|Copyright 1997 by the||Psychology in Spain, 1997, Vol. 1. No 1, 63-71|
|Colegio Oficial de Psicólogos|
This study examined occupational stress among mental health professionals (staff and trainees). The sample consisted of 36 physicians and 22 psychologists from different public centres in Galicia, Spain. The variables evaluated included stress situations at work, psychophysiological symptoms and some social and work indicators. Our results showed that psychiatrists experienced less stress than their colleagues, and suggest specific associations between situational stressors and state of health. In particular, frustration in carrying out their work is linked to high levels of self-reported symptoms in psychologists. For psychiatrists, certain complications in the treatment of patients, or their progressive deterioration, are important.
Se analiza el estrés que experimentan profesionales clínicos (especialistas y residentes) que trabajan en el campo de la salud mental. Componen la muestra 36 médicos y 22 psicólogos de distintos dispositivos asistenciales públicos de Galicia. Se recogieron los situaciones laborales más estresantes, síntomas psicofisiológicos y diversos indicadores sociolaborales. Los resultados indican que los psiquiatras experimentan menos estrés que otros colegas, detectándose asociaciones significativas entre ciertos estresores y el nivel de salud. En concreto, en los psicólogos la frustración en el desempeño de su trabajo se relaciona con el nivel de síntomas autoinformados. Para los psiquiatras son importantes determinadas complicaciones en el tratamiento o el deterioro que sufren algunos pacientes.
The original Spanish version of this paper has been previously published in Clínica y Salud, 1996, Vol. 7 No 2, 163-180
*Correspondence concerning this article should be addressed to Víctor M. Torrado Oubiña. Servicio de Psiquiatría. Complejo hospitalario "Xeral de Galicia-Gil Casares". Santiago de Compostela. C/Galeras s/n. 15705 Santiago de Compostela. Spain.
linical work in mental health may be considered stressful for a variety of reasons. Certain stress-generating situations are also common in work with other, non-psychiatric patients. Classic examples of these kinds of demands are overwork and the limited capacity to alter the course of certain disorders. Meanwhile, day-to-day work demands a more or less profound empathic relationship with patients over what may be a lengthy period (Pines and Maslach, 1978).
Specifically, the mental health field offers some peculiarities. First, psychiatry has been considered as one of the most stressful medical specialities, whose common link would be attention to severely ill patients with poor hopes of recovery and chronic or incurable diseases (Okinoura et al, 1990). However, there are two kinds of professional in this domain (clinical psychologists and psychiatrists) with an equivalent level of clinical-therapeutic responsibility who carry out their duties on the same sites. Most of the studies in the literature consist of samples of nurses or physicians and, while comparative studies amongst these groups are scarce, research comparing psychologists and physicians are practically non-existent (Alvarez y Fernández-Ríos, 1991).
Since Freudenberger (1974) used the term burnout, it has mainly been used to describe a state of physical and emotional exhaustion whose characteristics have been mostly applied to human services professionals, within which health staff is included. However, nowadays there is a trend to consider burnout as a specific form of occupational stress or a subclass of stress effects in individuals (Shinn, Rosario, Morch and Chesnut, 1984; Moreno, Oliver y Aragoneses, 1991).
According to Lazarus and Folkman (1984), we consider that stress arises when an individual estimates that environmental demands override his/her own adjustment resources. Hence, given existent empirical evidence on the relationship between stress and several negative psychobiological disorders (Skalar, 1981; Bomon, 1988; Herbert and Cohen, 1993), it would be interesting to identify those work situations which are potentially powerful stressors, as well as the nature of their consequences for health. The design of the study should allow us to examine the potential specific differences among occupations. Such matters constitute the objective of our work.
Following Ullrich and Fitzgeralds (1990) framework, stress was evaluated through a series of problematic or conflictive work situations, and through a list of psychophysiological symptoms. A problem situations list was expected to reflect the subjective stress level experienced by professionals, hypothesising that stress would be a good predictor of state of health.
