Dutch Health Care Inspectorate’s Supervision System for Suicides of Mental Health

An Examination of the Dutch Health

Care Inspectorate’s Supervision System

for Suicides of Mental Health Care Users

Annemiek Huisman, M.Sc. Paul B. M. Robben, Ph.D. Ad J. F. M. Kerkhof, Ph.D.

OObjective: This study examined characteristics of suicides among men- tal health care users reported between 1996 and 2006 to the Dutch Health Care Inspectorate and the inspectorate’s follow-up responses. The aims were to determine whether follow-up was associated with par- ticular characteristics and whether the responses could be improved in accordance with guidelines for treatment of suicidal patients. Methods: Information about patient and treatment characteristics was collected from a sample of 505 of the 5,483 suicide notifications between 1996 and 2006. The 1996–2005 sample included an equal number of cases to which the inspectorate did and did not respond. The 2006 sample in- cluded the first 205 notifications in that year. Results: For 2006 notifi- cations the response rate was 37%. The responses most frequently ad- dressed how and whether the suicide was evaluated and the adequacy of treatment for the psychiatric disorder. A follow-up response was more likely when the suicide involved a young patient or a patient treat- ed in a mental health care setting for less than a year or when the noti- fication was accompanied by the mental health institution’s plans for im- proving its policies. A response was less likely when the patient was dis- charged from inpatient care in the three months before the suicide. Since 2002 responses have more frequently emphasized the importance of suicide risk assessment, in accordance with guidelines. Conclusions: The inspectorate might improve its supervision system by placing greater emphasis on addressing suicidal impulses and treating older and chronically suicidal patients and patients soon after inpatient dis- charge. (Psychiatric Services 60:80–85, 2009)

Several studies have shown that suicide. Audits in the United King-

suicide and mental illness are

closely linked (1). Psychiatric patients are a priority group in sever- al national suicide prevention strate- gies in various countries (2,3), but only a small number of studies have examined the clinical care that psy- chiatric patients received before their

dom (4) and Australia (5) found sev- eral treatment-based risk factors for suicide among users of mental health services, including inadequate assess- ment and treatment of psychiatric disorders and psychosocial problems, problems with inpatient observation, and poor continuity of care. About

20% of the suicides of these mental health service users were considered to be preventable.

The Netherlands is one of the few European countries with a continu- ous national supervision and audit procedure for suicides of patients re- ceiving mental health care. The sys- tem has been in operation since 1984. Whenever such a suicide occurs, the therapist responsible for the patient and the medical director must write a notification to the Health Care In- spectorate, which is an independent organization under the Minister of Health, Welfare, and Sport.

The notification must include de- tails of the suicide and the mental health care delivered and an evalua- tion of policies in place for dealing with suicidal patients. The inspector may ask for more information and in some cases may require the health care service to improve the care that is offered to suicidal patients. In gen- eral, the aim of this procedure is not to evaluate individual suicide notifica- tions but to identify structural prob- lems in mental health care services. Some 550 suicide notifications are submitted per year, which account for 36% of all suicides annually in the Netherlands.

The supervision system of the Health Care Inspectorate is designed to improve the quality of care for sui- cidal patients and ultimately to pre- vent suicide. However, its effective- ness has never been evaluated. The study presented here is a preliminary step in this evaluation. The aim was to describe the management of sui-

Ms. Huisman and Dr. Kerkhof are affiliated with the Department of Clinical Psychology, VU University, Amsterdam, Van der Boechorststraat 1, Amsterdam 1081BT, Netherlands (e-mail: a.huisman@psy.vu.nl). Dr. Robben is with the Health Care Inspectorate, Utrecht, Netherlands.

80

PPSYCHIATRIC SERVICES ' ps.psychiatryonline.org ' January 2009 Vol. 60 No. 1

cide notifications by the inspectorate and to compare its responses with re- cent guidelines for suicide preven- tion (6) to determine whether super- vision could be improved. In addi- tion, changes in the manner in which the inspectorate has responded to sui- cide notifications over time was ex- amined. The results will be used in further studies to assess the impact of the inspectorate’s supervision.

