ORIGINAL PAPERS

Sources of stress, psychologicaldistress and burnout inpsychiatristsComparison of junior doctors, senior registrars andconsultants

Elspeth Guthrie, Teresa Tattan, Edwina Williams, Dawn Black andHimant Bacliocotti

Aims To assessthe degree of psychological morbidityand burnout in 138 psychiatrists in three Manchesterteaching hospitals.Results The results for senior house officers (SHOs),registrars, senior registrars and consultants werecompared. The overall response rate was 76.8%. Therewas no significant difference in psychological morbiditybetween the three training grades, but SHOs andregistrars reported significantly higher levels of burnoutthan either senior registrars or consultants. Dealing withviolent patients was stressful for all psychiatrists, nomatter what the grade.Clinical implications Factors related to job stress inpsychiatry need to be addressed. In particular, theprovision of safer working environments needs to beconsidered for psychiatrists at all levels of training.

Recent studies have reported high levels ofpsychological distress in both junior hospitaldoctors (Firth-Cozens, 1987), and consultants(Ramirez et al, 1996; Deary et al 1996a). Out ofall the medical specialities, psychiatrists havethe most intense form of interpersonal contactwith patients, dealing on a regular basis withextremely distressed or disturbed individuals.Although, consultant psychiatrists work fewerhours on average than consultant physicians orsurgeons, they report more emotional exhaustion and more severe depression as a result ofworkload (Deary et al, 1996b) than eitherphysicians or surgeons. This difference may bedue to the particular nature of the work thatconsultant psychiatrists undertake, or due toparticular personality characteristics of psychiatrists which make them vulnerable to job stress.Psychiatrists have higher neuroticism scoresthan physicians or surgeons (Deary et aL1996b), and medical students expressing apreference for psychiatry have a higher score on

'complexity traits' (Walton, 1969) and a moreabstract approach to problem-solving (Neyet al,1990) than other students.

Worryingly, psychiatrists have higher rates ofsuicide than other specialities, although thestudies in this area are rather old (Rich & Pitts,1980).

There is little work which has directly compared the prevalence of psychological morbidityin psychiatrists at different stages of theirtraining, or studied the different kinds ofStressors faced by psychiatrists as they progressup the career ladder.

The aim of this study was to compare: thedegree of psychological distress, the degree ofburnout and the kind of Stressors reported bypsychiatrists at different stages of training inthree hospitals in the Manchester area.

The studyAll senior house officers, registrars, seniorregistrars and consultants at three large hospitals in Manchester were asked to complete aconfidential self-report questionnaire. The following assessments were employed.

General Health Questionnaire (GHQ-12;Goldberg & Wiííiams.1988)This is a widely used self-report measure whichwas designed as a screening device for estimating minor psychiatric disorder in the generalpopulation. It has been found useful for assessing stress in occupational settings and as anestimate of psychiatric caseness. The usualthreshold score is either 1/2 or 2/3. In thepresent study, the GHQ method of scoring was

Psychiatric Bulletin (1999), 23. 207-212 207

ORIGINAL PAPERS

used and a high threshold score of 3/4 wastaken as an indication of probable caseness.

Maslach Burnout Inventory (Maslach &Jackson. 1981)This 22-item self-report questionnaire is a wellrecognised and widely used measure of burnoutin relation to occupational stress. It has threesub-scales: personal accomplishment (feelingsofcompetence and achievement) which is measured by eight items; depersonalisation (feelingdetached and uninvolved with patients) which ismeasured by nine items; and emotional energy(feeling emotionally drained by work) measuredby five items. Respondents are asked to rate eachitem according to frequency on a seven-pointscale from 'never' to 'every day'. The total scorefor each sub-scale is categorised 'low', 'average',or 'high' according to predetermined cut-offscores based on normative data from a sampleof American health professionals. A high degreeof burnout is indicated by high scores on theemotional energy and depersonalisation sub-scales and low scores on the personal accomplishment sub-scale.

Stress Incident Record (Keenan & Newton,1985)This measure examines acute Stressors and hasbeen used in previous studies of medicalstudents and hospital doctors (Firth-Cozens,1987). It requires respondents to "describebriefly a real event which has occurred in thepast month and which has been stressful to you(Note:a stressful event is one which has arousedyour feelings in some negative way)". Therespondents are then asked to describe feelingsinvoked by the incident by rating a list of 15feelings on a five-point scale from 'not at all' to'extremely'. The incidents described by the

psychiatrists were subjected to a content analysis by grouping together similar descriptions.This was carried out by two independent raters.(F. W. and T. T.).