A selection was made from the total psychiatrist, clinical psychologist and trainee population in these disciplines working in public health care services in the Comunidad Autónoma Gallega1 (Autonomous Community of Galicia, NW Spain) during 1994. From this population, the following were excluded:
a) Professionals working in centres in which attention is restricted to a particular type of problem, and whose team did not have a psychiatrist (for example, drug-addiction units and family counselling centres).
b) Professionals representing a kind of attention which does not involve multidisciplinary work, and whose working hours are fewer than the rest (for example, limited numbers consulting).
c) Professionals with less than one years experience.
Finally, the group of subjects who agreed to participate comprised the studys sample, which consisted of 58 subjects (41 males and 17 females), whose socio-demographic characteristics are described in Table 1. Sample mean age was 36.6 years, and experience in clinical work ranged from 2 to 28 years (mean = 10.45). Most subjects professed to having an eclectic approach to psychotherapy (20%).
A pilot-questionnaire was drawn up from the most frequently-mentioned problems in both health staff in general and personnel in the psychiatric domain, collected from previous studies in the literature (Payne and Firth-Cozens, 1987; Ullrich and Fitzgerald, 1990; Reig and Caruana, 1990; Koeske, Kirk and Koeske, 1993). This questionnaire was given to a small group of professionals whose comments were useful for refining the instrument and adapting it to the target population. The final questionnaire comprised the following aspects:
Conflict situations: This consists of a series of 58 items describing typical occupational stress situations. Subjects had to indicate whether the situation constituted a stress problem in their job (item relevance) and, if the answer was affirmative, to state the degree of emotional impact perceived (item severity), for which they were given a graded numerical scale from 1 ("slight stress") to 5 ("intense stress").
As they were presented to subjects, items were provisionally grouped in conflict situation subscales, or stressful situations, as they will be termed from now on. A series of items analysis provided by the statistical programme led to a small re-assignment of these. Most of the definitive scales were reliable and significant (see Table 2).
Sociodemographic aspects. This section contains matters related to age, sex, work centre, experience and type of post, among others. Some of the variables obtained in this section were used to analyse representativeness of the obtained sample with respect to total professionals population.
Health. Subjects completed a conventional list of 21 psychophysiological symptoms frequently related to burnout and stress, developed from Aros (1981) Stress Symptoms Scores and Cronkite and Moos (1984) Physical Symptoms -see Appendix A. Symptom occurrence during the previous 12 months was evaluated according to 4 response categories: Always, Continually, Quite often, Sometimes and Never. In order to obtain a global score that took into account number as well as frequency of the reported symptoms, responses were progressively scored, so that "Always" category was scored with 3 and "Never" with 0 (zero). The algebraic sum of scores in all the symptoms made up the global symptoms or physical complaints score.
Job dissatisfaction. A measure of occupational dissatisfaction was obtained by using Cooper, Watts, Balioni and Kellys (1988) Job Dissatisfaction Measure, translated by the authors of the present work (Appendix B). This scale consists of 5 statements that must be evaluated through a Likert-type 5-point scale ranging from "Totally agree" to "Totally disagree". A high score in this scale indicates high job dissatisfaction.
After a name list had been made, a total of 158 questionnaires were mailed to the chosen professionals according to the already-described criteria. Questionnaires were personally addressed to the workplace of every professional interviewed, with a letter asking for their co-operation and guaranteeing anonymity. Mailing back the questionnaire was facilitated by free postage.
Sixty questionnaires were received, two of which were not statistically suitable. 37% of questionnaires were answered, with a slightly higher rate for males (39%) than for females (32%). By occupations, the percentage of returned questionnaires was: 30% for psychiatrists, 38% for psychologists and 54% for resident internal staff. In child and youth mental health units (MHU) and psychiatric hospitals, questionnaire return rate was similar (34% to 36%) and greater than in general hospitals (24.1%).
Statistical analyses were made using the SPSS/PC+ (v.40) computer program. Results for conflict situations scales are based on perceived stress severity scores (when a subject marked an item as irrelevant, zero score was assigned). Reliability subprogram in SPSS provides adequate procedures for calculating reliability in assessment and item analysis instruments.
