In line with previous findings, in a recent randomized controlled trial (RCT), we found that home treatment (HT) for acute mental health care can reduce (substitute) hospital use among severely ill patients in crises. This study examined whether the findings of the RCT generalize to HT services provided under routine care conditions.

We compared patients who received HT during the RCT study phase with patients who received the same HT service after it had become part of routine mental health services in the same catchment area. Sociodemographic and clinical characteristics as well as service use (HT and hospital bed days) were compared between the RCT and the subsequent routine care study period.

Compared to patients who received HT during the RCT, routine care HT patients were more often living with others, less often admitted compulsorily, more often diagnosed with anxiety and stress-related disorders (ICD-10 F4) and less often diagnosed with schizophrenia spectrum disorders (F2). When compared to patients who were exclusively treated on hospital wards, involvement of the HT team in patients’ care was associated with a clear-cut reduction of hospital bed days both during the RCT and under routine care conditions. However, unlike during the RCT study period, involvement of HT was associated with longer overall treatment episodes (inpatient + HT days) under routine care conditions.

HT seems to reduce the use of hospital bed days even under routine care conditions but is at risk of producing longer overall acute treatment episodes.

Home treatment has been proposed as an alternative to acute in-patient care for mentally ill patients. However, there is only moderate evidence in support of home treatment.

To test whether and to what degree home treatment services would enable a reduction (substitution) of hospital use.

A total of 707 consecutively admitted adult patients with a broad spectrum of mental disorders (ICD-10: F2–F6, F8–F9, Z) experiencing crises that necessitated immediate admission to hospital, were randomly allocated to either a service model including a home treatment alternative to hospital care (experimental group) or a conventional service model that lacked a home treatment alternative to in-patient care (control group) (trial registration at ClinicalTrials.gov: NCT02322437).

The mean number of hospital days per patient within 24 months after the index crisis necessitating hospital admission (primary outcome) was reduced by 30.4% (mean 41.3 v. 59.3, P<0.001) when a home treatment team was available (intention-to-treat analysis). Regarding secondary outcomes, average overall treatment duration (hospital days + home treatment days) per patient (mean 50.4 v. 59.3, P = 0.969) and mean number of hospital admissions per patient (mean 1.86 v. 1.93, P = 0.885) did not differ statistically significantly between the experimental and control groups within 24 months after the index crisis. There were no significant between-group differences regarding clinical and social outcomes (Health of the Nation Outcome Scales: mean 9.9 v. 9.7, P = 0.652) or patient satisfaction with care (Perception of Care questionnaire: mean 0.78 v. 0.80, P = 0.242).

Home treatment services can reduce hospital use among severely ill patients in acute crises and seem to result in comparable clinical/social outcomes and patient satisfaction as standard in-patient care.