Additional ratings of StressorsThe psychiatrists were also asked to rate 30work-related items on a four-point scale (0=notstressful to 3=extremely stressful). The scale wasadapted from a scale used by Firth-Cozens tomeasure stress in junior doctors. Additionalitems, specifically relevant to psychiatry wereadded. These included: "dealing with violentpatients", "dealing with angry patients", "dealingwith sexual problems". The scores for thedifferent grades of psychiatrist were comparedfor each item. In addition, the mean score fromthe scale was used as an indicator of overallstress.

The data were analysed using the StatisticalPackage for the Social Sciences (SPSS). Non-parametric statistics were employed. Data fromthe three groups were compared using theKruskal-Wallis one-way analysis of variance.The proportion of psychiatrists who rated different items on the 'work-related scale' as moderately or very stressful were compared for casesand non-cases on the GHQ-12 using the x2test.Differences in gender were compared using theMann-Whitney L/-test.

FindingsOne hundred and thirty-eight doctors were sentquestionnaires of whom 106 (76.8%) responded(51.1% were male, n=54). The response rates foreach grade were as follows:SHOs/registrars 31 /46 (67.4%). senior registrars 33/42 (78.6%) andconsultants 42/50 (84%). The gender distribution for the different grades was as follows:consultants (males) 59.6%, senior registrars

Table 1. GHQ scores and sub-scales of the Maslach Burnout Inventory

GHQ (high scores (>4))(%)Burnout

(%)High emotional

exhaustion scoresLow personal

accomplishment scoreHighdepersonalisationStress

(%)Total stressscore11

. Kruskal-Wallis test, medianSHO/

registrars(n=31)11

(36.6)17(54.8)

14 (45.2)

15(48.4)41

(29-50)Senior

registrar(n=33)6(18.2)1

1 (33.3)

15(45.4)

11(33.3)30(21-42)Consultant

(n=42)10(23.8)1

1 (26.2)

14 (33.3)

6(14.3)28

(19-34)X23.18.9

1.5

11.510.58d.f.22

2

2PP=0.21P=0.01

P=0.45

P<0.01P<0.01(interquartile

range), Kruskal-Wallis x2-

208 Guthrie et al

ORIGINAL PAPERS

(males) 60.7%, registrars/SHOs (males) 35.5%.The median GHQ score for the whole sample was1.0 (range 0-4), 27 (31.4%) doctors reportedscoring above threshold. There was no significant difference in GHQ scores by gender. Themedian and interquartile range scores for malesand females were as follows:males 1.0 (0.0-4.0).females 1.0 (0.0-4.2) (Z-0.23, P=0.81). Therewas no significant difference in GHQ-12 scoresacross the different grades (Table 1), althoughSHOs/registrars had higher scores than theother two groups.

On the Maslach Burnout Inventory 42 doctors(39.6%) out of the whole sample were highscorers on the emotional exhaustion sub-scale,33 (31.1%) were high scorers on the depersonal -isation sub-scale and 42 (39.6%) were lowscorers on the personal accomplishment sub-scale. There was, however, a significant difference between the three psychiatric grades on twoof the sub-scales of the Maslach BurnoutInventory (Table 1). A far greater proportion ofSHOs and registrars reported feelings of severeemotional exhaustion and depersonalisationthan either senior registrars or consultants. Allthree grades reported similar levels of satisfaction and personal accomplishment from theirwork.

Eighty-three (78.3%) of the sample reported astressful incident. The main kinds of acuteincident, according to grade, are shown in Table2. Examples of each kind of Stressor are alsoincluded in the table. The four most frequentlycited Stressors, by the group as a whole, were:

personal problems, problems involving patients(such as suicide or violence), career threat (suchas a formal complaint) and administrative problems (such as lack of beds). There was.however, a marked difference between the kindsof Stressors that were most frequently citedaccording to grade.

Administrative difficulties and career threatswere the most frequently cited Stressors by consultants. The emphasis upon community careand the closure of large numbers of psychiatricbeds without sufficient community resourceswas frequently commented upon. Consultantsdescribed spending hours trying to arrange bedsfor patients in hospitals over 100 miles awayfrom Manchester. They also described the stressof carrying medical responsibility for suicidaland homicidal patients without any power toinfluence service delivery.

