A Naturalistic Controlled Trial of Individual Placement and Control in Bodø (IPSBODO)

Nordlandssykehuset HF, University of Tromso, London School of Economics and Political Science, University of Melbourne, St George's Healthcare NHS Trust, The University of New South Wales

The investigators aim to test the effectiveness of Individual Placement and Support (IPS) on employment, welfare dependency, and public-sector health care utilization. This is a naturalistic controlled trial, where one municipality (Bodø in Norway) with about 50000 inhabitants get access to IPS services in public sector mental health services during the period 2013-2016. The target group for the intervention is patients with severe mental illness (SMI) in the age group 18-40 at time of treatment. Patients already receiving lifelong disability benefits will be excluded. The control group will be an average of 10 municipalities in Norway without IPS services. Data for outcomes will be based on public registries available for research.

The investigators aim test the hypotheses that Individual Placement and Support (IPS) is effective in increasing employment, reducing welfare dependency, and reducing public-sector health care utilization. This is a naturalistic controlled trial, where one municipality (Bodø in Norway) with about 50000 inhabitants get access to IPS services in public sector mental health services during the period 2013-2016. The follow-up period will last until 2022. The target group for the intervention is patients with severe mental illness (SMI) in the age group 18-40 at time of treatment, with an ambition of 200 received IPS services. The investigators aim to study the effectiveness on the societal (not individual) level, assuming spill-over effects within Bodø municipality as a result of system changes in the treatment municipality (Bodø). Patients already receiving lifelong disability benefits will be excluded. The control group will be 10 similar municipalities in Norway without IPS services and will be chosen on the basis of similarities in economy, demography and geography and collected from publicly available information. Approximately 77000 subjects will be enrolled in this study; 7000 subjects from Bodø municipality and 70000 from the other 10 municipalities (the control group). Data for outcomes will be based on public registries available for research. Outcome operationalizations are: Employment will be based on registry data from tax records with information both on employment (yes/no and period) and taxable income (NOK). Welfare dependency will be based on registry data including rehabilitation benefits, disability benefits, sickness absence, unemployment benefits, and social benefits. These will be analysed separate and combined, and the main hypothesis will be based on the latter. Health care utilization will be based on data from the Norwegian Patient Register. Data for in-patient and out-patients treatments for mental illness and other (separate analyses) will be used. These will be analysed separate and combined, and the main hypothesis will be based on the first. The data will be analyzed in two rounds; first in 2019, and a follow-up analysis in 2022. A difference-in-difference approach will be used to measure the effects.

Naturalistic controlled trial with one municipality (Bodø) as treatment group and 10 similar municipalities without IPS services as controls.

Rates of welfare dependency (rehabilitation benefits, disability benefits, sickness absence, unemployment benefits, and social benefits) in Bodø (intervention municipality) will be compared to rates of welfare dependency in 10 resembling municipalities (Kongsberg, Lier, Røyken, Horten, Tønsberg, Larvik, Færder, Porsgrunn, Grimstad og Steinkjer) without IPS-services (control municipalities) using Norwegian registry data.

Rates of employment in Bodø (intervention municipality) will be compared to employment rates in 10 resembling municipalities (Kongsberg, Lier, Røyken, Horten, Tønsberg, Larvik, Færder, Porsgrunn, Grimstad og Steinkjer) without IPS-services (control municipalities). The measurement tool will be Norwegian registry data from tax records with information both on employment (yes/no and time period) and taxable income (NOK).

Rates of employment in Bodø (intervention municipality) will be compared to employment rates in 10 resembling municipalities (Kongsberg, Lier, Røyken, Horten, Tønsberg, Larvik, Færder, Porsgrunn, Grimstad og Steinkjer) without IPS-services (control municipalities). The measurement tool will be Norwegian registry data from tax records with information both on employment (yes/no and time period) and taxable income (NOK).

Welfare dependency rates (rehabilitation benefits, disability benefits, sickness absence, unemployment benefits, and social benefits) in Bodø (intervention group/municipality) will be compared to 10 similar municipalities without IPS-services (control groups/municipalities).

Rates of welfare dependency (rehabilitation benefits, disability benefits, sickness absence, unemployment benefits, and social benefits) in Bodø (intervention municipality) will be compared to rates of welfare dependency in 10 resembling municipalities (Kongsberg, Lier, Røyken, Horten, Tønsberg, Larvik, Færder, Porsgrunn, Grimstad og Steinkjer) without IPS-services (control municipalities) using Norwegian registry data.

Rates of health care utilization in Bodø (intervention municipality) will be compared to health care utilization rates in 10 resembling municipalities (Kongsberg, Lier, Røyken, Horten, Tønsberg, Larvik, Færder, Porsgrunn, Grimstad og Steinkjer) without IPS-services (control municipalities). Health care utilization will be based on data from the Norwegian Patient Register. We will use data for in-patient and out-patients treatments for mental illness and other (separate analyses), collected from the Norwegian Patient Register as a measurement tool. These will be analysed separate and combined, and the main hypothesis will be based on the first.

Rates of health care utilization in Bodø (intervention municipality) will be compared to health care utilization rates in 10 resembling municipalities (Kongsberg, Lier, Røyken, Horten, Tønsberg, Larvik, Færder, Porsgrunn, Grimstad og Steinkjer) without IPS-services (control municipalities). Health care utilization will be based on data from the Norwegian Patient Register. We will use data for in-patient and out-patients treatments for mental illness and other (separate analyses), collected from the Norwegian Patient Register as a measurement tool. These will be analysed separate and combined, and the main hypothesis will be based on the first.

Inclusion Criteria: Severe mental illness Work assessment allowance Exclusion Criteria: Disability benefit