Bulletin on the effectiveness of health service interventions for decision makers NHS Centre for Reviews and Dissemination, University of York DECEMBER 1998 VOLUME 4 NUMBER 6 ISSN: 0965-0288 Effective Health Care Deliberate self-harm The contents of this bulletin are likely to be valid for around one year, by which time significant new research evidence may have become available. ■ Deliberate self-harm involves intentional self-poisoning or injury, irrespective of the apparent purpose of the act. It is one of the top five causes of acute medical admissions for both women and men in the UK. ■ All hospital attendance following deliberate self- harm should lead to a specialist psychosocial assessment. This should identify motives for the act, and those associated problems which are potentially amenable to intervention such as psychological or social problems, mental disorder, alcohol and substance misuse. ■ Direct discharge from A&E should only be contemplated if a psychosocial assessment and aftercare plan can be arranged prior to discharge. Aftercare arrangements should include advice on the services available. ■ GPs should have ready access to training and advice about the assessment and management of self-harm patients in primary care. ■ Accessible and comprehensive services need to include a mechanism for engaging people who do not attend routine clinic appointments. Access to follow-up needs to be rapid as repetition can occur soon after the episode. ■ Service providers should work to improve attitudes towards self-harming patients, for example through training aimed at increasing knowledge. ■ There is insufficient evidence to recommend a specific clinical intervention after deliberate self-harm. Further research is needed to establish the effectiveness of potential interventions.
Transcript
Bulletin on the effectiveness of health serviceinterventions fordecision makers
NHS Centre for Reviewsand Dissemination, University of York
DECEMBER 1998 VOLUME 4 NUMBER 6 ISSN: 0965-0288
EffectiveHealth Care
Deliberate self-harm
The contents of this bulletin are likely to be valid for around one year, by which time significant new research evidence may have become available.
■ Deliberate self-harm involvesintentional self-poisoningor injury, irrespective of theapparent purpose of theact. It is one of the top fivecauses of acute medicaladmissions for both womenand men in the UK.
■ All hospital attendancefollowing deliberate self-harm should lead to aspecialist psychosocialassessment. This shouldidentify motives for theact, and those associatedproblems which arepotentially amenable tointervention such aspsychological or socialproblems, mental disorder,alcohol and substancemisuse.
■ Direct discharge from A&Eshould only becontemplated if apsychosocial assessmentand aftercare plan can bearranged prior to discharge.Aftercare arrangementsshould include advice onthe services available.
■ GPs should have readyaccess to training andadvice about theassessment andmanagement of self-harmpatients in primary care.
■ Accessible andcomprehensive servicesneed to include amechanism for engagingpeople who do not attendroutine clinic appointments.Access to follow-up needsto be rapid as repetitioncan occur soon after theepisode.
■ Service providers shouldwork to improve attitudestowards self-harmingpatients, for examplethrough training aimed atincreasing knowledge.
■ There is insufficientevidence to recommend aspecific clinical interventionafter deliberate self-harm.Further research is neededto establish theeffectiveness of potentialinterventions.
A. BackgroundDeliberate self-harm involvesintentional self-poisoning or selfinjury, irrespective of the apparentpurpose of the act.1 Self-poisoning,for example an overdose, is the mostcommon form followed by cutting.
Deliberate self-harm is one of thetop five causes of acute medicaladmissions for both women andmen.2 Most cases of deliberateself-poisoning present to generalhospitals; in the UK there are about150,000 such attendances annually.The most common substancesingested are analgesics, particularlyparacetamol and paracetamolcontaining compounds.3
Over the last 50 years, there hasbeen a rise in incidence of self-harm, with a marked increase fromthe early 1960s. Rates levelled offin the late 1970s, there was amodest decline until the mid1980s, but since then rates haverisen continuously (see Fig.1).4,176 Itis difficult to get an accuratepicture of the epidemiology ofdeliberate self-harm; Oxford is theonly UK centre with a continuousmonitoring system. Nonetheless, areasonable estimate is that currentrates are around 400 per 100,000population per annum – similar tothe highest rates of the late 1970s.3
This incidence is higher than mostothers recorded in Europe.5
Of known risk factors for completedsuicide, deliberate self-harm has thestrongest association. In the yearafter an episode of deliberate self-harm, the suicide rate is 100 timesthat of the general population.33
About a quarter of all suicidesattend a general hospital after anon-fatal act of self-harm in the 12months before they die.6,7 Effectiveintervention after deliberate self-harm, if it were available, couldtherefore be an important meansof achieving the targets forreduction of the suicide rate whichare outlined in the Health of theNation 8 and in the recent GreenPaper Our Healthier Nation.9
Although there were oncebetween 2 or 3 times as many
episodes in females, the sex-specific rates have steadily drawncloser together, so that self-harm isnow only slightly more commonamong women than men.2,3,10
Some general hospitals now dealwith more referrals of men thanwomen.11 This trend is worthnoting, because it is among youngmen that the suicide rate has beenincreasing in the last ten years.The mean age of the self-harmpopulation is in the early 30s forboth sexes, the peak age forpresentation being 15–24 years forwomen and 25–34 years for men.12,13
In most cases, people report thatthey have taken an overdose inresponse to social problems.14
Common problems includedifficulties with housing,unemployment, debt, poor personalhealth, and conflict or loss inpersonal relationships.15 There issome evidence that repetition ofself-harm may occur despiteresolution of personal problems.16,17
Following an episode of deliberateself-harm, about 30–40% of generalhospital attenders are given apsychiatric diagnosis, and about athird have had prior contact withthe psychiatric services.18 The mostcommon diagnosis is some form ofdepressive disorder.19 Alcohol dep-endence is diagnosed in about 10%of cases.20–22 Mental illnesses suchas schizophrenia and bipolar disorderare diagnosed in less than 10% ofepisodes of deliberate self-harm.23
Enduring psychologicalcharacteristics associated with self-harm include: hopelessness, whichas a character trait may occur
independently of depression,hostility to others, antisocialbehaviour and deficient problem-solving abilities.24–26 Thesepsychological characteristics maybe associated with self-harmbecause they confer vulnerabilityto mental disorder or socialproblems, or they may increaserisk of self-harm in their own right.
A number of features whichpredict repetition or eventualsuicide can be identified after anepisode of self-harm, the bestestablished are listed in Table 1.27–37
Risk of repetition is not uniformlydistributed, and some peoplerepeat self-harm on numerousoccasions. In one study, 15% ofpeople admitted to a poisons unithad taken at least 5 overdoses.38
Although it is often assumed that
2 EFFECTIVE HEALTH CARE Deliberate self-harm DECEMBER 1998
Fig. 1 Rates of deliberate self-harm 1962 to 1996 4,176
Pairs of years. Note that 1962/63 to 1973/74 are rates from Edinburgh,while 1977/78 onwards are rates from Oxford City
For non-fatal repetitiona history of self-harm prior to the currentepisode;psychiatric history, especially as an inpatient;current unemployment;lower social class;alcohol or drug-related problems;criminal record;antisocial personality;uncooperativeness with general hospitaltreatment;hopelessness;high suicidal intent.
For suicideolder age;male;previous attempts;psychiatric history;unemployment;poor physical health;living alone.
Table 1 Features which predict non-fatal repetitionof deliberate self-harm or eventual suicide
those who repeat self-harmfrequently are predominantlywomen, the excess of womenamong chronic repeaters is probablyno greater than among the self-harming population as a whole.38
Little is known about multiplerepeaters, except for a sub-groupof women who meet criteria forborderline personality disorder, manyof whom have been subject to abuse(not always sexual) in childhood.39
B. Nature ofevidenceTwo systematic reviews evaluatingthe effectiveness of interventionsafter deliberate self-harm haverecently been published.40,41 Thereview by Hawton et al. has beenused in this bulletin to evaluatethe effectiveness of suchinterventions, and has beenupdated to include two additionaltrials published after the reviewwas completed.42,43 The review is aversion of the Cochrane reviewavailable on the Cochrane Library(Update Software, Oxford). Areview of the research evidence onthe characteristics of an effectiveclinical service for the assessmentand aftercare of people who present
following an episode of deliberateself-harm was also undertaken. Asummary of the methodology isincluded in the appendix.
The literature on deliberate self-harm is limited in two ways. First,the data come largely from studieson general hospital attenders,although up to a third of episodesmay not lead to medical contact.44
Second, most research has beenconducted on deliberate self-poisoning rather than other formsof self-harm such as cutting. Thereis some overlap between thesebehaviours, but caution should betaken about generalising.
C. Outcomefollowingdeliberateself-harmC.1 Suicide subsequent todeliberate self-harmTwenty four studies reportedsuicide rates for the first year afteran episode of self-harm.17,31,34,45–65
These ranged from 0% to 6%(median 1%, interquartile range0% to 1.8%). The quality of thesestudies was variable. The mostrecent rigorous UK study33 (anunselected patient group, adequateascertainment of death, andsurvival analysis) reports a 1-yearsuicide rate of 1%, very similar tothe median for all the reviewedstudies, and 100 times the suiciderate of the general population.