Methods

Suicide files were made available by the Health Care Inspectorate for the period 1996–2006. All suicide notifi- cations from this period were identi- fied (N=5,483), and a total of 505 were selected. For 1996–2000, a total of 100 notifications were selected, and 200 were selected for 2001–2005. For 2006 the first 205 suicide notifi- cations submitted that year were ob- tained. A relatively large number of cases from recent years were exam- ined, because these were considered to be most representative of the cur- rent procedures of the inspectorate. Files from earlier years were studied to gain insight into historical develop- ments in the management of suicide notifications.

For 1996–2005 an equal number of suicide notifications with and without a response from the inspectorate were randomly selected. A response was defined as further questions, re- marks, or suggestions by the inspec- tor after the initial notification or a personal conversation between the inspector and the person who sent the notification. No follow-up was de- fined as a simple letter from the in- spectorate acknowledging receipt of the notification with no further ques- tions or remarks. A total of 227 sui- cide notifications had a response from the inspectorate, and 278 notifica- tions had no follow-up. The selection of notifications was conducted in this manner to permit comparison of cas- es with and without a response and to determine whether inspectors re- sponded more frequently to certain patient or treatment characteristics. A pen-and-paper instrument was used to gather data on relevant characteris- tics, including patients’ demographic characteristics and the responses of the inspectorate.

Responses to suicide notifications by inspectors were examined both quantitatively and qualitatively. The nature of the response was classified into three categories: a request for ad- ditional information, remarks or sug- gestions for improvement, and further contact with the mental health service or other involved services.

All responses were also subjected to a detailed qualitative analysis, facil- itated by ATLAS.ti software. An open coding scheme was derived from the questions and remarks of the inspec- tors, and every response was as- signed a preliminary code independ- ently by the first two authors. The codes were further refined, and a clear definition was generated for each until a comprehensive coding scheme was created that accurately reflected the responses. By use of this coding scheme, each response was reviewed independently by the second author, and inconsistencies in coding were discussed until agree- ment was reached.

The responses of the inspectorate to suicide notifications were then compared with the American Psy- chiatric Association’s (APA’s) Prac- tice Guideline for the Assessment and Treatment of Patients With Sui- cidal Behaviors (6) to establish whether the responses were in ac- cord with the guideline. The guide- lines note that the most important aspects are frequent suicide risk as- sessments on the basis of protective and risk factors, treatment planning to reduce suicide risk, continuity of care, and a restrained use of no-sui- cide contracts.

The relationship between charac- teristics of the suicide notifications and the likelihood of a response by the inspectorate was examined by cross-classifying whether or not the inspectorate responded on the basis of patient and treatment variables (age, gender, diagnosis, suicide meth- od, inpatient versus outpatient sta- tus, discharge from inpatient care, duration of treatment, warning sig- nals of suicide, discussion of suicidal- ity with the therapist, and lessons learned as a result of the suicide). Chi square tests were computed on that distribution, and the signifi- cance threshold was set at .01 to

compensate for the possibility of finding significance by chance when conducting such a large number of comparisons.

To determine whether the manage- ment of suicide notifications by the inspectorate had changed between 1996 and 2006, responses from re- cent years (2002–2006) were com- pared with those from an earlier peri- od (1996–2001) by using chi square tests.

Results

Patient characteristics

Demographic and clinical characteris- tics of the sample are summarized in Table 1. A typical patient was a mid- dle-aged male in outpatient treatment who had a diagnosis of depression.

The number of responses to sui- cide notifications in each of the three categories is shown in Table 2. In- spectorate responses for 227 of the 505 suicide notifications were exam- ined. The inspectorate responded to 75 of the 205 notifications (37%) in 2006.

Qualitative analysis of responses

The inspectorate’s responses to the suicide notifications were classified into 13 broad categories (Table 3). The greatest number of questions or remarks concerned evaluation of the care provided to the patient. The most common question in this re- spect was, “Has the suicide been evaluated and what were the results of the evaluation?” Other frequently asked questions involved whether the patient received adequate treat- ment for his or her psychiatric disor- der. Questions also addressed the na- ture, purpose, and progress of the treatment and the clinician’s deci- sions about appropriate treatment settings.

