true effectiveness because they do not objectively measure the overall impact of training. The current study is somewhat of a mix – the fact that the assessments were carried out after three months implies that apart from measuring a delayed response, it was also intended to assess learning, at least in the short term.
Analysis of questionnaire
The questionnaire was based on three standardized tests – perception of aggression, tolerance measure and the IMPACS, measuring the impact of aggression on the carers.
The perception of aggression scale (PSOS-A) formulated by Needham et al. (2004) measures nurses’ perceptions of their patients’ aggressive behaviour on a bipolar scale of 6-30, with ‘functional, understandable’ behaviour at one end, and ‘undesirable, dysfunctional’ behaviour at the other end of the spectrum. Although this is a validated measurement tool, its applicability to the current situation must be questioned. Firstly, individual perceptions vary widely for any given circumstance, and how someone views aggression is shaped by a host of complex factors, often stemming from one’s life experience starting from childhood. Inevitably some of us view such behaviour with more positivity than others, and these deep-seated attitudes are difficult to change with a short course of training. Secondly, as most healthcare workers would agree, the context of aggression has a significant effect on the perception it produces. Someone under acute psychological stress as a result of victimization is much more likely to be perceived as displaying ‘functional’ behaviour compared to someone who is displaying identical aggression under normal circumstances. The context can vary between different types of institutions, and the type of care provided, as well as the length of time someone has been under their care.
Somewhat similar considerations apply to the Tolerance scale (Whittington, 2002) that has been utilized. It has been mentioned that this scale has not been used to evaluate a program’s success in influencing attitude. However, there could be several reasons for it. First and foremost, individual capacities for tolerating adverse situations vary tremendously, and this would be reflected in the scores. Secondly, because attitudes are deeply ingrained, relatively few participants are expected to display anything but a moderate shift in their tolerance. Moreover, as participants in the study, which is being carried out by internal departments, it is quite possible there is some indirect pressure to make the intervention a success, and staff would be keen to demonstrate anyway that training has been beneficial for their personal learning. This ‘pressure’ to conform might be an important reason why tolerance questionnaires might not be applicable in the current situation – few would state that their attitudes ‘did not change’ despite being trained to do so.
Finally, the ‘Impact of Patient Aggression on Carer’s Scale’ (IMPACTS) is a relatively new metric designed by the authors of the study themselves. It is designed to measure the impact of ‘the actions of handling adverse circumstances’ on the carers themselves, and training it is hoped makes a big difference in the way such incidents are ‘professionally’ handled. This is perhaps the most measurable amongst the three parameters being measured – to demonstrate that training makes a difference.
Data Analysis
All the data in the study and control groups were compared statistically using appropriate tests – to prove or disprove the null hypothesis beyond reasonable doubt of Type I error. However, this does not remove the possibility of numerous biases creeping in.
Cook and Campbell (1979) pointed out that in a significant proportion of controlled studies, threats to internal validity remain as the study and control groups are not strictly comparable. In the current study, as we have pointed out before, there are areas where there confounding factors that make the groups difficult to compare.
All three variables, in the end, failed to make any statistically significant impact on attitudes towards aggression, and the non-significance of the study raises several issues. This has been observed despite the fact that the nurses themselves were appreciative of the training program and found it to be beneficial. Obviously, the feeling of usefulness did not translate to statistical difference, but it may be a reflection of the study characteristics rather than the usefulness of the training program. Whether the training program in question was aimed at specifically increasing awareness about the various issues related to the aggressiveness that nurses face at work remains unanswered. It is quite possible that the program itself was too compact and hurried to make a significant difference in attitudes. The authors have themselves considered the possibility that the impact on attitudes that they were looking for perhaps only happens in the longer term. Interestingly another quasi-experimental study has been mentioned in the same region in Switzerland using the same training program (Hahn, 2003), which also did not find any detectable difference in the attitudes of healthcare workers towards aggression. Perhaps more studies are required to assess whether training programs do cause any detectable change in attitudes compared to other long term approaches like changing organizational culture or create ongoing educational programs and activities that are targeted towards attitudinal change. The authors admit that three months was probably too optimistic to anticipate any real change. However, there was no continuity in the training program, too – and in this situation, only a few motivated participants would perform better than the rest. The recruitment of nurses for this study also is another important issue – it does not appear they were volunteers, but all in a hospital were invited to attend as their institution was randomly chosen. This means that compared to ‘enthusiastic’ volunteers who could have possibly scored higher in the post-test results, this was an unmixed population with less polarized results. The other factor that has been mentioned is that the mean age of the nurses is 38, and as a result, possibly they were not as enthusiastic as a younger cohort – perhaps this statement is somewhat ageist to assume that training has lesser effects on older cohorts: compared to this a more appropriate statement would have been ‘perhaps the training modality was not suitable for experienced professionals.’
In addition, both arms of the study contained too few subjects – sample sizes of only 25-30 per arm. Although a power analysis has been done, this seems to be too small a sample size to detect any difference.
Secondly, the sensitivity of the instruments used in this study also needs to be questioned closely. Of the three questionnaires, IMPACS has never been used to evaluate the impact of any intervention – therefore, its sensitivity is difficult to assess.