Twenty two studies with follow-upof between 1 and 5 years after self-harm reported suicide rates rangingfrom 0.6% to 11% with a medianof 3%.33-35,50,53,57,59,61,66–79 For 16studies with follow-up of between5 and 10 years, rates ranged from2% to 13%, with a median of3%.21,24,32,34,50,61,80–89 For three studiesreporting follow-up longer than 10years rates ranged from 4.7% to18% with a median of 7%.90–92
The suicide rate is therefore highestin the first year after an episode ofdeliberate self-harm (see Fig.2),
and within that year it is highest inthe first months.34 However, deathsby suicide are not confined to thefirst year, and studies with longer-term follow-up show that suicidedeaths after an episode of deliberateself-harm remain well above therate for the general population.
C.2 Repetition of non-fataldeliberate self-harmThirty seven studies report a 1-yearrepetition rate for deliberate self-harm.2,3,5,25,27,31,35–37,40,45–49,51,54,55,61,63,93–109
Excluding 6 studies that only dealwith selected sub-groups 48,54,55,105,106,109
and one study of 44 patients108
identified by frequent presentations,the remaining 30 studies reportedrepetition rates from 6% to 30%(median 16%, interquartile range13% to 18%). The wide variationin repetition is apparently accountedfor partly by differences in sampleselection, but also by differences inrates in different places.Repetition usually occurs early,particularly when there have beenprevious episodes. Median time torepetition among those with a historyof self-harm is about 72 days.37,97
D. Currentservices Guidelines for the assessment andmanagement of deliberate self-harm were produced by the DHSS in1984,110 and more recently by theHealth Advisory Service1 1 1 and by theRoyal College of Psychiatrists.112
These are not evidence-based, butthey provide a view of professionalconsensus in this area.
D.1 AssessmentThe DHSS guidelines110 recommendthat every patient should have aspecialist psychosocial assessment.The purpose of the assessment isto identify factors associated withsuicidal behaviour, to determinemotivation for the act, to identifypotentially treatable mentaldisorder, and to assess continuingrisk of suicidal behaviour. Basedon the assessment, a plan foraftercare should be made beforedischarge from hospital.
Deliberate self-harm EFFECTIVE HEALTH CARE 3DECEMBER 1998
Fig. 2 Suicide after non-fatal deliberateself-harm (dsh) according to duration offollow-up. N refers to number of publishedstudies in group.
1 ›1 to 4 ›5 to 9 10+Years of follow-up after non-fatal dsh episode
12
11
10
9
8
7
6
5
4
3
2
1
0
N=24
N=22
N=16
N=3
Suicidal intent is the degree towhich the person wished to die atthe time of the act. It is difficult toassess, because most people areambivalent and because reportedintent may change fairly quickly.113
The most widely used standardisedmeasure is the Suicide Intent Scale,28
which assesses the circumstancesof the act (such as planning andattempts to avoid rescue) and theperson’s reported intention to die.Suicidal intent is positively (butweakly) associated with themedical seriousness of the act.1
Motivation, other than the desireto die, may be assessed by askingthe patient or by inference fromthe circumstances. Examplesinclude a desire to communicatedistress, to obtain temporaryescape from intolerable worries, toeffect a change in the behaviour ofanother person, or to expressanother emotion such as anger.Assessment of motivation isunreliable as there are no widelyused standardised measures
The DHSS guidelines110
acknowledged that assessment andaftercare planning could beundertaken by staff other thanpsychiatrists – social workers orpsychiatric nurses for example –providing they had proper trainingand supervision. A number ofstudies have shown that thecontent and the quality of theirassessments are comparable withthose made by traineepsychiatrists.46,114–116 Non-medicalstaff, such as social workers, takelonger over assessments thanpsychiatrists, and recommendpsychiatric follow-up more often.116
One trial compared outcomesdepending on whether anassessment and managementdecision was made by apsychiatrist or junior medical(non-psychiatric) staff.45 There wasno significant difference inrepetition rate; 38/140 (27%)repeated within the year afterassessment by the general medicalteam, compared with 43/133 (32%)after psychiatric assessment.However, caution needs to beapplied in generalising from this
trial as it has not been replicated.In addition, junior medical staffreceived levels of training andsupervision that are not availablein many places, and not all staff oreligible patients participated.
Observational studies suggest thatwhen Accident and Emergency(A&E) Department staff makeassessments in routine clinicalpractice, the quality of note keepingis poor, and important informationsuch as assessment of mental stateor continuing suicidal thinking isfrequently not recorded.20,117–120
In practice, most assessments areundertaken by junior psychiatristsworking on a rota. Standards oftraining and supervision arepatchy,121 although it is unclearwhat effect this has on outcomes.
About three quarters of peoplewho harm themselves arrive athospital in the evening.122 It hasbeen recommended that suchpatients should be admittedovernight, with a view topsychosocial assessment in thedaytime.123 The advantages of thispolicy are that assessment is likelyto be of higher quality, and thataftercare is easier to arrangeduring office hours. Since thepatient has consumed alcohol inabout a half of episodes,20,22 or hisor her judgement may be impairedby the drugs ingested, thisargument has some appeal, but itis not widely accepted. In manyhospitals, more than half ofattenders are discharged from theA&E department.124–126 Patientswho leave hospital direct fromA&E, and especially those wholeave without a psychosocialassessment, are less likely to havebeen offered follow-up.57,124,127,128
D.2 AftercareSpecialist aftercare, when it isarranged, usually involves referralto psychiatric outpatients andsocial services.107,129,130 About 5–10%of cases lead directly to psychiatricadmission. In about a quarter ofhospitals there is a dedicatedmultidisciplinary self-harm team,but such teams follow-up only asmall minority of cases.124,131,132
There is no evidence comparingthe effectiveness of self-harmteams with that of generic services.
Non-statutory agencies,particularly in larger cities, mayoffer help not otherwise providedto people who self-harm. The bestknown of these agencies is theSamaritans. Early evaluations ofthe Samaritans producedconflicting evidence on itseffectiveness.133,134 There has beenno recent formal evaluation of thenon-statutory agencies which offerhelp to self-harming patients.
D.3 Deficiencies and variationsin practiceThe DHSS guidelines on deliberateself-harm are not followed in manyareas. Only about a half ofhospital attenders receive aspecialist psychosocial assessmentbefore they leave,124 many goinghome directly from the A&Edepartment as soon as they areassessed as physically fit to doso.135,136 Fewer than half are offeredany follow-up beyond the advicethat they might see their generalpractitioner. Reports from severalUK cities indicate that directdischarge without specialistassessment is becomingincreasingly common.117,124,132,137–139
There are large variations inpractice between services indifferent regions, and also betweenclinical teams in the samedistrict.126,128,137,140 For example,there are 3–4 fold differences inrates of discharge directly from theA&E department,124,126 and in ratesfor offering any form of psychiatricfollow-up.
One trial has compared theoutcomes after discharge fromA&E with those after hospitaladmission.141 There were nosignificant differences in outcome;3/27 admitted and 4/25 dischargedfrom A&E repeated within 16weeks. However these numberswere small and only a smallproportion of eligible patientsparticipated. Observational studieshave examined rates of repetitionin the following year for those whoare admitted, compared to those
4 EFFECTIVE HEALTH CARE Deliberate self-harm DECEMBER 1998
who go home from A&E.57,64 Thosewho are discharged withoutfollow-up have fewer known riskfactors for repetition, but their ratesof repetition are the same as thoseadmitted, suggesting that admissionmay confer some benefit.5 7 , 6 4
People who harm themselves arenot popular with health servicesstaff.142–147 Similar negative attitudesare also found in the psychiatricservices.148 They suffer from thestigma of psychiatric problems, andthey are often seen as undeservingand detracting from the clinicalcare of others whose illnesses arenot perceived as self-inflicted.People who harm themselvesrepeatedly, particularly those whocut themselves, may feel especiallysusceptible to this problem.
E. EffectivenessE.1 Are there effectiveinterventions?The results of the systematic reviewconducted by Hawton et al.41 aresummarised in Table 2, which alsocontains details of two trials reportedsince the review was published.42,43
The main interventions whichhave been evaluated in the trialsare: a brief psychological therapy(problem-solving therapy); moreintensive but conventionalpsychiatric care (special clinics,outreach, continuity of therapist,routine general hospital admission,longer-term contact); provision ofa crisis card; intensivepsychological therapy (dialecticbehaviour therapy, inpatienttherapy) and drug therapy(antidepressants, flupenthixol).
The trials varied in both thenature of the intervention andtheir aims. For example, theyincluded medical treatment aimedat reducing depressive symptomsor impulsivity; psychologicaltreatment designed to enhanceproblem-solving skills or to helpthe patient control self-injuriousbehaviours; and provision ofinformation intended to encourageeffective use of standard services
during a crisis. This heterogeneityin aims, coupled with widelyvarying study populations andinterventions, meant that littlepooling of data was possible.