The importance of adherence to treatment guidelines was stressed by the inspectorate in 36% of the re- sponses. In the Netherlands there are national guidelines for the treat- ment of various psychiatric disorders but not for suicidality. The responses often contained questions about whether the clinician adhered to the treatment guidelines for the psychi- atric disorder in question. Another question addressed whether guide-

PSYCHIATRIC SERVICES ' ps.psychiatryonline.org ' January 2009 Vol. 60 No. 1

81

Table 1

Characteristics of 505 users of mental health services who died by suicide in the Netherlands, 1996–2006

Characteristic N %

Gender Male 280 55 Female 225 45

Age 15–20 11 2 21–30 58 12 31–40 113 22 41–50 134 27 51–60 95 19 >60 92 18

DSM diagnosis Primary axis I diagnosis

Depressive disorder 218 43 Schizophrenia or other

psychotic disorder 141 28 Bipolar disorder 36 7 Substance use disorder 41 8

Alcohol 22 4 Drug 41

Both Anxiety disorder Other disorder

Comorbid secondary diag- nosis of a substance use disorder

Alcohol Drug Both

Treatment status Inpatient

Outpatient

15 3 22 4 47 9

81 16 35 7 32 6 14 3

154 30 351 70

lines for the treatment of suicidal pa- tients were in place at the agency, and if so, whether the clinician ad- hered to them. In the case of re- sponses to notifications of suicides by inpatients, questions were asked about policies on safety, patient priv- ileges, and monitoring.

In 29% of the responses the ques- tions or remarks addressed collabo- ration with other practitioners or services, and in 15% continuity of care was a focus. These questions and remarks were made most fre- quently in responses that involved suicide among patients who had been recently discharged from inpa- tient care or who had changed treat- ment setting. The responses ad- dressed transfer of information and consultation between therapists or services involved in the patient’s care and the frequency of aftercare ap- pointments.

dressed in 27% of the 227 responses. Most common were questions about whether and how risk assessment took place and whether suicidality was discussed periodically with the patient. Specific remarks addressed the importance of telling the patient about an elevated suicide risk in the first weeks of using antidepressant medication, taking the expression of suicidal ideation or behavior serious- ly, and communicating about suicide risk with other therapists involved in the patient’s care.

Questions about medication (in- cluded in 27% of responses) and about psychiatric assessment (18%) were usually straightforward: what medication was prescribed, or what was the DSM psychiatric diagnosis? Other questions involved the assess- ment of the primary psychiatric disor- der and any comorbid disorders, the accuracy of diagnoses, and the appro- priateness of the medication.

In 14% of the responses the ques- tions and remarks specifically re- ferred to management of the pa- tient’s suicidal impulses. Generally, this involved a question about whether and how the therapist had managed this risk.

Fifteen percent of responses in- cluded questions or remarks about the involvement of the patient’s fam- ily, and 7% addressed aftercare for the bereaved relatives. The inspec- torate stressed that mental health services must involve the patient’s family in the assessment and treat- ment of suicidal patients and must offer aftercare to the bereaved fami- ly after a suicide.

The significance of the role of the psychiatrist was emphasized in 12% of the notifications. In some cases the patient had not been seen by a psy- chiatrist. The inspectorate took the

position that a psychiatrist must exer- cise responsibility and see patients personally, especially in assessment of psychiatric disorders and suicide risk and prescription of psychotropic drugs.

Seven percent of responses con- cerned the patient’s treatment non- compliance and issues regarding in- voluntary hospitalization. In these cases, the patient usually refused mental health care or regularly missed appointments, to which the therapist did not take an active ap- proach. The inspectorate recom- mended in these cases that noncom- pliant patients must be approached more actively by mental health servic- es. Other responses included a ques- tion about whether involuntarily hos- pitalizing the patient was considered and whether it would have been bet- ter to have done so.

Critical remarks and suggestions for improvement The inspectors made critical re- marks in 106 notifications. They ad- dressed the lack of guidelines for suicide prevention, the lack of suffi- cient continuity of care and collabo- ration between therapists involved, insufficient involvement of a psychi- atrist in suicide risk assessment and prescription of medication, inade- quate assessment of suicide risk, in- adequate psychiatric treatment and inaccurate psychiatric diagnosis, and insufficient attention to communica- tion and signals from relatives of the patient.

