Effect of telephone contact on further suicide attempts in patients discharged from an emergency department: randomised controlled study

BMJ 2006; 332 doi: http://dx.doi.org/10.1136/bmj.332.7552.1241 (Published 25 May 2006)
Cite this as: BMJ 2006;332:1241
  1. Guillaume Vaiva, psychiatrist (gvaiva{at}chru-lille.fr)1,
  2. Guillaume Vaiva, psychiatrist2,
  3. François Ducrocq, psychiatrist1,
  4. Philippe Meyer, psychiatrist4,
  5. Daniel Mathieu, professor of crisis care medicine1,
  6. Alain Philippe, researcher in sociodemography3,
  7. Christian Libersa, professor of pharmacology1,
  8. Michel Goudemand, professor of psychiatry1
  1. 1 University Hospital of Lille, School of Medicine, France
  2. 2 University Hospital of Lille, School of Medicine, 6 rue du Professeur Laguesse, 59037 Lille cedex, France; Institut National de la Santé Et de la Recherche Médicale (INSERM U513), School of Medicine, Créteil, France
  3. 3 INSERM U513, School of Medicine, Créteil, France
  4. 4 General Hospital of Valenciennes, France
  1. Correspondence to: G Vaiva
  • Accepted 8 March 2006

Abstract

Objective To determine the effects over one year of contacting patients by telephone one month or three months after being discharged from an emergency department for deliberate self poisoning compared with usual treatment.

Design Multicentre, randomised controlled trial.

Setting 13 emergency departments in the north of France.

Participants 605 people discharged from an emergency department after attempted suicide by deliberate self poisoning.

Intervention The intervention consisted of contacting patients by telephone at one month or three months after discharge from an emergency department for attempted suicide to evaluate the success of recommended treatment or to adjust treatment. Control patients received treatment as usual, in most cases referral back to their general practitioner.

Main outcome measures The primary outcome measures were proportion of participants who reattempted suicide, number of deaths by suicide, and losses to follow-up at 13 months' follow-up. Secondary outcome measures were types and number of contacts with health care.

Results On an intention to treat basis, the three groups did not differ significantly for further suicide attempts, deaths by suicide, or losses to follow-up: contact at one month (intervention 23% (34/147) v controls 30% (93/312), difference 7%, 95% confidence interval − 2% to 15%), three months (25% (36/146) v 30%, difference 5%, − 4% to 14%). Participants contacted at one month were less likely at follow-up to report having reattempted suicide (12% v 22% in control group, difference 10%, 2% to 18%).

Conclusion Contacting people by telephone one month after being discharged from an emergency department for deliberate self poisoning may help reduce the number of reattempted suicides over one year.

Introduction

People who have attempted suicide are at high risk of further attempts (12%-30%) or completing suicide (1%-3%), within a year.18 Most patients discharged directly from emergency care are considered at the lowest short term psychopathological risk or best looked after by their family. When patients are referred to community services, however, they are not assessed for compliance with treatment or treatment success.

A review of controlled studies of treatment strategies for people who have attempted suicide found few methods that significantly reduced the repetition of suicidal behaviour.9 10 A non-significant lower rate of repeated self harm was found in people who had attempted suicide for the first time and were offered easy access to a trained psychiatrist.11 A controlled study that assessed the usefulness of telephone contact at four and eight months after suicide attempts, in addition to usual treatment, found no significant difference between the intervention and control groups in number of further suicide attempts at one year.12 The researchers concluded that the method was nevertheless useful because it offered patients who had never received psychiatric care before their index suicide attempt the chance of contact with health professionals.

We systematically contacted patients by telephone one or three months after an attempted suicide by deliberate self poisoning to determine the effect on reattempted suicide over one year. The intervention also included an attempt to enhance compliance with treatment and brief crisis intervention if needed.