The methodological quality of thereviewed randomised controlledtrials was poor. In particular,many studies were small, and noneincluded enough participants togive a reliable answer to theimportant question about theeffect of intervention on repetitionrates. Not all trials were analysedusing an intention-to-treatanalysis. Few used standardisedmeasures of outcomes (such asmood or quality of life) other thanrepetition. The trials recruitedhighly selected patient groups, andtheir results cannot be readilygeneralised to routine clinicalpractice. For example, clinicaltrials do not include patients wholeave hospital early, and often onlyinclude those who have acceptedpsychiatric referral; thus theirsamples are not representative ofthe self-harm population.
No intervention produced astatistically significant reduction inrepetition, although for some therewas a trend in that direction.However, three interventions seempromising.
The first of these involvesproviding patients with a crisiscard which carries advice aboutseeking help in the event of futuresuicidal feelings. In the Bristolstudy,55 which included onlypeople who had taken their first-ever drug overdose, possession ofthe card enabled its holder tospeak to a psychiatrist at shortnotice and to request psychiatricadmission in a crisis. Although themajority did not avail themselvesof any of the offers on the card,there was a suggestion of areduction in repetition. However,an attempted replication (not yetpublished) has produced anegative result, perhaps becauserepeaters were also included in theintervention.149 From a clinicalperspective, it is reasonable toexpect that people who attendhospital after an episode of self-
harm should be given advice aboutlocal services which could be usedin a crisis or when self-harm iscontemplated. But because thebest mode of delivering this advice(or its likely benefits) is unknown,further research is needed.
The second intervention isproblem-solving therapy. This is abrief treatment aimed at helpingthe patient to acquire basicproblem-solving skills, by takinghim or her through a series ofsteps: identification of personalproblems; constructing a problemlist which clarifies and prioritisesthem; reviewing possible solutionsfor a target problem; implementingthe chosen solution; reappraisingthe problem; reiterating theprocess; training in problem-solving skills for the future.150
This usually involves about 6 onehour sessions, with some readingmaterials, and work to beundertaken between sessions. Itcan be delivered by any experiencedmental health professional, withsuitable training and supervision.Standardisation can also be improvedby using a therapy manual.
Problem-solving has a theoreticalbasis since there is evidence thatpeople who harm themselves arepoor problem-solvers, which maybe linked to other importantcharacteristics such as hopelessness.Problem-solving therapy has beenshown to be an effective treatmentfor depression in other settings,150
and in self-harm studies it has ledto improvement in other relevantoutcomes such as mood and socialadjustment.11 It may therefore besuitable for some individuals,although the scope of itsapplicability is unclear from theexisiting evidence.
The third intervention, dialecticbehaviour therapy, was introducedas a method of helping those whoengage in chronic repetitive self-harm, particularly when they haveassociated borderline personalitycharacteristics.39 It is intensive,involving in its full form a year ofindividual therapy, group sessions,social skills training and access tocrisis contact. The interest it has
DECEMBER 1998 Deliberate self-harm EFFECTIVE HEALTH CARE 5
6 EFFECTIVE HEALTH CARE Deliberate self-harm DECEMBER 1998
Study
Gibbons et al.(UK, 1978) 96
Hawton et al.(UK, 1987) 49
Salkovskis et al.(UK, 1990) 164
McLeavey et al.(Ireland, 1994) 165
Chowdhury et al.(UK, 1973) 166
Welu (USA, 1977) 167
Hawton et al.(UK, 1981) 48
Allard et al.(Canada, 1992) 68
Van Heeringen et al.(Belgium, 1995)156
Van der Sande et al.(Netherlands, 1997)65
Morgan et al.(UK, 1993)55
Cotgrove et al.(UK, 1995)168
Linenan et al.(USA, 1991)169
Liberman and Eckman(USA, 1981)170
Details of participants
Patients over 17 years who presentedto A&E department after deliberateself-poisoning; repeaters (1or moreattempt) and first timers; 71% female
Patients over 16 years admitted togeneral hospital for self-poisoning;31% repeaters; 66% female
Patients aged 16–65 years (mean27.5) referred by duty psychiatristafter antidepressant self-poisoningassessed in A&E department; allrepeaters with high risk of furtherrepetition; 50% female
Patients aged 15–45 years (mean24.4) admitted to A&E departmentafter self-poisoning; 35.6%repeaters: 74% female
Patients (all repeaters) admitted togeneral hospital after deliberateself-harm; 57% female
Suicide attempters over 16 yearsbrought to A&E department; 60%repeaters; % female not given
Patients aged 16 years and over(mean 25.3) admitted to generalhospital after deliberate self-poisoning;32% repeaters; 70% female
Patients seen in A&E departmentfor suicide attempt; 50% repeaters;55% female
Patients aged 15 years and overtreated in A&E department aftersuicide attempt; 30% repeaters;43% female
Patients aged 16 years and over(mean 36.3) admitted to hospitalafter suicide attempt; 73%repeaters; 66% female
Mean age 30 years; patientsadmitted after first episode ofdeliberate self-harm; % female notgiven
Patients aged 18–45 years whohad self-harmed within 8 weeksbefore entering study; all female;all multiple repeaters of self-harm
Patients aged 18–47 years (mean29.7) all repeaters; patientsreferred by psychiatric emergencyservice or hospital A&Edepartment after deliberate self-harm; 67% female
Interventions
Experimental (n=200): crisis orientated, time limited, task centredsocial work at home (problem solving intervention). Control(n=200): routine service – 54% GP referral, 33% psychiatricreferral, 13% other referral
Experimental (n=41): outpatient problem orientated therapy bynon-medical clinicians. Control (n=39): GP care (e.g., individualsupport, marital therapy) after advice from clinician
Experimental (n=12): domiciliary cognitive behavioural problemsolving treatment. Control (n=8): treatment as usual (GP care)
Experimental (n=71): special aftercare – regular outpatientappointments; patients also seen without appointments; home visitsto patients who missed appointments; emergency 24 hour telephoneaccess. Control (n=84): normal aftercare – outpatient appointmentwith psychiatrist and/or social worker; non-attenders not pursued
Experimental (n=63): special outreach programme – community mentalhealth team contacted patient immediately after discharge; homevisit arranged; weekly/twice weekly contact with therapist. Control(n=57): routine care – appointment for evaluation at the communitymental health centre next day at request of treating physician
Experimental (n=48): domiciliary therapy (brief problem orientated)as often as therapist thought necessary; open telephone access togeneral hospital service. Control (n=48): outpatient therapy once aweek in outpatient clinic in general hospital
Experimental (n=76): intensive intervention – schedule of visits wasarranged including at least one home visit; therapy provided whenneeded; reminders (telephone or written) and home visits made ifappointments missed. Control (n=74): treatment by another staffteam in the same hospital
Experimental (n=258): special care – home visits by nurse to patientswho did not keep outpatient appointments, reasons for not attendingdiscussed and patient encouraged to attend. Control (n=258):outpatient appointments only; non-compliant patients not visited
Experimental (n=140): brief psychiatric unit admission,encouraging patients to contact unit on discharge; outpatienttherapy plus 24 hour emergency access to unit. Control (n=134):usual care – 25% admitted to hospital, 65% outpatient referral
Experimental (n=101): standard care plus green card (emergencycard indicating that doctor was available and how to contactthem). Control (n=111): standard care – for example, referral backto primary healthcare team, psychiatric inpatient admission
Experimental (n=47): standard care plus green card (emergencycard); acted as passport to readmission into paediatric ward inlocal hospital. Control (n=58): standard follow-up treatment fromclinic or child psychiatry department.
Experimental (n=32): dialectical behaviour therapy (individualand group work) for 1 year; telephone access to therapist. Control(n=31): months treatment as usual: 73% individual psychotherapy
Experimental (n=12): inpatient treatment with behaviour therapy.Control (n=12): inpatient treatment with insight orientated therapy;both groups received individual and group therapy plus aftercareat community mental health centre or with private therapist
Experimental
27/200 (13.5)
3/41 (7.3)
3/12 (25.0)
2/19 (10.5)
17/71 (23.9)
3/62 (4.8)
5/48 (10.4)
22/63 (34.9)
21/196 (10.7)
24/140 (17.1)
5/101 (5.0)
3/47 (6.4)
5/19 (26.3)
2/12 (16.7)
Control
29/200 (14.5)
6/39 (15.4)
4/8 (50.0)
5/20 (25.0)
19/84 (22.6)
9/57 (15.8)
7/48 (14.6)
19/63 (30.2)
34/195 (17.4)
20/134 (14.9)
12/111 (10.8)
7/58 (12.1)
12/20 (60.0)
3/12 (25.0)
Problem-solving therapy v standard aftercare
Intensive care plus outreach v standard care
Proportion (%) ofparticipants who repeatedbehaviour during follow-up
Emergency card v standard aftercare
Table 2 Summary of participants, interventions, size of trial, and proportion (%) of participants who repeated behaviour during follow-up
Inpatient behaviour therapy v inpatient insight orientated therapy
Dialectical behaviour therapy v standard aftercare
provoked is due to the suggestionthat it leads to a reduction in self-harming behaviour in a group ofpeople for whom the services havelittle or nothing else to offer.However, because it is an intensiveintervention, better evidence of itsapplicability and cost-effectivenessis required, but it does offer aninteresting model for the care ofpeople who have problems whichare among the most intractable inpsychiatry.