Characteristics associated with inspectorate responses Characteristics of the suicide notifica- tions to which inspectors responded more or less frequently are summa- rized in Table 4.

Suicide risk assessments were ad-

Table 2

Categories of 227 inspectorate responses to notifications of mental health service

users who died by suicide in the Netherlands, 1996–2006

Category N %

Additional information requested Remarks or suggestions for improvement Contact or discussion with the therapist, medical

104 21 106 21

director, or services involved

17 3

PSYCHIATRIC SERVICES ' ps.psychiatryonline.org ' January 2009 Vol. 60 No. 1

82

Changes in responses 1996–2006

In the 2002–2006 period inspectors were significantly more likely to em- phasize the importance of suicide risk assessment than in the 1996– 2001 period (37% and 19%; χ2=6.4, df=1, p= .01). In addition, the con- tent of responses about risk assess- ment appears to have changed. In earlier years questions were simpler and mainly addressed whether risk was assessed. In the later period questions were more elaborate and more often required a detailed as- sessment on the basis of risk factors for suicide.

For all other variables (listed in Table 3) no differences were found.

Correlation between responses and APA guidelines In general, the inspectorate’s 1996– 2006 responses were in line with APA guidelines. Important aspects of the responses were adequate psychiatric treatment, cooperation with other therapists involved in the patient’s care, continuity of care, and provision of aftercare for the bereaved family. In addition, responses about suicide risk assessment corresponded in- creasingly with the guidelines in the most recent years examined.

However, use of no-suicide con- tracts was addressed only once in all 227 of the inspectorate’s responses to suicide notifications, although 23% of the notifications indicated that a no- suicide agreement was arranged with

the patient, including patients who had a diagnosis of a psychotic disor- der, highly impulsive patients, and those with serious addiction.

Suicide notifications without follow-up The 278 suicide notifications without follow-up by the inspectorate were studied qualitatively to gain more in- sight into possible reasons for a lack of follow-up and to determine whether there were, despite the lack of follow-up, indications of structural problems in the mental health care provided. From the perspective of the APA guidelines, several possible signs of shortcomings in the mental

health care provided were observed and are discussed below.

Incomplete or inadequate risk as- sessment. In 59 notifications without follow-up (21%), therapists underes- timated the risk of suicide despite the presence of several risk factors. Previ- ous suicide attempts were labeled as “merely a cry for help” in nine notifi- cations, and thus the suicide risk was estimated to be low. In addition, in nine cases mental health care workers were unaware of the suicidal history of the patient or knew nothing about suicidal intent expressed by the pa- tient to family members or fellow pa- tients by the patient.

Another problem in the area of

Table 3

Issues addressed in 227 inspectorate responses to suicide notifications involving

suicides of mental health service users in the Netherlands, 1996–2006a

Issue N %

Evaluation of the suicide Treatment of the psychiatric disorder Treatment guidelines Collaboration with other practitioners or services Suicide risk assessment Medication Psychiatric assessment Continuity of care Involvement of the patient’s family in treatment Treatment of suicidality Role of the psychiatrist Aftercare for relatives Noncompliance and involuntary hospitalization

135 60 86 38 82 36 66 29 62 27 61 27 40 18 33 15 33 15 32 14 27 12 15 7 15 7

a Many responses addressed more than one issue.

Table 4

Characteristics of suicide notifications involving 505 mental health service users in the Netherlands, 1996–2006, by

whether or not the characteristic was present and whether the inspectorate responded

Characteristic present

N of Characteristic responses %

Characteristic absent

N of responses % χ2a p

Patient was less than 35 years old Patient was treated in mental health care for less than a year Patient was still in an initial registration procedure Fellow patients had signals of an imminent suicide in the

months before the suicide Unclear whether the clinician discussed suicidality with the patient Notification contained plans to improve mental health care

as a result of the suicide Patient was discharged from inpatient care in the 3 months

before the suicide

68 52 159 72 53 155 24 67 203

13 68 214 44 57 183

76 59 151 54 38 173

43 3.46 .06 42 4.39 .04 43 7.39 .01

44 4.40 .04 43 5.46 .02

40 14.41 <.01 48 4.52 .03

a df=1

PSYCHIATRIC SERVICES ' ps.psychiatryonline.org ' January 2009 Vol. 60 No. 1

83

poor risk assessment was that atten- tiveness to suicide risk had waned, es- pecially for patients who had a history of severe suicidality but who did not report current suicidal ideation (N=15) and for patients who were chronically suicidal (N=15).