For SHOs and registrars, dealing with patientsand managing personal problems, were the mostcommon Stressors. In particular, junior psychiatrists commented upon the difficulties posed bypersonal life events influencing the ability to dealwith patients' emotional problems. Examples ofthis included the deaths of relatives and in onecase the suicide of a psychiatrist's closerelative.

Registrars scored significantly higher on theoverall stress score from the 30-item stressquestionnaire (Table 1).Table 3 shows the itemswhere there was a difference in scoring acrossthe grades and it also shows other items whichwere rated highly by all three grades. The

Table 2. Types of Stressor identified by the stress incident record

Type of StressorSHO/registrar, n=30 Senior registrar, n=33 Consultants, n=36n (%) n (%) n (%)

Careerthreat(e.g.job interview,complaint)Administration(no

beds)Clinical41113.33.33.344312.112.19.169216.625.05.5

(on call, use of Mental Health Act.acting up)

Academic 0 0.0(lack of research, obtaining funding)

Patients 7 23.3(violence, threats, homicide,

suicide)Trainee/trainer conflict 1 3.3Peers 1 3.3

(covering for illness,deathof colleague)

Personal 10 33.3(loss/illness,fertility issues,

relationship problems, personalillness,family conflict,moving house)

3.0

12.1

3.00.0

15.2

8.3

13.8

2.713.8

13.8

Stress in psychiatrists 209

Table 3. Ratings of job-relatedmoderately or verystressfulStressorWard

roundsOn-callUnexpectedcoverViolent

patientsAngrypatientsPatients'sexualproblemsUnable

tohelppatientsDealing

withissuestodowith

dyingPatients'relativesTime

offforsicknessAcademic

workRelationshipswithjunior

doctorsRelationswithnursesAdministrationRelationshipswith

consultantsstress:

number and percentage of doctors rating work itemasSHO/registrar,

n=31 Senior registrar, n=33 Consultants, n=42n (%) n (%) n(%)1820232425231361172132227(62)(68.9)(79.3)(82.7)(86.6)(79.3)(44.8)(20.7)(37.9)(24.1)(72.4)(10.3)(6.9)(75.8)(24.1)158142522241631161913205(45.5)(24.2)(42.4)(75.8)(66.6)(72.7)(48.5)(9.1)(33.3)(18.2)(57.6)(3)(9.1)(60.6)(15.2)555252461791317142112117(13.8)(13.8)(13.8)(69.4)(66.6)(16.6)(47.2)(25)(36.1)(47.2)(38.8)(5.5)(30.5)(58.3)(47.2)X216.6323.7928.071.543.9132.30.083.050.147.647.41.468.482.449.1d.f.222222222222222PP<0.001P<

0.00001P<0.00001P<0.46P<142P<0.00001P=0.96P=0.22P=0.92P<0.05P<0.05P=0.47P<0.05P=0.29P<0.05

number and percentage of doctors rating eachItem as being 'moderately stressful' or 'verystressful' is compared.

Dealing with violent patients was rated asbeing stressful by more psychiatrists, no matterwhat grade, than any other item on the stressquestionnaire: with 82.7% of SHOs/registrars.75.8% of senior registrars and 69.4% of consultants rating this item as being moderately orseverely stressful. There were marked differencesbetween the grades, however, in the ratings ofother items on the stress questionnaire. Wardrounds, on-call, unexpected cover and academicwork were all rated as being more stressful bySHOs and registrars compared with consultants,with scores for senior registrars falling in-between. Consultants rated time off work forsickness, relationships with nurses and relationships with other consultants as being significantly more stressful than either SHOs/registrars or senior registrars. Interestingly,senior registrars were never the highest scorerson any item of the stress questionnaires, andwere either intermediate scorers between SHOs/registrars and consultants or the lowest scorers.

In an attempt to ascertain which work-relatedfactors might be most closely associated withpsychological morbidity in psychiatrists, thegroup was divided into 'cases' (those scoringabove threshold on the GHQ-12) (n=27) and

'non-cases' those scoring below threshold (n=79).There was no difference between cases and non-cases on the total scores from the stressquestionnaire or on any individual item withexception of covering for colleagues (cases 66% v.non-cases 36% x2 P=0.01) or speaking withrelatives (cases 60% v. non-cases 30%, P=0.02).