E.2 Can services be restricted tocertain high risk groups?Many services restrict follow-up tothose perceived as being at highrisk of repetition or suicide. Is thereevidence to support this practice?
Scales to predict suicide haveextremely weak predictive power,because the absolute risk of
suicide is so low. Scales forprediction of repetition of non-fatal self-harm seem a betterproposition, but unfortunately,their performance is not of asmuch practical value as might behoped, for two main reasons.
First, self-harm repetition scalesare not accurate because theindividual risk factors whichconstitute them have poor positivepredictive value. For example, inone study of around 1000 patients,which found four of the above riskfactors for repetition to besignificantly more frequent amongthose who repeated, the best riskfactor (past psychiatric contact)had a positive predictive value ofonly 21%.57 Adding the individualitems to produce a composite riskscore does not add sufficiently tothis predictive value.
Second, scores on risk of repetitionscales show a positively skewednormal distribution, that is, theyare not evenly distributed throughthe range but show an extended“tail” towards the high scores. As aconsequence, the apparently goodpositive predictive value from ahigh score does not mean riskscales are accurate in predictingrepetition for the wholepopulation. Although high scorersfrequently repeat, only a fewpeople score high. Most repeatsinvolve the much larger numberof people at lower apparent risk.
What this means is that usingexisting risk assessments, thesmaller high-risk group willcontain no more people who willeventually repeat than does thelarger low-risk group.25 Aneffective or equitable service
Deliberate self-harm EFFECTIVE HEALTH CARE 7DECEMBER 1998
Torhorst et al.(Germany, 1987)171
Waterhouse and Platt(UK, 1990)141
Montgomery et al.(UK, 1979)172
Hirsch et al. (UK,1982)173 R Draper,S Hirsch (personalcommunication)
Montgomery et al.(UK, 1983)174
Verkes et al.(Netherlands, 1998)42
Torhorst et al.(Germany, 1988)175
Harrington et al.(UK, 1998)43
Patients referred to toxologicaldepartment of TechnicalUniversity Munich after deliberateself-poisoning; 48% repeaters;62% female
Patients aged 16 years and over(mean 30.3) admitted to A&Edepartment for deliberate self-harm; 36% repeaters; 63% female
Patients aged 18–68 years (mean35.3) admitted after suicidal act;all repeaters; 70% female
Patients aged 16–65 years admittedafter deliberate self-poisoning; %repeaters and % female not given
Patients with personality disorders(mean age 35.7 years) admitted tomedical ward after deliberate self-harm; all repeaters; 66% female
Adults referred after self-poisoningwhich was not their lifetime first,who did not have major depression.Analysed according to number ofprevious episodes
All patients repeaters who haddeliberately self-poisoned; %female not given
All children aged 16 years orless, admitted to a paediatricward after deliberate self-poisoning, and referred forpsychiatric assessment; 90% girls
Experimental (n=68): continuity of care – therapy with sametherapist who assessed patient in hospital after attempt. Control(n=73): change months of care – therapy with different therapistthan seen at hospital assessment
Experimental (n=38): general hospital admission. Control (n=39)discharge from hospital. On discharge both groups advised tocontact GP if they needed further help
Experimental (n=18): 20mg intramuscular flupenthixol deconatefor 6 months. Control (n=19): placebo for 6 months
Experimental (n=76): antidepressants – either 30–60mg mianserinfor 6 weeks or 75–150 mg nomifensine for 6 weeks. Control(n=38): placebo for 6 weeks
Experimental (n=17): mianserin 30mg for 6 months. Control(n=21): placebo
Experimental (n=46) paroxetine 40mg/day, control (n=45)placebo for 12 months
Experimental (n=40): long-term therapy – one therapy session amonth for 12 months. Control (n=40): short-term therapy – 12weekly therapy months sessions for 3 months; all participants hadbrief crisis intervention (3 days) in hospital
Experimental (n=85) 5 sessions home-based family therapy.Control (n=77) received treatment as usual in child psychiatryclinic, averaging 3.6 sessions
12/68 (17.6)
3/38 (7.9)
3/14 (21.4)
16/76 (21.1)
8/17 (47.1)
15/46 (33)
9/40 (22.5)
11/74 (15)
4/73 (5.5)
4/39 (10.3)
12/16 (75.0)
5/38 (13.2)
12/21 (57.1)
21/45 (47)
9/40 (22.5)
11/75 (15)
Same therapist (continuity of care) v different therapist (change of care)
General hospital administration v discharge
Flupenthixol v placebo
Antidepressants v placebo
Long-term therapy v short-term therapy
Family therapy v standard care
Table 2 Continued
cannot be based on application ofexisting risk assessment as ameans of identifying cases. This isnot to say that identification ofrisk is pointless in clinical practice,simply that it is insufficientlyaccurate to serve as the basis forinterventions aimed at tackling thepublic health problem representedby deliberate self-harm.
E.3 What style of serviceprovision should be adopted?
Services in the general hospital
Even when aftercare is arranged, itis not always taken up. Accordingto the type of service reported,30–70% of those offeredpsychiatric follow-up either do notattend at all or drop out after theirfirst appointment.49,97,151–153 This istrue even when the referral is to aspecialist service such as an alcoholand addictions service, or whenthe clinic is arranged in the A&Edepartment so that the patient isreturning to the place (perhaps tosee the same person) where theoriginal assessment wasundertaken.154
Three methods have been suggestedto improve contact rates. Writtenprompts are easy to provide, butare relatively ineffective.155
Motivational interviewing, whichaims to encourage a rationalapproach to health-relatedbehaviour, has been widely usedin other settings, but has thedisadvantage that it requirestraining to administer.54 The bestrates of contact are achieved byoutreach programmes48,156 whichare the only means of maintainingcontact with the 20–30% ofpatients who will not attend clinicappointments.
Aftercare through the usualpsychiatric services is unsatisfactorybecause repetition of self-harmtends to occur early (see above); aquarter of those with a history ofpast attempts will repeat within 3weeks.97 Few routine clinics canoffer new appointments within thistimescale, particularly for thenumbers of people for whom itwould be required.
The general practitioner and deliberateself-harm
Around 50–60% of patients havevisited their general practitioner inthe month before an episode ofself-harm.18,157,158 For this reason,attention has turned to thepossibility of basing primary orsecondary prevention in generalpractice. However, up to half ofGP consultations before a self-harm episode are not for overtlypsychosocial reasons,159 so theopportunities for detection andprimary intervention at thiscontact may not be as great as issometimes supposed.
The most frequent managementdecision made after assessment isthat the patient should return tosee his or her GP. Around half ofpatients do visit their GP in the1–2 months after anepisode.45,157,160,161 As noted above,even when specialist aftercare isproposed, there are difficulties inarranging predictable follow-upwith psychiatric services. Thisraises the question of the role ofthe GP in the management of self-harm. Even in a trial assessingintervention in primary care, nearlyhalf of those scheduled to receiveGP counselling had not seen theirGP within two months of theoriginal episode.49 These figuresshow that any intervention ingeneral practice would need to havea component aimed at achievinghigher attendance rates than areachieved through routine practice.
There is no research evidence whichanswers the question of whatintervention should be offered byGPs. In Sweden, for GPs who weretaught skills in recognising andtreating depression there was anapparent reduction in the suiciderate,162 but the relevance of thisstudy to management after anepisode of self-harm is unclear.
E.4 What are the financialimplications? There are no detailed UK dataconcerning the costs of providingself-harm services, and none of thetrials reviewed above included acost-effectiveness analysis.
In one hospital, it was estimatedthat self-harm absorbed only about0.4% of the hospital budget.130
From the results of this study it ispossible to estimate the generalhospital costs of deliberate self-harm at around £45–50 millionannually (at 1998 prices). This isbecause, despite its importance asa reason for admission, mostinpatient episodes last only 24–48hours and incur relatively lowtreatment costs. Treatment onintensive care units accounts forless than 10% of hospital costs ofdeliberate self-harm.130
A study from Australia examinedthe costs associated withcentralising self-harm services andarranging routine general hospitaladmission and psychosocialassessment.163 Costs for thecentralised comprehensive servicewere lower than for other servicesin the state, mainly as a result offewer inpatient days arising fromself-harm.
F. Implicationsfor practice■ All hospital attendance
following deliberate self-harmshould lead to a psychosocialassessment. This should aim toidentify motives for the act,and associated problems whichare potentially amenable tointervention such aspsychological or socialproblems, mental disorder, andalcohol and substance misuse.
■ Since assessments undertakenas part of routine clinicalpractice are incomplete and ofvariable quality, staff whoundertake assessments shouldreceive specialist training andhave supervision available.
■ Direct discharge from A&Eshould only be contemplated ifa psychosocial assessment andaftercare plan can be arrangedin A&E prior to discharge.
■ Aftercare arrangements shouldinclude the provision of verbal
DECEMBER 19988 EFFECTIVE HEALTH CARE Deliberate self-harm
DECEMBER 1998 Deliberate self-harm EFFECTIVE HEALTH CARE 9
and written information onservices available for people whoare contemplating self-harm.