Other studies in the past have assessed other modalities like the patients’ responsibility for aggressive action (Ryan and Poster 1989) or nurses’ reactions to aggressive behaviour ( Collins 1994). The issue of whether a change in attitude towards aggressiveness is essential for bringing in change in working behaviour has never been fully explored. The desired endpoint for any training is to improve one’s effectiveness at work, and in this respect, attitudinal change is definitely harder to achieve compared to changes in behavioural response on the part of carers through training. The purpose for this study is, therefore, somewhat unclear – whether attitudinal change is desired as a prerequisite for change ineffectiveness at work or is a desired end-point in itself. If it was clear, maybe the study design would have been better planned to reflect those changes.
Overall conclusion about the study
If we consider all the parameters we have discussed so far, perhaps the experiment itself is somewhat experimental. Changes in attitude are difficult to quantify and even more difficult to compare, as there are so many variables that cannot be controlled. In real-life organisations, change in attitudes only come through change management programs that are specifically designed to do so – incorporating newer paradigms, change in organisational culture and often introducing radically new models and ideas. It is somewhat simplistic to assume that a mere five-day course would result in a significant attitudinal change for inexperienced nurses regarding something they had experienced in their entire careers and lives. Perhaps the aim of the study was just that – to demonstrate the relative ‘uselessness’ of short courses.
Randomized controlled studies have certain pre-conditions to be meaningful, and the presence of successful statistical analysis does not in itself lend any merit to the paper. One of these important pre-conditions – i.e. comparable groups, is somewhat suspect in the study. The control group, for instance, without any training, could perhaps be better substituted by a different sort of control, e.g. audiovisual training, or attending an interactive, informal workshop. Alternatively, a placebo group who were delivered a training program unrelated to the attitude question against aggression, e.g. how to deal with stress at work, could also serve as logical comparatives.
Training healthcare staff to deal with aggressiveness is extremely important – and research should ideally try to answer the question about how to best administer such training. Whether attitude change precedes a change in work practice or it works the other way round is an important consideration for educators. But for nurses and other related healthcare professionals, the important question is: what is the best approach to deal with problems related to aggressiveness at the workplace? Is it through formal training or through a long process of personal learning in small groups, lifelong?
The main value of this paper is that it places attitudes above quick-fix training: attitudes that can only be imbibed with self-reflection and learning through experience. Attitudes towards aggressiveness are described as ‘more stable than thought’, echoing the sentiment that one’s negative attitudes towards difficult work issues can only be resolved through personal effort, maturity and above all, experience. Even the most ‘helpful’ and elegant training methods cannot always successfully substitute for the above skills, which can only be earned.
References
Alliger, G.M., Tannenbaum, S.I., Bennett, W., Jr., Traver, H., & Shotland, A. (1997). A meta-analysis of the relations among training criteria. Personnel Psychology, 50, 341–358.
Arthur,W., Jr., Bennett,W., Jr., Edens, P.S.,&Bell, S.T. (2003). Effectiveness of training in organizations: A meta-analysis of design and evaluation features. Journal of Applied Psychology, 88, 234–245.
Beech, B., & Leather, P. (2006). Workplace violence in the health care sector: A review of staff training and integration of training evaluation models. Aggression and Violent Behavior, 11, 27–43.
Collins, J., 1994. Nurses’ attitudes towards aggressive behaviour, following attendance at ‘‘The prevention and management of aggressive behaviour programme’’. Journal of Advanced Nursing 20, 117–131.
Cook, T.D., & Campbell, D.T. (1979). Quasi-experimentation: Design and analysis issues for field settings. Boston: Houghton Mifflin.
Hahn, S., 2003. Effekt einer Schulung auf die Haltung von Pflegepersonen zu Ursache und Management von Patientenaggression. In: Faculty of Health Sciences. Rijksuniversiteit Limburg, Maastricht, 59 pp.
Holt, R.W., Boehm-Davis, D.A., & Beaubien, J.M. (2001). Evaluating resource management training. In E. Salas, C. Bowers, & E. Edens (Eds.), Improving teamwork in organisations: Applications of resource management training (pp. 165–190). Mahwah, NJ: Erlbaum.
Kirkpatrick, D.L. (1994). Evaluating training: The four levels (2nd ed.). San Fransisco: Berrett-Koehler.
Needham, I., Abderhalden, C., Dassen, T., Haug, H.J., Fischer, J.E., 2004. The perception of aggression by nurses: psychometric scale testing and derivation of a short instrument. Journal of Psychiatric and Mental Health Nursing 11 (1), 36–42.
Oud, N.E., 1997. Aggression and psychiatric nursing. Broens and Oud: partnership for consulting and training, Amsterdam, 33 pp.
Ryan, J.A., Poster, E.C., 1989. The assaulted nurse: short-term and long-term responses. Archives of Psychiatric Nursing 3 (6), 323–331.
Whittington, R., Higgins, L., 2002. More than zero tolerance? Burnout and tolerance for patient aggression amongst mental health nurses in China and the UK. Acta Psychiatrica Scandinavia 106 (Supplementum 412), 37–40.