Brief admission by self-referral, a novel crisis intervention designed to reduce suicide and self-harm in adults, was adopted for adolescents in paediatric psychiatry in Malmö, Sweden, in 2018. We aimed to investigate changes in utilisation of emergency psychiatric care. We did an observational longitudinal cohort study in The University Hospital in Malmö, Sweden, which provides the only psychiatric emergency unit with 24 h psychiatric facilities in Region Skåne. Eligible patients were those aged 13-17 years who were admitted to the psychiatric facility, who had at least one emergency visit or admission during the 6 months before admission, and had prominent features of instability and self-harm, corresponding to at least three of the nine criteria for borderline personality disorder as per the DSM-5 as assessed by a paediatric psychiatrist during the admission. Patients with intellectual disabilities, psychosis, or language barriers were excluded. Patients who signed a brief admissions contract between April 1, 2018, and April 30, 2021, were eligible for inclusion in the study. A brief admissions contract allows patients to admit themselves to psychiatric emergency care for a transitory time. The primary outcome measures were the number of emergency visits, emergency admissions, inpatient days, and episodes of coercive (involuntary) care, compared at individual level before and after signing the brief admissions contract until end of follow-up. The number of visits and days were modelled using random-effects Poisson regression models, and the relative changes in the expected numbers of days per time unit were reported as rate ratios (RRs). Of the 928 patients admitted to the psychiatric facility between April 1, 2018, and April 30, 2021, 60 were excluded, and a further 801 did not meet the inclusion criteria for age, previous emergency visits, or having at least three of the nine criteria of borderline personality disorder. 67 patients were eligible for inclusion, but four patients did not sign a contract. 63 patients were included in the study, including 60 females (95%) and three (5%) males, with a mean age of 14·8 years (SD 1·7). Ethnicity data were not collected. Patients were followed up for a median of 13·5 months (IQR 9·2 -19·6). After signing the contract, there was a decrease in the number of emergency visits (RR 0·22 [95% CI 0·15-0·32]; p<0·0001), emergency admissions (RR 0·26 [0·19-0·35]; p<0·0001), inpatient days (RR 0·29 [0·26-0·32]; p<0·0001), and inpatient days including brief admissions (RR 0·44 [95% CI 0·40-0·48]; p<0·0001). Episodes of coercive care did not change significantly (RR 0·99 [95% CI 0·40-2·43]; p=0·98). Psychiatric evaluation due to persistent suicidal ideations immediately after discharge was required for five patients. Our findings suggest that brief admissions can be successfully implemented in paediatric psychiatry and appear to be an effective crisis management method for adolescents, associated with reduced demand for emergency care. Future randomised controlled trials are warranted. Region Skåne Health Care Authority.