■ There is insufficient evidenceto recommend a specificclinical intervention afterdeliberate self-harm. However,brief psychological therapiessuch as interpersonal therapyand problem-solving therapyare effective in the treatment ofdepression in similar clinicalsettings, and the latter has beenshown to have benefits (if notreducing repetition) after self-harm. Opportunities for referralfor such therapies should beavailable to suitable patients.
■ GPs should have ready accessto training and advice aboutthe assessment andmanagement of self-harmpatients in primary care.
■ Accessible and comprehensiveservices will need a mechanismfor engaging people who do notattend routine clinicappointments. Access to follow-up needs to be rapid becauserepetition occurs soon after theepisode.
■ Service providers should workto improve attitudes towardsself-harming patients, forexample through training aimedat increasing knowledge aboutself-harm, and perhaps throughcontact with service users.
G. Implicationsfor research■ Research is needed to determine
the effect of discharge directlyfrom the A&E department afterpresentation with deliberateself-harm; whether it reducesthe quality or outcomes ofpsychosocial assessment, theeffect it has on subsequentcontact with services, and onoutcomes.
■ Research is needed to establishthe clinical and cost-effectiveness of potentialinterventions. Trials should belarge enough to determine
whether the intervention reducesrepetition, but should examineother relevant outcomesincluding use of health andsocial services, quality of life,mood, interpersonal problemsand social functioning.
■ Trials might focus on specificsubgroups, such as chronicrepeaters or those sufferingfrom alcohol dependence, iflarge enough sample sizes canbe recruited. Alternatively, ifthe subjects are to berepresentative of all self-harmpatients, they should includeall hospital attenders, and notonly patients recruited frompsychiatric services or patientswho visit their GP.
■ Research is needed into formsof self-harm other than drugsoverdosage, and in particularinto cutting – its causes,outcomes, and effectivetreatments.
Research MethodsFor the review of trials of intervention afterdeliberate self-harm, a literature search was carriedout of the following databases: Medline (1966-Aug1998) PsycLit (1974-Aug 1998) Embase (1980-Aug1998) and the Cochrane Controlled Trials Register(1998). The search used the CochraneCollaboration search strategy for identifying clinicaltrials on Medline, with modifications for the otherdatabases. Ten specialist journals in psychiatry andclinical psychology were also hand-searched.
Trials were included if the participants had engagedin deliberate self-harm shortly before entry into thetrial, if there was clear evidence of randomisation totreatment and control groups and repetition of self-harm was reported as an outcome measure.Concealment of allocation was rated independentlyby two reviewers. Data were extractedindependently by two reviewers.
For the other elements of the bulletin, the abovesearches were made again of Medline, Embase,PsychLit and Cinahl. Articles were included if theywere in English and if they reported on populationsin which deliberate self-harm was the presentingfeature. Excluded were reports concerned solelywith children and adolescents, with self-injury inpeople with learning difficulties, or with self-injuryin prisons. Data were extracted by two reviewers,and secondary references cross-checked by a third.
For studies describing the rate of suicide followingan episode of non-fatal self-harm: those publishedearlier than 1970 are of uncertain relevance tocurrent practice because of substantial recentchanges in the epidemiology of self-harm -including for example the switch from tranquillisersto analgesics as the main substances ingested andthe increasing incidence among men. Studies wereincluded if they were published since 1970, andthey followed-up a sample which was likely to berepresentative of general hospital attenders. Thusthey were excluded if they followed-up onlyselected subgroups such as children andadolescents, or people who were identified followinginpatient admission to specialist psychiatric orresearch units. The latter groups are likely to be athigh risk, but because admission criteria are unclearor not reproducible, it is not possible to generalisefrom reports on their outcomes.
References1. Hawton K, Catalan J. Attempted suicide: A
practical guide to its nature and management.Oxford: Oxford University Press, 1987.
2. Hawton K, Fagg J. Trends in deliberate selfpoisoning and self injury in Oxford, 1976–90.BMJ 1992;304:1409–11.
3. Hawton K, Fagg J, Simkin S, et al. Trends indeliberate self- harm in Oxford, 1985-1995:Implications for clinical services and theprevention of suicide. Br J Psychiatry1997;171:556–60.
4. Hawton K, Fagg J, Simkin S, et al. Deliberateself-harm in Oxford 1996. Oxford: OxfordUniversity Department of Psychiatry 1996.
5. Schmidtke A, Bille-Brahe U, De Leo D, et al.Attempted suicide in Europe: rates, trends andsociodemographic characteristics of suicideattempters during the period 1989–1992.Results of the WHO/EURO Multicentre Studyon Parasuicide. Acta Psychiatr Scand1996;93:327–38.
6. Foster T, Gillespie K, McClelland R. Mentaldisorder and suicide in Northern Ireland. Br JPsychiatry 1997;170:447–52.
7. Owens D, House A. General hospital servicesfor deliberate self harm. J R Coll PhysiciansLond 1994;28:370–1.
8. National Health Service ManagementExecutive. Health of the Nation Key AreaHandbook: Mental Illness, 1993.
9. Department of Health. Our Healthier Nation: acontract for health, 1998.
10. McLoone P, Crombie I. Hospitalisation fordeliberate self-poisoning in Scotland from1981 to 1993: trends in rates and types ofdrugs used. Br J Psychiatry 1996;169:81–5.
11. House A, Owens D, Storer D. Psycho-socialintervention following attempted suicide: Isthere a case for better services? Int RevPsychiatry 1992;4:15–22.
12. Charlton J, Kelly S, Dunnell K, et al. Trends insuicide deaths in England and Wales. PopulTrends 1992;69:10–6.
13. Charlton J, Kelly S, Dunnell K, et al. Suicidedeaths in England and Wales: trends in factorsassociated with suicide deaths. Popul Trends1993;71:34–42.
14. Bancroft J, Skrimshire A, Casson J, et al. Peoplewho deliberately poison or injure themselves:their problems and their contacts with helpingagencies. Psychol Med 1977;7:289–303.
15. Platt S, Kreitman N. Long term trends inparasuicide and unemployment in Edinburgh,1968–87. Soc Psychiatry Psychiatr Epidemiol1990;25:56–61.
16. Sakinofsky I, Roberts R. Why parasuicidesrepeat despite problem resolution. Br JPsychiatry 1990;156:399–405.
17. Sakinofsky I, Roberts R, Brown Y, et al. Problemresolution and repetition of parasuicide. Aprospective study. Br J Psychiatry 1 9 9 0 ; 1 5 6 : 3 9 5 – 9.
18. Morgan HG, Burns CC, Pocock H, et al.Deliberate self-harm: clinical and socio-economic characteristics of 368 patients. Br JPsychiatry 1975;127:564–74.
19. Newson-Smith J, Hirsch S. Psychiatricsymptoms in self-poisoning patients. PsycholMed 1979;9:493–500.
21. Suokas J, Lonnqvist J. Suicide attempts inwhich alcohol is involved: a special group ingeneral hospital emergency rooms. ActaPsychiatr Scand 1995;91:36–40.
22. Wylie K, House A, Storer D, et al. Deliberateself-harm and substance dependence: Themanagement of patients seen in the generalhospital. J Ment Health Admin1996;23:246–52.
24. Cullberg J, Wasserman D, Stefansson C. Whocommits suicide after a suicide attempt? An 8to 10 year follow up in a suburban catchmentarea. Acta Psychiatr Scand 1988;77:598–603.
DECEMBER 199810 EFFECTIVE HEALTH CARE Deliberate self-harm
25. Kreitman N, Foster J. Construction andselection of predictive scales, with specialreference to parasuicide. Br J Psychiatry1991;159:185–92.
27. Bancroft J, Marsack P. The repetitiveness ofself-poisoning and self-injury. Br J Psychiatry1977;131:394–9.
28. Beck AT, Schuyler D, Herman J. Developmentof suicidal intent scales. in Beck AT, ResnickH, Lettieri D, editors The prediction of suicide:Bowie, Maryland, 1974.
29. Beck AT, Steer R, Kovacs M, et al.Hopelessness and eventual suicide: a 10-yearprospective study of patients hospitalized withsuicidal ideation. Am J Psychiatry1985;142:559–63.
30. Beck A, Steer R. Clinical predictors of eventualsuicide: a 5–10-year prospective study ofsuicide attempters. J Affect Disord1989;17:203–9.
31. Buglass D, Horton J. The repetition ofparasuicide: A comparison of three cohorts. BrJ Psychiatry 1974;125:168–74.
32. Ekeberg O, Ellingsen O, Jacobsen D. Suicideand other causes of death in a five-yearfollow-up of patients treated for self-poisoningin Oslo. Acta Psychiatr Scand 1991;83:432–7.
33. Greer S, Bagley C. Effect of psychiatricintervention in attempted suicide: a controlledstudy. BMJ 1971;1:310–2.
34. Hawton K, Fagg J. Suicide, and other causes ofdeath, following attempted suicide. Br JPsychiatry 1988;152:359–66.
35. Morgan H, Barton J, Pottle S, et al. Deliberateself-harm: a follow-up study of 279 patients. BrJ Psychiatry 1976;128:361–8.