Insufficient continuity and intensity of care. Continuity of care was not al- ways adequate. In 11 notifications without follow-up, the patient com- mitted suicide while on a waiting list for treatment or while involved in a registration procedure that lasted sev- eral months, despite the patient’s se- vere psychiatric symptoms or crisis. In 14 cases that were not followed up by the inspectorate, follow-up appoint- ments after discharge from inpatient care took weeks or months (range of three weeks to three months). In 17 other cases the emergency service did not assess suicide risk in time or did not make an appointment with the pa- tient within a few days of referral or of the patient’s initial contact with servic- es, and the patient committed suicide before being seen.

Unwarranted trust in no-suicide contracts. In 28 notifications without follow-up, a no-suicide contract was arranged with a patient and the ther- apists involved considered the sui- cide risk to be reduced. In seven cas- es the patient’s willingness to enter a contract was sufficient to result in transfer to an open ward. Moreover, in five cases arrangement of a no- suicide contract seemed to be the only safety measure taken; other measures, such as more intensive care or a safety plan, were not car- ried out.

Inadequate decisions about hospi- talization. In 14 cases without follow- up, patients in crisis weren’t hospital- ized because hospitalization was thought to be risk enhancing, presum- ably because these patients had a per- sonality disorder. In addition, seven patients committed suicide while on a waiting list for inpatient admission.

Inadequate communication. Inad- equate communication between mental health care workers, especial- ly about suicidality, may have led to insufficient transfer of information and suicide risk management in 19 notifications for which no follow-up was received.

Insufficient monitoring of severely depressed or psychotic patients. For six notifications that involved suicide of a hospitalized patient and for which no follow-up was received, the patient was able to run away from a closed ward on repeated occasions.

Inadequate communication with the patient’s family. In 16 cases with- out follow-up, the patient’s relatives were either unable to discuss their concerns about the suicidality of their relative with the therapists or they were not involved in treatment de- spite the patient’s severe suicidality.

Discussion

This study was undertaken as a first step in a research program to evaluate the suicide notification procedure ad- ministered by the Health Care In- spectorate in the Netherlands. The results show that in 2006 approxi- mately 37% of all mental health work- ers who reported a suicide received further questions or remarks from the inspectorate. Inspectors’ responses were mostly focused on the thorough evaluation of circumstances and care surrounding the suicide. Another main point of interest to the inspec- torate was the treatment of psychi- atric disorders in accordance with treatment guidelines. Compared with responses to suicide notifications be- tween 1996 and 2001, recent respons- es have more often stressed the im- portance of conducting suicide risk assessment, which is in line with APA guidelines.

Certain aspects of the notifications led to more or less frequent respons- es. Inspectors’ responses depended on the treatment status of the patient who died by suicide and tended to depend on the patient’s age and time in treatment. The proportion of re- sponses was larger for patients who were young or at the beginning of treatment, and it was smaller for pa- tients who were recently discharged from inpatient care. These findings suggest that the inspectorate focuses especially on patient groups and time periods for which suicide pre- vention efforts are considered most effective. Apparently inspectors be- lieved that there were few opportu- nities for prevention among elderly persons, those with chronic illnesses,

and those in the postdischarge peri- od, although patients in the postdis- charge period are widely recognized to be at high risk of suicide (7). There may be opportunities for the inspectorate to emphasize more ef- fective suicide prevention in the postdischarge period.

Inspectors tended to pay special attention to suicides in which fellow patients had noticed signals of an im- minent suicide in the months before and when it was unclear whether the clinician had discussed suicidality with the patient or whether the pa- tient had been treated as suicidal. These aspects were apparently re- garded as important considerations for suicide prevention. Moreover, these findings may demonstrate the gradually growing awareness in the field and within the inspectorate that suicidal impulses need specific at- tention in addition to the usual treat- ment for psychiatric disorders. The inspectorate could further promote such awareness, as recommended in the APA guidelines.