In some of the statistical analyses only two groups are mentioned: PCH (psychiatrists) and CPS (clinical psychologists), in which cases training staff in psychiatry (MIR) and psychology (PIR) are included within them. In order to analyse differences in percentages, a Chi-square test was used (with Yates correction when a 2x2 table resulted and a Fisher test in cases where expected frequencies were lower than 5). To study mean differences, a Student t test was used. The remaining statistics are explained in the text.
Responses to the conflict situations questionnaire are summarised in Table 2. As expected, nearly every item identified problem situations for the majority of subjects. Similarity among professionals was high: less than 5% of the items showed significant differences in relevance between the two groups, according to Chi2.
As far as severity is concerned, a series of variance analyses was made following a fixed effect model for studying the influence of occupation on mean stress levels; no significant differences among groups were found in any of the subscales. Stress level profiles were obtained by reducing the sample to two groups (see Figure 1). Thus, PCH group shows stress levels slightly higher than CPS group in several subscales, but differences are only significant for "Identification with the patient" (t = 2.0; p= 0.05).
The two groups coincide on the three subscales reflecting the most important stress situations, though their order is inverted for CPS group ("Therapeutic decisions", "Job criticism" and " Deterioration and complications"), compared to PCH group.
Figure 2 compares symptoms levels between the two groups, with 4 complaints being statistically more frequent among CPS than PCH members: Tiredness/weakness, loss of libido, abdominal pain and nausea/vomiting.
A non-parametric analysis of variance on global scores for physical complaints was made, with occupation as the independent variable. The Kruskal-Wallis (K-W) H statistic provides an analysis of variance based on ordering of ranks. As can be observed in Table 3, psychiatrists are seen to be less affected by physical symptoms in comparison to the other three groups.
Other variables considered in the study, such as therapeutic approach, age, experience and sex, did not show any relation to global symptom level. There is a trend towards higher scores on physical complaints among females than among males (p = 0.06, according to Mann-Whitney U).
Distribution of scores on job dissatisfaction showed a positive asymmetry and a narrow and thin form which does not fit the theoretical distribution of the normal curve according to the Kolgomorov-Smirnov test (p = 0.045). Overall, job dissatisfaction level was not high, and nor were there obvious differences among occupations (K-W: Chi2 = 1.32; p = 0.72). However, this variable maintained some positive correlations with physical symptoms level, which was found to be significant for the established staff but not for the residents (trainees) (Spearmans r = 0.74; p<0.01 among psychologists; 0.44; p<0.05 among psychiatrists). Two stress subscales theoretically related to job dissatisfaction -job criticism and daily work- showed significant covariation with this variable.
The remaining correlations among symptoms levels and conflict situations questionnaire subscales may be seen in Table 4. Clinical psychologists presented high correlations in several subscales, in particular, Job criticism.
Considering specialist psychiatrists separately, the most important subscale is Deterioration/Complications (r=0.37; p<0.05), while for residents it is Daily work (r=0.53; p<0.05).
In order to study the capacities of the different conflict situation subscales for predicting symptoms, a 10-step multiple regression analysis (stepwise) was made.
Among clinical psychologists, the stress derived from the Job criticism scale and the Therapeutic decisions scale explains 55% of variance in self-reported symptoms, though with one important exception: once the first effect of the first variable had been controlled, the "stress" effect generated by therapeutic decisions becomes opposite to that of the first variable. Thus, items receiving high scores on subjective stress do not necessarily generate undesired consequences in the form of symptoms (see Table 5)
As could be predicted from the correlation analysis, in the case of psychiatrists, it was not possible to obtain any predictive variable, neither considering the group as such, nor even grouping them with the residents (MIR-Psychiatry).
In the analysis made with the total sample (N=58), just one variable was significant for predicting 18% of psychophysiological symptoms: Identification with patient. A similar finding was obtained when professionals working in mental health units (MHU) were considered, although 46% of variance was explained in this case. Probably, the group working in these centres is the most homogeneous as regards the way they experience the phenomenon of stress and its sources.