CommentThe results from this study must be treated withcaution due to the small numbers and thereliance upon a self-reort measure. The responserate was reasonable for each grade but lowest forSHOs/registrars. There was no way of determining whether the responders were different fromthe non-responders because of the importance ofmaintaining anonymity.

The overall prevalence of psychological distress in psychiatrists was high, with approximately one-quarter scoring above threshold onthe GHQ. This figure, however, is very similar toprevalence rates found in other studies of stressand psychological morbidity in doctors in general(Firth-Cozens, 1987: Ramirez et al, 1996). Unlikeother studies, we did not find a difference ingender. Female psychiatrists were no more likelyto suffer from psychological distress than theirmale counterparts.

210 Guthrie et al

ORIGINAL PAPERS

It is clear that junior psychiatrists reportedmuch higher levels of psychological distress andjob-related stress than consultants, and morejuniors appeared to experience symptoms ofburnout than consultants. This difference couldnot be explained by the higher proportion ofwomen in the SHO/registrar grade comparedwith consultant grade, as there was no differencein GHQ scores between males and females. Thescores for consultant psychiatrists on the GHQand Maslach Burnout Inventory are very similarto those obtained by Ramirez et al (1996) forconsultants in medicine and surgery, and appearto be representative of doctors working at asenior level.

It is worrying that junior psychiatrists reportedmore symptoms of burnout than their seniorcolleagues and in particular felt less able to feelcompassionately towards patients than seniorcolleagues (depersonalisation sub-scale).

It is not possible from the design of this studyto identify the causes of such high stress inyoung psychiatrists, but acute work relatedstress appeared to be more closely associatedwith psychological morbidity in SHOs and registrars than the other grades.

The results of this small study suggest thatjunior psychiatrists face particular Stressors inrelation to work that are different to those facedby more senior colleagues. Acute on-call work inpsychiatry has become more stressful in recentyears, particularly in the inner-city areas.Although these Stressors affect psychiatrists atall grades juniors are in the front line and facethe greatest pressures. It is also possible thatconsultants report less stress than juniors, as bydefinition they represent a group of self-selectedhigh achievers, who have chosen psychiatry as acareer and have coped with psychiatric trainingwithout dropping out.

One of the most striking findings of this studywas that many of the spontaneously generated'stressful factors' on the Stress Incident Recordwere personal and not related to work. Studieson 'stress-in-the-workplace' often overlook theimportance of non-work related Stressors. Inpsychiatry, personal life events (e.g. bereavement) may independently contribute to increasedstress, or may make it more difficult forpsychiatrists to function at work, particularly ifthey have to help others who have sufferedsimilar events to themselves. The subtle waysin which job-related factors and personal factorsimpinge upon one another are difficult to study,but are important to consider.

Factors related to job stress in psychiatristsneed to be addressed. This must include bettersupervision for junior psychiatrists, with recognition from senior doctors of how personaldifficulties in a young psychiatrist's life mayimpinge upon work issues. Pressures upon

consultant psychiatrists are unlikely to changeunless the providers of mental health services,and in particular consultants who take medicalresponsibility for the most seriously ill patients,have some influence over the service provisionand the number of beds required to safely treatpatients. Ramirez et al (1996) have found thatautonomy and variety of work are importantfactors which contribute to consultant jobsatisfaction, and protect consultants from mental ill health. Further erosion of consultantautonomy in psychiatry is likely to lead to greaterjob stress.

Dealing with violent patients was the moststressful work-related item for psychiatrists of allgrades. This highlights the importance of training in relation to issues of safety but also theimportance of the provision of safer workingenvironments for psychiatrists by trusts whichemploy them.

Although in comparison to other medicalspecialities, the overall workload of psychiatristsmaybe less onerous, the particular nature of thework is very different, and it is sometimespersonally threatening. Particular job-relatedcharacteristics need to be considered in futurework which attempts to compare doctors working in different specialities.

ReferencesDEARY.I. H.. BLENKIN.H.. Aoius. R. M., et al (1996a) Models

of job-related stress and personal achievement amongconsultant doctors. British Journal of Psychology, 87.3-29.

—¿�,AGIUS,R. M. & SADLER,A. (1996b) Personality and stressIn consultant psychiatrists. International Journal ofSocial Psychiatry. 42. 112-113.

FIRTH-COZENS.J. (1987) Emotional distress in junior houseofficers. British Medical Journal. 295. 533-536.

GOLDBERG.D. & WILLIAMS.P. (1988) A User's Guide to theGeneral Health Questionnaire. Windsor: NFER-NelsonPublishing Co.