A notable result is that only one of the inspectorate’s 227 responses ad- dressed use of no-suicide contracts, although such contracts were used in about one in five of the cases re- viewed by the inspectorate. Con- tracts were made with patients who had addictive or psychotic disorders or who were highly impulsive, which is discouraged by APA guidelines for the treatment of suicidal patients (6).

The inspectorate was more likely to respond to a suicide notification when mental health institutions at- tached plans for improvement to the notification. In its responses the in- spectorate both supported the in- tended improvements and acknowl- edged the flaws in the mental health care delivery that the institutions themselves admitted. However, in some cases the inspectorate did not respond, although the notification contained indications of possible flaws in care delivery. This finding seems to indicate that the inspec- torate neglected to address some shortcomings. Moreover, inspectors did not respond in the same manner to all notifications involving the same themes, which suggests a somewhat arbitrary element.

84

PPSYCHIATRIC SERVICES ' ps.psychiatryonline.org ' January 2009 Vol. 60 No. 1

In general, mental health care providers are concerned about possi- ble disciplinary measures by the in- spectorate; however, the findings of this study show that in cases of suicide notifications, such measures seldom follow. In none of the inspectorate’s 227 responses to suicide notifications were disciplinary measures taken, and only a small percentage (3%) of suicide notifications led to an exten- sive inquiry into the case.

Some limitations should be noted. The results of this study depend on the quality and comprehensiveness of suicide notifications. Additional re- search is in progress to evaluate these aspects of the notifications. The re- sults of the qualitative analyses are based on the authors’ interpretations of whether treatment was consistent with APA guidelines (6) and therefore are not conclusive. In addition, a rel- atively large number of tests were conducted, and it is possible that some associations were found by chance. Replication is needed to con- firm the factors that determine whether the inspectorate responds to a notification.

The notification procedure is

meant to provide supervision of the quality of health care service delivery and to improve care for suicidal pa- tients in the future. As such the in- spectorate’s procedure can be a pow- erful tool in promoting suicide pre- References vention. Further research is in

Acknowledgments and disclosures

This research was funded by the Health Care Inspectorate in the Netherlands.

The authors report no competing interests.

progress to examine the influence of the suicide notification procedure on the quality of care in mental health services and to examine how mental health services view the notification procedure.

Conclusions

The results show that supervision in mental health care can be optimized in accordance with guidelines for the treatment of suicidal patients. The inspectorate might enhance its review procedure by more consis- tent supervision, continuing empha- sis on systematic suicide risk assess- ment, more emphasis on the specif- ic treatment of suicidal impulses, more attention to the treatment of older patients who are chronically suicidal and to patients newly dis- charged from inpatient care, and more focus on a restrained use of no-suicide contracts.

1. Bertolote JM, Fleischmann A, De Leo D, et al: Psychiatric diagnoses and suicide: revisit- ing the evidence. Crisis 25:147–155, 2004

2. National Suicide Prevention Strategy for Eng- land. London, Department of Health, 2002

3. National Strategy for Suicide Prevention: Goals and Objectives for Action. Rockville, Md, US Public Health Service, 2001

4. Appleby L, Shaw J, Amos T, et al: Suicide within 12 months of contact with mental health services: national clinical survey. British Medical Journal 318:1235–1239, 1999

5. Burgess P, Pirkis J, Morton J, et al: Lessons from a comprehensive clinical audit of users of psychiatric services who commit- ted suicide. Psychiatric Services 51:1555– 1560, 2000

6. American Psychiatric Association: Practice guideline for the assessment and treatment of patients with suicidal behaviors. Ameri- can Journal of Psychiatry 160(Nov suppl): 1–60, 2003

7. Huisman A, Kerkhof AJFM, Robben PBM: Guidelines for the treatment of suicidal pa- tients: an overview [in Dutch], in Jaarboek Voor Psychiatrie en Psychotherapie 2007– 2008, 10th ed. Edited by Schene AH, Boer F, Jaspers JPC, et al: Houten, Netherlands, Bohn Stafleu van Loghum, 2007

PSYCHIATRIC SERVICES ' ps.psychiatryonline.org ' January 2009 Vol. 60 No. 1

85