36. Owens D, Dennis M, Read S, et al. Outcome ofdeliberate self-poisoning. An examination ofrisk factors for repetition. Br J Psychiatry1994;165:797–801.
37. Wilkinson G, Smeeton N. The repetition ofparasuicide in Edinburgh 1980–1981. SocPsychiatry 1987;22:14–9.
38. Kreitman N, Casey P. Repetition ofparasuicide: an epidemiological and clinicalstudy. Br J Psychiatry 1988;153:792–800.
39. Linehan MM. Cognitive-behavioural treatmentof borderline personality disorder. New York:Guilford, 1993.
40. Van der Sande R, Buskens E, Allart E, et al.Psychosocial intervention following suicideattempt: a systematic review of treatmentinterventions. Acta Psychiatr Scand1997;96:43–50.
41. Hawton K, Arensman E, Townsend E, et al.Deliberate self harm: systematic review ofefficacy of psychosocial and pharmacologicaltreatments in preventing repetition. BMJ1998;317:441–7.
42. Verkes R, van der Mast R, Hengeveld M, et al.Reduction by paroxetine of suicidal behaviourin patients with repeated suicide attempts butnot major depression. Am J Psychiatry1998;155:543–7.
43. Harrington R, Kerfoot M, Dyer E, et al.Randomized trial of a home-based familyintervention for children who havedeliberately poisoned themselves. J Am AcadChild Adolesc Psychiatry 1998;37:512–8.
44. Kennedy P, Kreitman N. An epidemiologicalsurvey of parasuicide (‘attempted suicide’) inGeneral Practice. Br J Psychiatry1973;123:23–34.
45. Gardner R, Hanka R, O’Brien VC, et al.Psychological and social evaluation in cases ofdeliberate self-poisoning admitted to a generalhospital. BMJ 1977;2:1567–70.
46. Gardner R, Hanka R, Roberts S, et al.Psychological and social evaluation in cases ofdeliberate self-poisoning seen in an accidentdepartment. BMJ 1982;284:491–3.
47. Goldacre M, Hawton K. Repetition of self-poisoning and subsequent death inadolescents who take overdoses. Br JPsychiatry 1985;146:395–8.
48. Hawton K, Bancroft J, Catalan J, et al.Domiciliary and out-patient treatment of self-poisoning patients by medical and non-medical staff. Psychol Med 1981;11:169–77.
49. Hawton K, McKeown S, Day A, et al.Evaluation of out-patient counsellingcompared with general practitioner carefollowing overdoses. Psychol Med1987;17:751–61.
50. Johnsson F, Ojehagen A, Traskman-Bendz L. A5-year follow-up study of suicide attempts.Acta Psychiatr Scand 1996;93:151–7.
51. Kennedy P. Efficacy of a Regional PoisoningTreatment Centre in preventing furthersuicidal behaviour. BMJ 1972;4:255–7.
52. Lonnqvist J, Achte K. Follow-up study on theattempted suicides among the elderly inHelsinki in 1973–1979. Crisis 1985;6:10–8.
53. Lonnqvist J, Ostama A. Suicide following thefirst suicide attempt: A five year follow-upusing a survival analysis. Psychiatr Fenn1991;22:171–9.
54. Moller HJ. Efficacy of different strategies ofaftercare for patients who have attemptedsuicide. J R Soc Med 1989;82:643–7.
55. Morgan H, Jones E, Owen J. Secondaryprevention of non-fatal deliberate self-harm.The green card study. Br J Psychiatry1993;163:111–2.
56. Ojehagen A, Danielsson M, Traskman-Bendz L.Deliberate self-poisoning: treatment follow-upof repeaters and non-repeaters. Acta PsychiatrScand 1992;85:370-5.
57. Owens D, Dennis M, Jones S, et al. Self-poisoning patients discharged from accidentand emergency: risk factors and outcome. J RColl Physicians Lond 1991;25:218–22.
58. Paerregaard G. Suicide among attemptedsuicides: a 10-year follow-up. Suicide.1975;5:140–4.
59. Pallis DJ, Gibbons J, Pierce D. Estimatingsuicide risk among attempted suicides II.Efficiency of predictive scales after the attemptI. Br J Psychiatry 1984;144:139–48.
60. Rosen DH. The serious suicide attempt. Five-year follow-up study of 886 patients. JAMA1976;235:2105–9.
61. Rygnestad T. A prospective 5-year follow-upstudy of self-poisoned patients. Acta PsychiatrScand 1988;77:328–31.
62. Rygnestad T. Prospective study of social andpsychiatric aspects in self poisoned patients.Acta Psychiatr Scand 1982;66:139–53.
63. Stocks R, Scott A. What happens to patientswho frequently harm themselves? Aretrospective one-year outcome study. Br JPsychiatry 1991;158:375–8.
64. Suokas J, Lonnqvist J. Outcome of attemptedsuicide and psychiatric consultation: riskfactors and suicide mortality during a five-year follow-up. Acta Psychiatr Scand1991;84:545–9.
65. Van der Sande R, Van Rooijen L, Buskens E, etal. Intensive in-patient treatment andcommunity intervention versus routine careafter attempted suicide. Br J Psychiatry1997;171:35–41.
66. Adam K, Isherwood J, Taylor G, et al.Attempted suicide in Christchurch: three-yearfollow-up of 195 patients. NZ Med J1981;93:376–81.
67. Adam K, Valentine J, Scarr G, et al. Follow-upof attempted suicide in Christchurch. Aust NZ J Psychiatry 1983;17:18–25.
68. Allard R, Marshall M, Plante M. Intensivefollow-up does not decrease the risk of repeatsuicide attempts. Suicide Life Threat Behav1992;22:303–14.
69. Bille-Brahe U, Jessen G. Repeated suicidalbehaviour: a two-year follow-up. Crisis1994;15:77–82.
70. Buglass D, McCulloch J. Further suicidalbehaviour: the development and validation ofpredictive scales. Br J Psychiatry1970;116:483–91.
71. Gharagozlou H, Hadjmohammadi M. Reporton a three-year follow-up of 100 cases ofsuicidal attempts in Shiraz, Iran. Int J SocPsychiatry 1977;23:209–10.
72. Hansen W, Wang A. Suicide attempts in aDanish region. Soc Psychiatry1984;19:197–201.
73. Hassanyeh F, O’Brien G, Holton A, et al.Repeat self-harm: an 18-month follow-up. ActaPsychiatr Scand 1989;79:265–7.
74. Pallis D, Pierce D. Recognizing the suicidaloverdose. J R Soc Med 1979;72:565–71.
75. Pallis DJ, Barraclough BM, Levey AB, et al.Estimating suicide risk among attemptedsuicides: I. The development of new clinicalscales. Br J Psychiatry 1982;141:37–44.
76. Pederson A, Awad G, Kindler A.Epidemiological differences between whiteand nonwhite suicide attempters. Am JPsychiatry 1973;130:1071-6.
77. Siani R, Garzotto N, Tansella CZ, et al.Predictive scales for parasuicide repitition:further results. Acta Pscychiat Scand1979;59:17–23.
81. Braftos O. Attempted suicide: a comparativestudy of patients who have attempted suicideand psychiatric patients in general. ActaPsychiatr Scand 1971;47:38–55.
82. Kotila L, Lonnqvist J. Adolescents who makesuicide attempts repeatedly. Acta PsychiatrScand 1987;76:386–93.
83. Lonnqvist J, Niskanen P, Achte KA, et al. Self-poisoning with follow-up considerations.Suicide. 1975;5:39–46.
84. Lonnqvist J, Karha E. Suicide attempts inHelsinki. Psychiatr Fenn 1984;15:135–45.
85. Palsson SP, Jonsdottir G, Petursson H. Themortality risk of psychiatric emergencypatients. A follow-up study. Nordic Journal ofPsychiatry 1996;50:207–16.
86. Pierce D. Suicidal intent and repeated self-harm. Psychol Med 1984;14:655–9.
87. Pierce DW. The predictive validation of asuicide intent scale: a five year follow-up. Br JPsychiatry 1981;139:391–6.
88. Rosenman SJ. Subsequent deaths afterattempted suicide by drug overdose in thewestern region of Adelaide, 1976. Med J Aust1983;2:496–9.
89. Sundqvist-Stensman U. Suicides amongpersons treated for self-poisoning at an ICU.Opuscula Medica 1988;33:71–6.
90. De Moore G, Robertson AR. Suicide in the 18years after deliberate self-harm. A prospectivestudy. Br J Psychiatry 1996;169:489–94.
91. Ekeberg O, Ellingsen O, Jacobsen D. Mortalityand causes of death in a 10-year follow-up ofpatients treated for self-poisonings in Oslo.Suicide Life Threat Behav 1994;24:398–405.
92. Mutzell S. Survival after attempted suicide inSweden. A follow-up after 20 years.International Journal of Adolescence and Youth1997;6:315–28.