Participants and methods

Our study was carried out in 13 emergency departments in the north of France. Eligible patients were those aged between 18 and 65 years who had attempted suicide by drug overdose, had been examined by a psychiatrist who agreed to their discharge from the emergency department, could give the name of their general practitioner, could be contacted by telephone, and gave written consent for being contacted. We excluded homeless patients and those addicted to illegal drugs.

Study design

We randomised one group of participants to telephone contact at one month after a suicide attempt and the other to contact at three months (figure). We compared the intervention groups with a control group that consisted of usual treatment (no telephone contact) during the 12 months of follow-up. For every patient in each of the intervention groups we randomised one patient to the control group.

Progress of participants through trial

Randomisation

In each department patients were allocated to a group according to the number in an opaque sealed envelope. The allocation sequence was provided by a statistician uninvolved in the assessment of patients. Patients were randomised to either telephone contact or usual treatment in blocks of eight (2:2:4 per block) on the basis of a computer generated list of pseudorandom numbers. We used two strata for the randomisation process: one for patients who had attempted fewer than four suicides in the past three years and one for those who had attempted more than four suicides in the past three years. For each stratum the patients were assigned by random allocation to telephone contact at either one month or three months. No change was possible after allocation, this being verified by the staff responsible for randomisation. The allocation list was stored in tamper proof envelopes in a locked cabinet, accessible only to authorised staff.

Sample size

To produce significant differences for change in suicidal behaviour after dialectic cognitive therapy, Linehan et al needed a population of 44 participants.13 More recently, using a similar method to ours, Guthrie et al reached statistical significance with 95 participants.9 Compared with these studies, the psychotherapy involved in our study was less specific, but our telephone interventions took place over a shorter period at defined times. To reach statistical significance we required a population four times larger than that of Guthrie et al (n = 400). Assuming that 30% of the participants would not be contactable by telephone and that 25% would be lost to follow-up, we included 600 people in our study.

Baseline assessments

At the baseline assessment data were collected on personal details; stay in the emergency department; details of the suicidal crisis; and mental disorders using the minimal international neuropsychiatric interview for Diagnostic and Statistical Manual of Mental Disorders, fourth edition, the revised symptom checklist 90, and Beck's hopelessness scale.1416

Interventions

Psychiatrists with at least five years' experience in managing suicidal crises telephoned the participants. None of the psychiatrists had met the participants. Telephone contact was abandoned if unsuccessful after three attempts on three different days and at two different times (midday or evening).

The principle of contacting participants was to go back over the treatment recommended in the emergency department: if treatment was difficult to follow a new one was suggested or if patients were considered at high risk of suicide an urgent appointment was made at the emergency department in which they had originally been treated. The psychotherapeutic approach used was psychological support, mainly based on empathy, reassurance, explanation, and suggestion. The experimental intervention therefore included an attempt to enhance compliance with treatment and to provide brief crisis intervention when needed.

The participants' general practitioners were given details of the telephone contact and its conclusions. If participants could not be contacted, their general practitioners were informed by letter.

Outcome data assessments

A specially trained research psychologist, blind to allocation group, assessed the outcome, by telephone.

Participants self reported any further suicide attempts, but these were recorded only if they met a standardised definition.7 Throughout the study period a clinical research assistant visited the 13 emergency departments to validate the data by checking the records on all suicide attempts, deaths, or further suicide attempts.

The assessment included information about the use of health services based on a well established method for recording data on health economics.17 If participants could not be contacted for the final assessment we telephoned their general practitioner, checked the attendance files and medical records of the emergency departments, and sent a letter to the registrar general's office to check if the participant had died.18

Data analysis

We analysed the data using SAS version 8.2. Firstly, we compared the groups on an intention to treat basis, including all participants, regardless of whether their assigned follow-up had taken place. We considered as adverse outcomes any further suicide attempts, deaths by suicide, and losses to follow-up. Secondly, we analysed the data of participants successfully contacted at one month and at three months, to assess the effect of telephone contact on further suicide attempts. We compared normally distributed variables using the t test and we compared qualitative data using Fisher's exact test.