The use of a mailed questionnaire allowed us to obtain a sample of mental health professionals which was considered sufficiently significant in terms of sex, occupation and place of work.
Potentially stressful situations were categorised into 9 subscales reflecting problematic areas in the therapist-patient relationship, in the professionals place of work and in his/her family. Our data indicate a considerable similarity among occupations in relation to perceived stress sources. Academic qualifications (medical doctor or psychologist) did make any difference for the recognition of a situation as stressful (relevance). Also, items with the highest scores on perceived stress intensity were virtually the same for the two professions: Therapeutic decisions and Deterioration/Complications. However, identification with patients suffering (given the often intense emotional involvement with others lives) is a significantly more stressing aspect for psychologists than for psychiatrists. In reference to this last result we may state that the singular characteristics of psychotherapeutic work usually carried out by psychologists considerably contributes to such an evaluation.
As regards physical complaints (tiredness, headaches, backache, etc.), the psychiatrists group generically shows a lower level than the other three groups (clinical psychologists, MIR and PIR).
Bearing in mind that any ranking of sources of stress is open to criticism -since the way the items are phrased may influence their mean values, in this work we tried to relate perceived stress areas to psychophysiological symptoms. We observed that in the multiple regression analysis made on the total sample, only one scale -Identification with patient- is associated with a small proportion of stress symptoms. This situation could occur if we were -as we predicted- looking at a group that was heterogeneous in the form of experiencing the stress phenomenon: that is, stressful situations differ in their power for predicting symptoms depending on the profession considered. In order to verify our hypothesis, we looked to the correlation analysis and, whenever possible, to multiple regression.
Thus, stress symptoms appear to be associated with daily work among residents in psychiatry, while, among psychiatrists, they result from deterioration and complications in patients. Bjorkstein, Sutherland, Miller and Stewart (1983) obtained similar results, noting that work overload and time pressure and their effects on private life are stressors frequently mentioned by medical residents, though this is also a common complaint in students in general. Also, relationships with other professionals, and particularly with established staff, are a well-known stress source during training or practical stages, and it is not unusual to see how certain professionals provide inadequate feed-back or act in a way that reduces the residents self-confidence, instead of boosting it (Torrado, 1995). In turn, it is more likely that psychiatrists experience stress symptoms due to serious complications in the attended patients (e.g., high suicide risk, frequent relapses, poor response to treatment or serious cognitive deterioration as a consequence of the disease).
It is interesting to note that, among experienced psychologists, the greatest source of physical and psychological discomfort is associated with the Job criticism subscale, a type of stress configured by items describing feelings of occupational frustration, reminiscent of the typical burnout syndrome. However, unexpectedly, activation evoked by therapeutic decisions -"eustress" according to Selyes (1956, 1974) terminology - somehow contributes to counteracting the negative effect of noxious stress sources summarised in the Job criticism items. How could this finding be interpreted?
Firstly, we must remember that the term "eustress" is used by some authors to underline the stress phenomenon as an adaptive experience, and makes reference to "situations and experiences in which stress has predominantly positive results and consequences, since it produces the necessary stimulation and activation to allow people to achieve satisfactory outcomes in their activities with manageable personal costs" (Peiró, 1993, p. 10). Secondly, it is interesting to discuss our results on the Therapeutic decisions scale in the light of the literature on coping strategies and styles. Certainly, any health professional actively concerned with generating alternatives for alleviating patients suffering would show a positive attitude focused on solutions, even though such a task is subjectively quite demanding. Several studies with different populations and stressors have detected a positive association between psychological well-being and coping strategies oriented to control versus those exclusively oriented towards avoidance (Mullen and Suis, 1982; Folkman and Lazarus, 1985; Endler and Parker, 1990; Koeske et al., 1993). However, we cannot confirm the existence of a coping style which is, a priori, more adaptive in every situation (partly because of specific contextual factors). Recently, significant associations have been found between emotional exhaustion and different combinations of coping styles in mental health workers (Leiter, 1991).