93. Arensman E, Kerkhof AJ, Hengeveld MW, et al.Medically treated suicide attempts: a four yearmonitoring study of the epidemiology in TheNetherlands. J Epidemiol Community Health1995;49:285–9.
94. Diekstra RF, van Egmond M. Suicide andattempted suicide in general practice, 1979-1986. Acta Psychiatr.Scand. 1989;79:268–75.
95. Garzotto N, Siani R, Tansella CZ, et al. Crossvalidation of a predictive scale for subsequentsuicidal behaviour in an Italian sample. Br JPsychiatry 1976;128:137–40.
96. Gibbons JS, Butler J, Urwin P, et al. Evaluationof a social work service for self-poisoningpatients. Br J Psychiatry 1978;133:111–8.
97. Gilbody S, House A, Owens D. The earlyrepetition of deliberate self harm. J R CollPhysicians Lond 1997;31:171–2.
98. Gutierrez E, Gamo E, Guindeo J, et al.Epidemiologic study of attempted suicideattended at the emergency in a generalhospital. Actas Luso-Espanolas De NeurologiaPsiquiatria Y Ciencias Afines 1997;25:172–81.
99. Hawton K, Fagg J, McKeown SP.Alcholism,alcohol and attempted suicide.Alchohol and Alcoholism 1989;24:3–9.
DECEMBER 1998 Deliberate self-harm EFFECTIVE HEALTH CARE 11
100. Hawton K, O’Grady J, Osborn M, et al.Adolescents who take overdoses: theircharacteristics, problems and contacts withhelping agencies. Br J Psychiatry1982;140:118–23.
101. Hjelmeland H. Repetition of parasuicide: apredictive study. Suicide Life Threat Behav1996;26:395–404.
103. Morton MJ. Prediction of repetition ofparasuicide: with special reference tounemployment. Int J Soc Psychiatry1993;39:87–99.
104. Myers ED. Predicting repetition of deliberateself-harm: a review of the literature in thelight of a current study. Acta Psychiatr Scand1988;77:314–9.
105. Nielsen B, Wang AG, Bille-Brahe U. Attemptedsuicide in Denmark. IV. A five-year follow-up.Acta Psychiatr Scand 1990;81:250–4.
106. Nowers M. Deliberate self-harm in the elderly:A survey of one London borough. Int J GeriatrPsychiatry 1993;8:609–14.
107. Platt S, Hawton K, Kreitman N, et al. Recentclinical and epidemiological trends inparasuicide in Edinburgh and Oxford: a tale oftwo cities. Psychol Med 1988;18:405–18.
108. Rodger CR, Scott AI. Frequent deliberate self-harm: repetition, suicide and cost after threeor more years. Scott Med J 1995;40:10–2.
109. Stenager EN, Jensen K. Attempted suicide andcontact with the primary health authorities.Acta Psychiatr Scand 1994:109–13.
110. Department of Health and Social Security. Themanagement of deliberate self-harm. HN (84):25 1984.
111. NHS Health Advisory Service. Suicideprevention: the challenge confronted. A manualof guidance for the purchasers and providers ofmental health care. London: HMSO, 1994.
112. Royal College of Psychiatrists. The generalhospital management of adult deliberate self-harm. London: Gaskell, 1994.
113. Salter D, Platt S. Suicidal intent, hopelessnessand depression in a parasuicide population:the influence of social desirability and elapsedtime. Br J Clin Psychol 1990;29:361–71.
114. Catalan J, Marsack P, Hawton KE, et al.Comparison of doctors and nurses in theassessment of deliberate self-poisoningpatients. Psychol Med 1980;10:483–91.
115. Newson-Smith JGB, Hirsch SR. A comparisonof social workers and psychiatrists inevaluating parasuicide. Br J Psychiatry1979;134:335–42.
116. Newson-Smith JGB. The use of social workersas alternatives to psychiatrists in assessingparasuicide. in Farmer R, Hirsch S, editors. Thesuicide syndrome: Croome Helm, 1988.
117. Black D, Creed F. Assessment of self-poisoningpatients by psychiatrists and junior medicalstaff. J R Soc Med 1988;81:97–9.
118. O’Dwyer FG, D’Alton A, Pearce JB. Adolescentself-harm patients: audit of assessment in anaccident and emergency department. BMJ1991;303:629–30.
119. Shepherd RM, Dent TH, Alexander GJ, et al.Prevalence of alcohol histories in medical andnursing notes of patients admitted with selfpoisoning. BMJ 1995;311:847.
120. Ebbage J, Farr C, Skinner DV, et al. Thepsychosocial assessment of patientsdischarged from accident and emergencydepartments after deliberate self-poisoning. J RSoc Med 1994;87:515–6.
121. Taylor S. Training and supervision ofdeliberate self-harm assessments. PsychiatrBull 1998;22:510–2.
122. Owens DW. The role of the accident andemergency department in the admission ofdeliberate self-poisoning patients. [MD Thesis].University of Leeds, 1990.
123. Kreitman N. Services for parasuicide and theplace of poisoning within. In: Farmer R, HirschS, editors. The suicide syndrome: CroomeHelm, 1988.
124. Kapur N, House A, Creed F, et al. Managementof deliberate self poisoning in adults in fourteaching hospitals: descriptive study. BMJ1998;316:831–2.
125. Termansen PE, Bywater C. S.A.F.E.R.: a follow-up service for attempted suicide in Vancouver.Can Psychiatr Assoc J 1975;20:29–34.
126. Thomas SH, Bevan L, Bhattacharyya S, et al.Presentation of poisoned patients to accidentand emergency departments in the north ofEngland. Hum Exp Toxicol 1996;15:466–70.
127. Suokas J, Lonnqvist J. Selection of patients whoattempted suicide for psychiatric consultation.Acta Psychiatr Scand 1991;83:179–82.
128. Gunnell D, Brooks J, Peters T. Epidemiologyand patterns of hospital use after parasuicidein the south west of England. J EpidemiolCommunity Health 1996;50:24–9.
129. Hall DJ. A psychiatric liaison service in ageneral hospital referrals and theirappropriateness. Scott Med J 1994;39:141–4.
130. Yeo HM. The cost of treatment of deliberateself-harm. Arch Emerg Med 1993;9:8–14.
131. Butterworth E, O’Grady TJ. Trends in theassessment of cases of deliberate self-harm.Health Trends 1989;21:61.
132. Owens DW, Jones SJ. The accident andemergency department management ofdeliberate self-poisoning. Br J Psychiatry1988;152:830–3.
133. Bagley C. The evaluation of a suicideprevention scheme by an ecological method.Soc Sci Med 1968;2:1-14.
134. Jennings C, Barraclough B, Moss J. Have theSamaritans lowered the suicide rate? acontrolled study. Psychol Med 1978;8:413–22.
135. Ryan J, Clemmett S, Perez-Avila C. Managingpatients with deliberate self harm admitted toan accident and emergency observation ward.J Accid Emerg Med 1996;13:31–3.
136. Ryan J, Rushdy A, Perez-Avila C, et al. Suiciderate following attendance at an accident andemergency department with deliberate selfharm. J Accid Emerg Med 1996;13:101–4.
138. Blake DR, Mitchell JR. Self-poisoning:management of patients in Nottingham, 1976.BMJ 1978;1:1032–5.
139. Linter C. Psychiatric involvement in cases ofdeliberate self-harm. British Journal of Socialand Clinical Psychiatry 1985;3:11–5.
140. Hawton K, James R. General hospital servicesfor attempted suicide patients: a survey in oneRegion. Health Trends 1995;27:18–21.
141. Waterhouse J, Platt S. General hospitaladmission in the management of parasuicide:a randomised controlled trial. Br J Psychiatry1990;156:236–42.
142. Ghodse AH. The attitudes of casualty staff andambulance personnel towards patients who takedrug overdoses. Soc Sci Med 1978;12:341–6.
143. Woodside M. Attempted suicides arriving at ageneral hospital. BMJ 1958:411–4.
144. Patel A. Attitudes towards self-poisoning. BMJ1975;2:426–30.
145. Barber J, Hodgkin G, Patel A, et al. Effect ofteaching on students’ attitudes to self-poisoning. BMJ 1975;2:431–4.
146. Ramon S, Bancroft J, Skrimshire A. Attitudestowards self-poisoning among physicans andnurses in a general hospital. Br J Psychiatry1975;127:257–64.
147. Creed F, Pfeffer J. Attitudes of housephysicians towards self-poisoning patients.Med Educ 1981;15:340–5.
148. Brogan R, Ullyatt K, Pelosi A. Psychiatricservices have limited role in deliberate self-poisoning. BMJ 1998;317: 416.
149. Evans M, Morgan HG, Hayward A. Arandomised controlled trial of the use of crisistelephone consultation for patients admittedafter deliberate self harm. Conferenceproceedings Bristol 1998.
150. D’Zurilla TJ. Problem-solving therapy: a socialcompetence approach to clinical intervention.New York: Springer, 1986.