Results

Thirteen emergency departments in the north of France recruited patients over a six month period, during the day but not at weekends. Overall, 605 patients were included. The mean age of participants did not differ from that of non-participants. The sex ratio of participants was three men for every seven women compared with four men for every six women for non-participants (P = 0.0001, χ2 = 47, df = 1).

Attrition

Overall, 70% of participants in both of the intervention groups were successfully contacted by telephone. Compared with the contacted participants, fewer of the non-contactable ones were depressed (35% (31/89) v 48% (98/204); P = 0.05, χ2 = 3.8, df = 1) or had a somatisation disorder (13% (n = 11) v 24% (n = 49); P = 0.03, χ2 = 4.8, df = 1). No difference was found for age, sex, personal factors, number of previous suicide attempts, or severity of psychopathological disorders.

At the end of the 13 month follow-up period we assessed all the included participants, regardless of whether their assigned telephone intervention had taken place. Fifty seven were lost to follow-up (9%). These were more often men (38% (22/57) v 27% (147/545); P = 0.07), had arrived at the emergency department on their own (59% (n = 34) v 42% (n = 229); P = 0.01), were more likely to be drunk (53% (n = 30) v 40% (n = 218); P = 0.05), and had fewer social phobias (2% (n = 1) v 13% (n = 71); P = 0.02).

Six participants died: two among the intervention groups (one a possible suicide from a road traffic incident and the other cancer), two in the control group (both by suicide), and two who were lost to follow-up (one from suicide and one from an ischaemic vascular attack).

The randomised groups did not differ for personal characteristics, psychopathological assessments, previous suicide attempts, or ongoing treatment in the emergency departments (table 1).

View this table:
Table 1

Baseline characteristics of participants assigned to telephone contact at one month or three months after attempted suicide or to usual treatment (controls). Values are numbers (percentages) unless stated otherwise

Effect of contact by telephone

On an intention to treat basis, the three groups did not differ significantly for proportion with an adverse outcome (P = 0.25, table 2): contact at one month (one death (cancer), 23% (34/147) v 30% (93/312) in control group, difference 7%, 95% confidence interval − 2% to 15%); contact at three months (one death from suicide, 5% (36/146) v 30% in control group, difference 5%, − 4% to 14%); comparison between contact at one month and contact at three months (23% v 25%, difference 2%, − 11% to 8%).

View this table:
Table 2

Results of intention to treat analysis in participants assigned to telephone contact one month or three months after attempted suicide or to usual treatment (controls). Values are numbers (percentages) unless stated otherwise

The number of participants contacted at one month who reattempted suicide was significantly lower than that of controls (12% (13/107) v 22% (62/280); χ2 = 4.7, df = 1, P = 0.03, difference 10%, 95% confidence interval 2% to 18%). This difference was seen over the first six months after telephone contact. No deaths from suicide occurred in this group.

For participants contacted at three months, the number who attempted further suicide was not significantly lower than that of controls (17% (16/95) v 22%; P = 0.27, difference 5%, − 4% to 14%).

Attendance for treatment

Participants in the interventions groups talked about their attempted suicide with their general practitioner more often than the controls (table 3).

View this table:
Table 3

Use of healthcare resources during 13 months' follow-up in participants contacted by telephone one month or three months after attempted suicide

Of the 107 participants contacted at one month, 72 were ordinary calls (lasting 5-10 minutes), 22 concerned crisis intervention (15-45 minutes), and 13 detected participants at high risk of suicide. Seven of the 72 participants who seemed alright at the time of contact attempted suicide during the following year. Of the 22 participants who required crisis intervention, five attempted suicide within a year. Thirteen participants were sent to the emergency department; 10 were considered by the psychiatrist as being at risk and eight of these were admitted to hospital. Only one of these 13 patients reattempted suicide, six months later.