When considering that the emotional exhaustion phenomenon may be optimally explained using a stage model (Golembiewski and Munzenrider, 1988), we must bear in mind that it is very likely that tension generated by therapeutic decisions may become harmful if that experience is excessive, uncontrolled or uncontrollable (Greenglas, 1991).
Based on the results obtained by Álvarez and Fernández-Ríos (1991), it appears that the stress symptoms experienced by psychiatrists, in contrast to clinical psychologists, would be more related to immediate work context variables. For example, the burnout level among clinical psychologists that could be explained by work atmosphere factors (pressure, autonomy, innovation...) was considerably lower than for other professional groups, such as psychiatrists. We then noted the potential existence of personal and formative mediating variables. In the same line, we have now found that, for psychiatrists, workplace defines the severity of the stress symptoms experienced better than profession, at least in some of the more common settings (such as Mental Health Units).
As regards the socio-demographic and occupational variables analysed in this study (sex, age, years of experience), none of them had a determining effect on the global symptoms level. Traditionally, the sex variable has not given unequivocal results on individual differences with respect to stress. Greenglas (1991) careful review considers that, although women may more frequently experience multiple sources of stress, they are consistently seen to be less affected by them. Despite biological differences in susceptibility to stress, it is obvious that there are different roles exposing each gender to different stressors and different coping strategies (Peiró and Salvador, 1993).
With respect to the influence of years of experience on stress, there are no conclusive data in the literature. Nor is there evidence that time, by itself, guarantees the acquisition of new coping strategies, nor the appropriateness of acquired responses. Some authors have suggested a hypothetical critical period, between the first 2 and 5 years of professional experience, during which emotional exhaustion could affect personal and professional competence, as well as decisions related to the job itself (Deckard, Meterko and Field, 1994).
Finally, it is interesting to note that occupational stress experiences cannot always be coped with in an efficient way if only the individual level is taken into account (Ganster, Mayes, Sime and Tharp, 1982; Shinn et al., 1984). It seems that in the occupational domain individual coping responses may be less useful than group and organisational strategies. As other authors affirm, the health service organisations maintain the potential to intervene and take action to reduce the occurrence of occupational stress, and so to reduce the negative consequences for individuals, organisations and the quality of attention provided (Golembiewski and Munzenrider, 1988).
Alvarez, E. and Fernández-Ríos, L. (1991). El síndrome de burnout o el desgaste profesional (II): Estudio empírico de los profesionales gallegos del área de Salud Mental. Revista de la Asociación Española de Neuropsiquiatría, 11, 267-273.
Aro, S. (198l). Stress, morbidity and health-related behavior. A five year follow-up study among meta industry employees. Scandinavian Journal of Social Medicine (Suppl 25), I - 130.
Bjorksten, O., Sutheriand, S., Miller, C. and Stewart, T. (1983). Identification of medical student problems and comparison with those of other students. Journal of Medical Education, 58, 759767.
Bomon, B. (1988). Stress and heart disease. En S. Fisher and J. Reason (Eds.), Handbook of Life Stress, Cognition and Health (301-316), Chichester: John Wiley & Sons.
Colexio Oficial de Psicólogos-Galicia (1992). Relación de recursos de Salud Mental para la aplicación del Programa PIR. Technical Internal Document.
Cooper, C.L., Watts, J., Balioni, A.J. and Kelly, M. (1988). Occupational stress among general practice dentists. Journal of Occupational Psychology, 61, 163-174.
Cronkite, R.C. and Moos. R.H. (1984). The role of pre-disposing and moderating factors in the stress-illness relationship. Journal of Health and Social Behavior. 25, 372-393.
Deckard, G., Meterko, M. and Field, D. (1994). Physician Burnout: An examination of personal, professional and organizational relationships. Medical Care, 32, 745-754.
Edwards, J.R. and Cooper, C.L. (1988). The impacts of positive psychological states on physical health: a review and theoretical framework. Social Science and Medicine, 27, 1447 1459
Endier, N.S. and Parker, J.D. (1990). Multidimensional assessment of coping: A critical evaluation. Journal of Personality and Social Psychology, 58, 844-854.