151. Hes JP, Dor T. Suicide attempt and theemergency room of the general hospital. MentHealth Soc 1978;5:14–22.
152. Malone K, McCormack G, Malone JP. Non-fataldeliberate self-poisoning in Dublin’s northinner city—an overview. Ir Med J1992;85:132–5.
153. Hawton K, Marsack P, Fagg J. The attitudes ofpsychiatrists to deliberate self-poisoning:comparison with physicians and nurses. Br JMed Psychol 1981;54:341–8.
154. Jones DR. A follow-up of self-poisonedpatients. J R Coll Gen Pract 1977;27:717–9.
155. Haynes RB. Strategies to improve compliancefor referrals, appointments and prescribedmedical regimes. In: Haynes RB, Taylor DW,Sackett DL, editors. Compliance in health care.Baltimore: Johns Hopkins University Press, 1979.
156. Van Heeringen C, Jannes S, Buylaert H, et al.The management of non-compliance withreferral to out-patient after care amongattempted suicide patients. Psychol Med1995;25:963–70.
157. Crockett AWB. Patterns of consultation andparasuicide. BMJ 1987;295:476-8.
158. Turner RM. Parasuicide in an urban generalpractice 1970–1979. J R Coll Gen Pract1982;32:273–81.
159. Petrie K. Recent general practice contacts ofhospitalised suicide attempters. N Z Med J1989;102:130–1.
160. Gorman D, Masterton G. General practiceconsultation patterns before and afterintentional overdose: a matched control study.Br J Gen Pract 1990;40:102–5.
161. Turner RJ, Morgan HG. Patterns of health carein non-fatal deliberate self-harm. Psychol Med1979;9:487–92.
162. Rutz W, Von Knorring L, Walinder J.Frequency of suicide on Gotland aftersystematic postgraduate education of generalpractitioners. Acta Psychiat Scandinavica1989;80:151–4.
163. Whyte IM, Dawson AH, Buckley NA, et al. Amodel for the management of self-poisoning.Med J Aust 1997;167:142–6.
164. Salkovskis PM, Atha C, Storer D. Cognitive-behavioural problem solving in the treatmentof patients who repeatedly attempt suicide. Acontrolled trial. Br J Psychiatry 1990;157:871–6.
165. McLeavey BC, Daly RJ, Ludgate JW, et al.Interpersonal problem-solving skills trainingin the treatment of self-poisoning patients.Suicide Life Threat Behav 1994;24:382–94.
166. Chowdhury N, Hicks RC, Kreitman N.Evaluation of an after-care service forparasuicide (attempted suicide) patients.Soc Psychiatry 1973:67–81.
167. Welu TC. A follow-up program for suicideattempters: evaluation of effectiveness.Suicide Life Threat Behav 1977;7:17–20.
168. Cotgrove AJ, Zirinsky L, Black D, et al.Secondary prevention of attempted suicide inadolescence. J Adolesc 1995;18:569–77.
169. Linehan MM, Armstrong HE, Suarez A, et al.Cognitive behavioural treatment ofchronically parasuicidal borderline patients.Arch Gen Psychiatry 1991;48:1060–4.
170. Liberman RP, Eckman T. Behaviour therapy vsinsight-oriented therapy for repeated suicideattempters. Arch Gen Psych 1981;38:1126–30.
171. Torhorst A, Moller HJ, Burk F, et al. Thepsychiatric management of parasuicidepatients: a controlled clinical study comparingdifferent strategies of outpatient treatment.Crisis 1987;8:53–61.
172. Montgomery SA, Montgomery DB, Jayanthi-Rani S, et al. Maintenance therapy in repeatsuicidal behaviour: a placebo controlled trial.In the proceedings of the Tenth InternationalCongress for Suicide Prevention and CrisisIntervention, 1979; Ottawa.
173. Hirsch SR, Walsh C, Draper R. Parasuicide: Areview of treatment interventions. J AffectDisord 1982;4:299–311.
174. Montgomery SA, Roy D, Montgomery DB. Theprevention of recurrent suicidal acts. Br J ClinPharmacol 1983;15:183–5.
175. Torhorst A, Moller HJ, Kurz A, et al.Comparing a three month and a twelve monthoutpatient aftercare programme forparasuicide repeaters. In: Moller HJ, SchmidtkeA, Welz R, editors. Current Issues ofSuicidology. Berlin: Springer-Verlag,1988:419–24.
176. Kreitman N (Ed) Parasuicide. London: JohnWiley & Sons 1977;22.
Subscriptions and enquiries:
Effective Health Care bulletins are published in association with FT Healthcare. The Department of Health funds a limited number of thesebulletins for distribution to decision makers. Subscriptions are available to ensure receipt of a personal copy. 1998 subscription rates, includingpostage, for bulletins in Vol. 4 (6 issues) are: £42/$68 for individuals, £68/$102 for institutions. Individual copies of bulletins from Vols 1–3 areavailable priced £5/$8 and from Vol. 4 priced £9.50/$15. Discounts are available for bulk orders from groups within the NHS in the UK and toother groups at the publishers discretion.
Please address all orders and enquiries regarding subscriptions and individual copies to Subscriptions Department, Financial Times Professional,PO Box 77, Fourth Avenue, Harlow CM19 5BQ (Tel: +44 (0) 1279 623924, Fax: +44 (0) 1279 639609). Cheques should be made payable toFinancial Times Professional Ltd. Claims for issues not received should be made within three months of publication of the issue.
Enquiries concerning the content of this bulletin should be addressed to NHS Centre for Reviews and Dissemination, University of York, YorkYO10 5DD; Fax (01904) 433661 email [email protected]
Copyright NHS Centre for Reviews and Dissemination, 1998. NHS organisations in the UK are encouraged to reproduce sections of thebulletin for their own purposes subject to prior permission from the copyright holder. Apart from fair dealing for the purposes of research orprivate study, or criticism or review, as permitted under the Copyright, Designs and Patents Act, 1988, this publication may only be produced,stored or transmitted, in any form or by any means, with the prior written permission of the copyright holders (NHS Centre for Reviews andDissemination, University of York, York YO10 5DD).
The NHS Centre for Reviews and Dissemination is funded by the NHS Executive and the Health Departments of Scotland, Wales and NorthernIreland; a contribution to the Centre is also made by the University of York. The views expressed in this publication are those of the authors andnot necessarily those of the NHS Executive or the Health Departments of Scotland, Wales or Northern Ireland.
Printed and bound in Great Britain by Latimer Trend & Company Ltd., Plymouth. Printed on acid-free paper. ISSN: 0965-0288
Effective Health Care BulletinsVol. 21. The prevention and treatment of
pressure sores2. Benign prostatic hyperplasia3. Management of cataract4. Preventing falls and subsequent injury
in older people5. Preventing unintentional injuries in
children and young adolescents6. The management of breast cancer7. Total hip replacement8. Hospital volume and health care
outcomes, costs and patient access
Vol. 31. Preventing and reducing the adverse
effects of unintended teenage pregnancies2. The prevention and treatment of obesity3. Mental health promotion in high risk
groups4. Compression therapy for venous leg ulcers5. Management of stable angina6. The management of colorectal cancerVol. 41. Cholesterol and coronary heart disease:
screening and treatment2. Pre-school hearing, speech, language and
vision screening3. Management of lung cancer 4. Cardiac rehabilitation5. Antimicrobial prophylaxis in colorectal
surgery
This bulletin is based on a reviewcarried out by the followingresearchers at the Department ofPsychiatry, University of Leeds.
■ Allan House
■ David Owens
■ Lesley Patchett
The bulletin was written and producedby staff at the NHS Centre for Reviewsand Dissemination, University of York.
Acknowledgements:
Effective Health Care would like toacknowledge the helpfulassistance of the following whocommented on the text:
■ Mark Baker, North Yorkshire Health Authority
■ Simon Balmer, Nuffield Institute for Health
■ Alison Evans, University of Leeds
■ Jenny Firth-Cozens, NHS Executive Northern and Yorkshire
■ David Gunnell, University of Bristol
■ Ian Hammond, Bedford and Shires Health Care Trust
■ Keith Hawton, University of Oxford
■ Anna Higgitt, Department of Health
■ Paul Hodgkin, Centre for Innovation in Primary Care, Sheffield
■ Rob Kehoe, Airedale NHS Trust
■ Glyn Lewis, University of Wales College of Medicine
■ Paul Marshall, Department of Health
■ Colin Pollock, Wakefield Health Authority
■ Anne Richardson, Department of Health
■ Deborah Sharp, University of Bristol
■ Colin Waine, Sunderland Health Authority
The Effective Health Care bulletins arebased on sys tematic rev iew andsynthesis of research on the clinicaleffectiveness, cost-effectiveness andacceptabili ty of health serviceinterventions. This is carried out by aresearch team using es tablishedmethodological guidelines, with advicefrom expert consultants for each topic.Great care is taken to ensure that thework, and the conclusions reached,fairly and accurately summarise theresearch findings. The University ofYork accepts no responsibility for anyconsequent damage arising from theuse of Effective Health Care.
DECEMBER 199812 EFFECTIVE HEALTH CARE Deliberate self-harm