KEENAN.A. & NEWTON. T. J. (1985) Stressful events.Stressors and psychological strains in youngprofessional engineers. Journal of OccupationalBehaviour. 6. 151-156.

MASLACH.C. & JACKSON.S. E. (1981) The Maslach BurnoutInventory. Palo Alto. CA: Consulting PsychologistsPress.

NEY. P.O., TAM,W. W. K. & MAURICE.W. L. (1990) Factorswhich determine medical student interest in psychiatry.Auslralian and New Zealand Journal of Psychiatry, 24.65-76.

RAMIREZ.A.. GRAHAM.J.. RICHARDS.M. A., et al (1996)Mental health of hospital consultants: the effects ofstress and satisfaction at work. Lancet. 347, 724-28.

RICH. C. L.. & PITTS,F. N. (1980) Suicide by psychiatrists: astudy of medical specialists among 18 730 consecutivephysician deaths during a five-year period 1967-72.Journai of Clinical Psychiatry. 41. 261-263.

WALTON.H. J.. (1969) Personality correlates of a careerinterest in psychiatry. British Journal of Psychiatry.115. 211-219.

Stress in psychiatrists 211

ORIGINAL PAPERS

*Elspeth Guthrie, Senior Lecturer in Psychiatry,University of Manchester, Rawnsley Building,CMHT, Manchester M13 9BX; Teresa Tattan,Research Fellow, University of Manchester,Edwina Williams, Specialist Registrar.Westminster and Chelsea Hospitals, London:

Dawn Black, Consultant Psychiatrist, HopeHospital. Salford: and Himant Bacliocotti,Specialist Registrar, Nottingham

•¿�Correspondence

First psychosis liaison unit familyeducation programmeEric Morris, Catie Morris and Lynley Nolon

Aims and methods Thisstudy presents an evaluation ofan education programme for families whose relativehas experienced a first episode of psychosis.Participants attended a five-week programme and

were asked to rate their level of knowledge andconfidence in managing the Illness before and afterthe education sessions.Results Two consecutive groups of relatives (n=9;n=ll) attended the programme. Comparison ofknowledge and confidence ratings before and afterthe education sessions demonstrated significant levelsof change at the final session for both groups.Clinical implications Education for relatives of peopleexperiencing psychosis for the first time is an Importantcomponent In community management. Such education improves the knowledge and confidence ofrelatives in coping with the illness, possibly leading toa more accepting and less stressful family environment,reducing the risk of relapse and promoting thepsychological health of the patient.

Assistance in the family management of psychosis can produce significant reductions in relapserate and produce benefits for the generalpsychological well-being of the individual with apsychotic illness (Falloon et ai, 1982). A reduction in subsequent relapses after a first episodeof illness may help prevent chronicity andassociated disability (Jackson & Birchwood,1996).

An important component of the family management process is education of the familyregarding the illness, the role of various treatment providers, problem-solving and managingcrises, and the variety of services available to thefamily and the individual (Tamer & Barrow-clough et al. 1987).

We describe a family education programmewhich is part of the first psychosis service (FirstPsychosis Liaison Unit; FPLU) based at the MillsStreet Centre, Perth, Western Australia. TheMills Street Centre services a socio-economicallydiverse catchment area of around 120 000adults. The FPLU, a community-based initiative,became operational in late August 1995 andaims to facilitate the provision of personalisedcomprehensive care for young people (aged 16 to35 years) experiencing a first episode of psychosis. This is achieved by fostering effective linksbetween relevant health, welfare, employmentand education-related agencies. FPLU providescase management for individuals in the firstepisode of a psychotic illness and operates as anadjunct to the existing multi-disciplinary mentalhealth service. Case management is conductedby clinical liaison officers, who are trainedmental health professionals from either socialwork, community mental health nursing oroccupational therapy backgrounds working ingeneric positions. Other components of the FPLUservice include individual and family support,community education, promotion of early identification of psychosis and clinical research.

The FPLU Family Education Programme aimsto provide family members with accurate information and to encourage the use of thisinformation to enhance familial adjustment tohaving a member with a psychotic illness. It isexpected that this will enable more appropriateuse of services, better communication betweenfamily members and treatment providers, andimprove the coping skills of the family. Inaddition the programme allows the participantsto meet other families who are facing the same

212 Psychiatric Bulletin (1999), 23, 212-214