Discussion

Contacting people by telephone one month after attempted suicide by deliberate self poisoning may help reduce the proportion of people who reattempt suicide. Telephone contact also enables the detection of people at high risk of further suicide attempts and timely referral for emergency care.

In our study, the proportion of participants who reattempted suicide in the control group, which received treatment as usual, is similar to the data in the literature.3 10 19 A systematic review noted no deaths or morbidity at one year in participants of eight of 10 studies; the remaining two studies showed a brief reduction in further suicide attempts.13 20 21 Recently, Guthrie et al assessed the efficacy of a series of psychoanalytically based consultations at the patient's home during the month after a suicide attempt.9 This study focused on patients who were not admitted to hospital after their suicide attempt, a population similar to ours. Although the reduction in number of further suicide attempts seems greater in Guthrie et al's study, the telephone contact used was easier and cheaper to set up than in our study.

We made no attempt to control for the non-specific effects of psychotherapy, since ours was a pragmatic study. We compared a specific intervention with treatment as usual in France for patients who harm themselves.

Three quarters of the eligible patients agreed to participate. Our recruitment rate was much higher than in previous studies, showing that the patients were more open to telephone contact than to an appointment in the psychiatric clinic.9

One limitation of our study was that 48 of the attempted suicides occurred before the telephone contact at one month. One option would have been to test the effect of earlier contact, between days 15 and 21 after the first suicide attempt, to see if a larger number of reattempted suicides could have been avoided.

Another limitation of our study was the number of patients who could not be contacted by telephone. We believe this to be the main reason for the negative result of the intention to treat analysis. In our study the decision to abandon telephone contact was based on the strict criterion of three unsuccessful attempts. When the final assessment was made, this criterion was lifted and so we were able to include the maximum number of patients possible. We therefore recommend that the number of telephone contacts should be unlimited, to ensure maximum contact with patients.

What is already known on this topic

Up to 15% of patients who attempt suicide by deliberate self poisoning eventually commit suicide

Few interventions of proved efficacy exist for these patients

The interventions are expensive and never concern all people who attempt suicide

What this study adds

Telephone contact of people one month after attempted suicide may reduce the number who reattempt suicide

Acknowledgments

We thank the following staff of the emergency departments: Sylvie Dupont (Catholic University of Lille), BenoÎt Averland (Douai), Marianne Babe (Roubaix), Claude Meurisse (Valenciennes), Jean Bernard Campagne (Dunkerque), Pierre Valette (Arras), Nadine Ruolt (Lens), Jean Michel Bourez (Tourcoing), Marianne Sandlars (Denain), Yves Fichaux (Seclin), Olivier Brochart (Douai), and Fabienne Saulnier (Lille); the psychiatrists who contacted the participants: Olivier Cottencin, Christian Bondu, Christophe Debien, and Virginie Boss; the clinical research assistants who helped with follow-up: Stéphane Duhem and Anne Laure Demarty; the French World Health Organization collaborating centre in mental health (Jean-Luc Roelandt, Lille); and the Groupement d'Etude et de Prevention du Suicide (Jean-Jacques Chavagnat, Poitiers).

Contributors: GV helped design the study, contacted the participants, reviewed the literature, and wrote the paper. He is guarantor. FD contacted the participants and reviewed the literature. PM supervised the conduct of the study and contacted participants. DM supervised the conduct of the study and assessed participants for eligibility. AP helped design the study and carried out the statistical analysis. CL helped design the study, and carried out the statistical analysis. MG helped design the study. All authors reviewed the manuscript.

Footnotes

  • Funding This study was funded by a hospital clinical research grant (PHRC98), a state region contract plan, a subsidy from the regional hospitalization agency,.

  • Competing interests None declared.

  • Ethical approval This study was approved by the ethical committee of the University Hospital Centre of Lille (CCPPRB 98/33).

References

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.

THIS WEEK'S POLL