The main findings of the present study were that many patients reported that no follow-up was offered, even though the hospital records show that they were registered with follow-up assessment. The satisfaction with treatment in the hospital and during follow-up was generally good. However, a significant proportion of the patients exhibited repeated self-harming behavior during the 3 months after discharge from the hospital and reported a considerably high level of psychiatric problems.
Many patients that had a hospital record of referral to follow-up reported that they had not been offered this assessment. These findings are in line with those of a Dutch study, in which 35% of suicide attempters did not remember whether aftercare arrangements had been made during their hospital stay when asked 7 days after discharge . The fact that almost one-third of patients had to wait 3 weeks for their first appointment is one possible explanation for this result. During an acute crisis, patients may perceive such a long waiting time as no follow-up. Conversely, several patients who were definitely transferred directly to a psychiatric ward also reported absence of follow-up. Information about the referral may not have been understood or remembered by the patient, they may not have received the message in the hospital or written confirmation of their appointment.
Although the patients included in this study were rather satisfied with the treatment they received both in the hospital and after discharge, in the psychiatric outpatient clinic, they were less satisfied with the plans for follow-up and with the time from discharge to the first appointment. Taylor and coworkers reviewed attitudes towards clinical services among people engaging in self-harm. They found similar results, as the participants in their study also pointed out the need to improve access to care after discharge .
To date, there is no clear evidence of interventions or treatments that are effective after self-poisoning or injury . However, because of the diversity and complexity of psychiatric problems, and their somatic and social character, it is important to provide coordinated, close, and systematic follow-up. Furthermore, a considerable number of patients drop out of treatment. Chain of care and early intervention after suicide attempts have yielded promising results, with lower rates of dropout from treatment and decreased number of readmissions . In Norway, these kinds of services have been implemented in some parts of the country, to varying degrees; 50% of the hospitals report cooperation of follow-up . As in many parts of the world, general practitioners are one of the cornerstones of the health care system. Almost one-third of the patients included in the present study considered that their general practitioner was most important in their care. This is supported by the results of another study , in which 64% of deliberate self-harm patients were satisfied with their general practitioner. Taken together, these findings suggest that follow-up performed by general practitioners may be helpful.
The results of the present study indicate that problems and a need for help were present 3 months after discharge, as a large group of patients felt that they needed help from their general practitioner or a specialist, or even hospitalization. The findings reported in a case-control study performed by Appleby and coworkers  indicate that a considerable number of the individuals who had completed suicide had their care reduced at the final appointment before they committed suicide. Thus, a correlation has been found between suicide risk and reduced level of care. Furthermore, the low scores observed on the GSE (25.2) show that the participating individuals' belief in their own ability to cope with novel or difficult situations and to deal with a large variety of stressors is low . As a comparison, the mean GSE in the American general adult population is 29.5 .
The BHS scores obtained in our study were also high in the whole group (10.1). The mean BHS score in the general population of Ireland was 4.45  versus 4.5 in the general population of Norway  and 10.2 among suicide attempters in a somatic hospital .
The depression score on the BDI observed in the present study was 23.3, which was elevated compared with that of suicide attempters in Bærum (19.4) . In that study, a matched control group from the general population had a BDI score of 5.1. This indicates that the level of hopelessness and depression is much higher among patients with self-poisoning 3 months after discharge than in the general population. It should be noted that the study of Dieserud and coworkers, which we have used as a comparison for the sample of the present study, excluded drug-related overdoses. The comparison of the group classified as drug-related poisoning in the present study with the general population showed that the levels of depression and hopelessness were much higher in the former. Together with the considerable level of repetition of self-poisoning, cutting, and injuries reported, these results indicate a need for more active treatment strategies.
The assessment of intention is often performed in consultation or in accordance with the assessment of psychiatric personnel, and it is likely that the classification will influence decisions regarding follow-up plans at the time of discharge. Bjornaas and coworkers found that the agreement between the patients' self-reported intention and the physician's evaluation of intention, using the same categories as in the present study, was high . In a 20-year follow-up study of all self-poisoning patients in Oslo, a suicidal motive upon admission were found to be the only predictor of later suicide as the risk were 3.1 times higher . Considering the findings of the present study, this point must be emphasized further if the clinical evaluations are one of the main contributors to the choice of follow-up that is offered at the time of discharge, especially as the psychiatric symptoms were at similar levels in the appeal and suicide attempt groups 3 months after discharge.
Strengths and limitations
This study was part of a multicenter study covering the capital of Norway and the largest surrounding municipality and therefore provides knowledge from one population with both self-poisoning and substance use related poisoning. It also provided the possibility of describing and comparing the respondents with the non-respondents regarding many variables. The use of validated and known instruments allowed the comparison of results. The knowledge about repeated acts based on self-reports from the patients who exhibited several repeated episodes of self-poisoning and cutting was another strength of the study, as many will not get in contact with health care services every time. It could be argued that there was a low response rate with regard to this question, but in spite of this, the amount of repeated self-harm and self-poisoning was considerable and clinically significant. Despite the rather low response rate obtained, the main conclusions of the study seem robust because even if the response rate were low, there is still a considerable group with high depression scores, long waiting times and further self-harming behavior.
Given the low response rate observed, the external validity of the present results may be questioned. The many similarities between respondents and non-respondents, however, support the validity of the main findings. A previous Swedish report indicated that the study of this patient group is difficult . Nevertheless, almost one-third of the current patients answered the questionnaire. Unfortunately, we were not given permission from the Regional Ethics Committee to call the patients. In Norway, we have not until very recently had the opportunity to link patient data to register data, although verification of the results might have been useful and strengthened the findings. Previous research found that suicidal behavior, morbidity, and mortality are relatively high among people with no fixed place of abode . It was impossible to avoid this selection bias, as postal addresses were not available. The term 'appeal' is controversial in the field of suicidology, as some fear the use of the term might implicate an underestimation of the seriousness of such acts. From clinical practice, there is a wide range between medically serious suicide attempts with an outspoken wish to die, and acts of self-poisonings which was never life threatening, but where the patient wished to communicate an unbearable situation to others in order to get help. In our study the main distinction between suicide attempters and appeals were their suicidal intent. Therefore, the terms 'moderate to high suicide intent' and 'low or no suicide intent' may have been used instead. We chose to keep the terms used in the original study forms used by the clinicians.