Folkman, S. and Lazarus, R.S. (1985). If it changes it must be a process: Study of emotion and coping during three stages of a college examination. Journal of Personality and Social Psychology, 48, 150-170.
Freudenberger, H.J. (1974). Staff burn-out Journal of Social Issues, 30, 159-165.
Ganster, D.C., Mayes, B.T., Sime. W.E. and Tharp, G.D., (1982). Managing occupational stress: A field experiment. Journal of Applied Psychology, 67, 533-542.
Golembiewski, R.T. and Munzenrider, R.F., (1988). Phases of Burnout. Developments in concepts and applications. New York: Praeger.
Greenglas, E.R., (1991). Burnout and gender: Theoretical and organizational implications. Canadian Psychology, 32,562-572.
Herbert, T.B. and Cohen. S. (1993). Stress and Immunity in humans: A meta-analytic review. Psychosomatic Medicine, 55, 364-79.
Koeske, G.F., Kirk, S.A. and Koeske, R.D. (1993). Coping with job stress which strategies work best? Journal of Occupational and Organizational Psychology, 66, 319-335.
Lazarus, R.S. and Folkman, S. (1984). Stress, appraisal and coping. New York: Springer Publishing Company.
Leiter, M.P. (199l). Coping patterns as predictors of burnout: The function of control and escapist coping. Journal of Occupational Behavior, 12,123-144.
Moreno, B.; Oliver, C. and Aragoneses, A. (1991). El burnout, una forma específica de estrés laboral. In G. Buela-Casal y V. Caballo (Eds). Manual de Psicología Clínica Aplicada (271-284). Madrid: Siglo XXI.
Mullen, B. and Suis, J. (1982). The effectiveness of attention and rejection as coping styles: A meta-analysis of temporal differences. Journal of Psychosomatic Research. 26,43-49.
Olkinuora, M., Asp, S., Juntunhen, J., Kauttu, K., Strid, L. and Äärimaa, M.(1990). Stress symptoms, burnout and suicidal thoughts in Finnish physicians. Social Psychiatry and Psychiatric Epidemiology, 25,81-86.
Payne, T. and Firth-Cozens, J. (1987) (Eds). Stress in health professionals. New York: John Wiley and Sons.
Peiró, J.M. (1993). Desencadenantes del estrés laboral. Madrid: Eudema.
Peiró, J.M. and Salvador, A. (1993). Control del estrés laboral. Madrid. Eudema.
Pines, A. and Maslach, C. (1978). Characteristics of staff burnout in Mental Health settings. Hospital and Community Psychiatry, 29, 233-237.
Reig, A. and Caruana, A. (1990). Estrés ocupacional y patrón A de comportamiento como predictores de insatisfacción laboral y del estado de salud en personas de enfermería de hospital. Psicología del Trabajo y de las Organizaciones, 6, 146-162.
Selye, H. (1956). The Stress of Life. New York: McGraw Hill.
Selye. H. (1974). Stress without distress. London: Hodder & Stoughton.
Shinn, M., Rosario, M., Morch, H. and Chestnut, D.E. (1984). Coping with job stress and burnout in the human services. Journal of Personality and Social Psychology, 46, 846-876.
Skalar, L.S. and Anisman, H. (1981). Stress and cancer. Psychological Bulletin, 89,369-406.
Torrado, V. M. (1995). Fuentes de estrés en los profesionales de la salud. Siso/Saúde. 26, 25-29.
Ullrich, A. and Fitzgerald, P. (1990). Stress experienced by physicians and nurses in the cancer ward. Social Science and Medicine, 31, 1013-1022.
Xunta de Galicia (1989). Guía de recursos asistenciales en Saúde Mental da Comunidade Autónoma de Galicia. Colección Saúde Mental. Serie Reforma Psiquiátrica-Guías, nº 3. Santiago de Compostela: Autor.
1 Given the variety of administration and functional social services centres at the time of this study, a professionals name listing was made by looking at different sources. In general terms, assistance resources offered (COP-Galicia, 1992; Xunta de Galicia, 1989), the report on experienced professionals in every health area and centres staff listing have been consulted.