Predictors of suicidality and self-harm related admissions to psychiatric and general hospitals

Abstract

Background and aims: Although suicide risk is a frequent reason for acute psychiatric admissions, there is a lack of knowledge on the prevalence and characteristics of such admissions and very little is known about the short- and long-term outcome in relation to severe self-harm after psychiatric discharge. The overall aim of this thesis was to describe the prevalence of suicide risk, suicidal ideation, non-suicidal self-harm, and suicide attempt as the main or contributing reasons for acute psychiatric admissions: as well as to study these variables together with other clinical, sociodemographic, diagnostic, and treatment related variables as predictors of acute psychiatric readmissions and self-harm induced somatic admissions. Methods: This thesis is based on the results from two prospective, observational and longitudinal cohort studies, and one interview study. The studies are based on data from patients admitted to a psychiatric acute unit at a hospital, which has a catchment area of about 400 000 inhabitants. Papers I (n=1245) and II (n=2827) used data obtained by the standardised assessments of all consecutive admissions during one and three years respectively. Study I examined the data assessed at the index admission (each patient’s first admission to the psychiatric acute unit during the inclusion period) as predictors of readmission due to any cause, and readmissions due to suicide risk, within a mean follow-up time of 1.5 years. Study II analysed data assessed at index admission and possible readmissions, as well as follow-up data on psychiatric outpatient treatment as predictors of selfharm induced somatic admissions within a mean follow-up time of 2.3 years. Paper III included 308 patients who were randomly selected from non-psychotic patients consecutively admitted to the psychiatric acute unit, due to suicide risk. A multiple regression model was used to examine if post-traumatic stress disorder (PTSD) predicted the number of self-harm induced somatic admissions within 6 months, when adjusted for borderline personality disorder (BPD). Secondly, a structural model comprising two latent BPD factors, ‘dysregulation’ and ‘relationship problems’, as well as PTSD and several other diagnostic variables was applied to examine the associations between these variables and the number of post-discharge self-harm induced somatic admissions. Results: Suicide risk was the main or contributing reason for 54% of the total index admissions and 62% of the total readmissions. In the most frequently admitted patients, 80% of the admissions were related to suicide risk. Of the total cohort, about one in ten patients had at least one self-harm induced somatic admission during follow-up. The proportion of patients having their first such admission within the first six months after psychiatric discharge was 48%, while 42% of the patients had their first self-harm induced somatic admission within the subsequent six months. Important predictors of self-harm induced somatic admission, were having had the most recent psychiatric admission related to nonsuicidal self-harm or to a suicide attempt. Other significant predictors were having a history of psychiatric hospitalisation before the index admission, psychiatric readmissions during follow-up, an increasing number of psychiatric outpatient consultations during follow-up, and having a diagnosis of recurrent depression, BPD, substance use disorders, or PTSD. Only about half of the somatic hospital admissions identified by the researchers as caused by self-harm, received an ICD-10 diagnosis of intentional self-harm. In the subgroup of patients with suicide risk related acute psychiatric admissions one in five patients had at least one self-harm induced somatic admission within 6 months after discharge. Among these patients, PTSD and BPD predicted the number of self-harm induced somatic admissions with nearly the same estimates. An emotional dysregulation factor based on the BPD criteria was significantly associated both directly and indirectly, via PTSD, with the number of somatic self-harm admissions; even when controlled for other relevant psychiatric disorders and symptoms. Conclusions: Suicide risk is a frequent reason for acute psychiatric admissions, and about one in ten patients have at least one self-harm induced somatic admission during the year following psychiatric discharge. The risk of such severe self-harm is high over an extended period after discharge. Treatment of the underlying disorder(s) may not be sufficient to prevent self-harm, and the repetition of self-harm in particular. In addition, there is a need for evidencebased psychosocial and behavioural interventions that may directly and transdiagnostically address suicidal thoughts and behaviours. Treatments that help patients to better deal with problems related to emotional dysregulation may contribute to preventing or reducing self-harm behaviour in several subgroups of patients admitted to psychiatric acute units.