Prevention of suicide and attempted suicide in Denmark. Epidemiological studies of suicide and intervention studies in selected risk groups.

The suicide rates in Denmark have been declining during the last two decades. The decline was relatively larger among women than among men. All age groups experienced a decline except the very young with stable rates and the very old with increasing rates. The Universal, Selective, Indicated (USI) model recommended by Institute of Medicine was used as a framework for the thesis. Universal preventive interventions are directed toward the entire population; selective interventions are directed toward individuals who are at greater risk for suicidal behaviour; and indicated preventions are targeted at individuals who have already begun self-destructive behaviour. At the universal level, a review was carried out to highlight the association between availability of methods for suicide and suicide rate. There were mostly studies of firearms, and the conclusion of the review was that there was clear indication of restricted access to lethal means was associated with decline in suicide with that specific method, and in many cases also with overall suicide mortality. Restricting access is especially important for methods with high case fatality rate. Our own study indicated a beneficial effect on suicide rates of restrictions in access to barbiturates, dextropropoxyphen, domestic gas and car exhaust with high content of carbon monoxide. Although a range of other factors in the society might also be of importance, it was concluded that restrictions in access to dangerous means for suicide were likely to play an important role in reducing suicide rates in Denmark, especially for women. At the selective level, there are several important risk groups such as psychiatric patients, persons with alcohol and drug abuse, persons with newly diagnosed severe physical illness, all who previously attempted suicide, and groups of homeless, institutionalized, prisoners and other socially excluded persons. The thesis focused on homeless persons and psychiatric patients, especially patients with schizophrenia and related disorders. The thesis contains a review of the risk of suicide in homeless. In all the studies included, increased suicide mortality was found, and in the studies that evaluated suicide risk in different age groups, the excess suicide mortality was most dominant in younger age groups. Our own study revealed an increased risk of suicide, and in univariate analysis, significant predictors for suicide were found to be associated with shortest stay in hostel less than 11 days and more than one stay during one year. The thesis also contains a review of the risk of suicide in first-episode patients with schizophrenia, and it was concluded on the basis of the identified studies that long-term risk of suicide was not 10 percent as previously accepted, but lower. Risk factors for suicide among patients with schizophrenia were evaluated in case control studies, in nested case control studies, and in prospective studies. The following risk factors were the most important and frequently observed predictors: male gender, young age, short duration of illness, many admissions during last year, current inpatient, short time since discharge, previous and recent suicide attempt, co-morbid depression, drug abuse, poor compliance with medication, poor adherence to treatment, high IQ, and suicidal ideations. The results of analyses of psychotic symptoms as risk factor for suicide were contradictory, but a recent meta-analysis concluded that both hallucinations and delusions seemed to be protective; however, there was a non-significant tendency that command hallucinations were associated with higher suicide risk. Prevention of suicide in schizophrenia must especially focus on improving assessment of risk of suicide during inpatient treatment and the first week after discharge, and special attention must be paid to patients with one or more of the identified risk factors. There is a need for large randomised clinical trials evaluating the effect on suicide and suicide attempt of psychosocial and pharmacological treatment in schizophrenia. In our own study, we did not find any effect of integrated treatment on attempted suicide, but there was an effect on hopelessness and a trend toward lower prevalence of depression among patients in the integrated treatment. There were four suicides and one probable suicide (drowning) in standard treatment and one suicide in integrated treatment at two-year follow-up, but the study did not have sufficient power to detect these differences in proportion to who committed suicide; more than one thousand patients should have been in each treatment condition in order for these differences in proportion to be significant. At the indicated prevention level, a literature review was carried out regarding risk of suicide attempt and suicide in short-term, medium-term and long-term follow-up of persons who attempted suicide. It was concluded that the risk of repetition in short- and medium-term follow-up studies was approximately 16 percent, with lower risk among "first-evers" compared to repeaters. There was a large variation in repetition rate. The proportion who committed suicide in medium-term follow-up studies was 2.8 percent and in long-term follow-up studies was 3.5 percent (weighted mean) with clearly higher proportions in the Nordic studies than in the studies from UK. Risk factors for attempted suicide were previous suicide attempt, alcohol and drug abuse, depression, schizophrenia, previous inpatient treatment, self-discharge before evaluation, sociopathy, unemployment, frequent change of address, hostility, and living alone. Several of the predictors are overlapping and most of them were already identified in early studies of factors predictive of repetition of suicide attempt. Predictors of suicide were male gender, increasing age, previous suicide attempt, serious suicide attempt, alcohol and substance abuse, somatic disease, mental illness, and planning of suicide attempt, high suicidal intent score, violent suicide attempt or suicide attempt with severe lethality, and ongoing or previous psychiatric treatment. In our follow-up study from Bispebjerg Hospital, we found that the risk of suicide during a ten-year follow-up period among patients admitted in 1980 after self-poisoning was 30 times greater than in the general population. We also found increased mortality by all other causes of death. Predictors of suicide were several previous suicide attempts, living alone and increasing age. There are not many randomised clinical trials of psychosocial interventions aiming to reduce risk of repetition among suicide attempters. A Cochrane review concluded that evidence was lacking to indicate the most effective forms of treatment for deliberate self-harm patients. A recent randomised controlled trial showed a positive influence of cognitive behavioural therapy on repetition rate. Our own quasi-experimental study of effectiveness of two weeks' inpatient treatment in a special unit of young persons who had severe suicidal thoughts or who had attempted suicide showed that risk of repetition was reduced in the intervention group, and that the intervention group obtained a significantly greater improvement in Beck's Depression Inventory, Hopelessness Scale, Rosenberg Self-Esteem Scale and CAGE-score. The study of emergency outreach indicates that there are many persons in the community that experience a suicidal crisis, and that this group is an important target group for psychiatric emergency outreach. In our study of registration and referral practice in Copenhagen Hospital Cooperation, we conclude that not all suicide attempts were registered as such in the National Patient Register - in fact, only 37 percent. It must be concluded that the quality of the Danish Patient Register must be improved with regard to registration of suicide attempt. We found that psychiatric evaluation was planned in relation to almost all suicide attempts, but that it must be recommended to pay attention to escorting patients to psychiatric emergency in order to ensure that the patient actually attends the planned consultation. We found that patients who were referred after psychiatric evaluation to psychiatric treatment at outpatient facilities only received the planned treatment in approximately two-thirds of the cases; therefore, like Hawton et al. [Hawton et al., 1998; Hawton et al., 1999], we recommend that outpatient facilities adopt an assertive approach to patients who have attempted suicide. Danish suicide research is strong, primarily due to the possibilities for linking complete national registers providing detailed data and large sample sizes for suicide research, which is so far unique for the Nordic countries. This, combined with skilful use of epidemiological methods, had resulted in a remarkable series of papers highlighting risk of suicide in different risk groups, risk factors and protective factors. This activity must continue. In this work it is important to be aware of limitations in naturalistic studies such as the risk of interchanging cause and effect and the necessity to carry out control for confounders. Meta-analysis is a strong tool for summing up results of previous research. Meta-analyses can be used in reporting the evidence for effectiveness of interventions, but also for determining risk or identifying risk factors. A meta-analysis of risk factors of repetition of suicide attempt has not been carried out, and the quality of the identified studies did not allow a formal meta-analysis. Large randomised clinical trials examining the effectiveness of interventions on reducing rate of suicide attempt and suicide should have high priority. Suicide is a major public health problem and should be given high priority with regard to prevention and research. (ABSTRACT TRUNCATED)