Study setting and participants
According to the idea of BCTs, the aim of ODS was to study the impact of the selected intervention in real life setting of psychiatric secondary services. The comparisons were made with a control group representing as similar patient population as possible treated with TAU methods in the psychiatric services of the same area. This setting allowed the assessment of differences in outcomes between health care units treating similar patients with different methods (more structured brief intervention vs. treatment as usual).
ODS patients were recruited in five psychiatric outpatient clinics and one psychiatric hospital ward in the South Ostrobothnia hospital district of Finland (population 200,000) during October 2009–October 2013. Consecutive patients who were referred to adult psychiatric services because of depressive symptoms, anxiety, self-destructiveness, insomnia or substance-related problems were screened using the Beck Depression Inventory (BDI, version 1A) [
14] . Those with a BDI score ≥ 17 (corresponding to at least moderate-level depression) at the screening phase were included in the intervention group (
n = 242). In addition to having depressive symptoms, these patients had various other psychiatric diagnoses which is typical in psychiatric secondary services. Patients with a likely or verified psychotic disorder (other than psychotic depression, International Classification of Diseases [ICD]-10, F2*.**) [
15] or organic brain disease were excluded.
The control group (
n = 205) was recruited from the South Ostrobothnia hospital district database of psychiatric outpatient clinics not participating in the ODS study (time period: October 2009–December 2012) and from the same psychiatric hospital ward before the start of the ODS (i.e., before 30 September 2009). Patients with a new referral to adult psychiatric services were selected in chronological order if their BDI score was ≥17 at admission and the Alcohol Use Disorders Identification Test (AUDIT) [
16] score at admission was available. Patients who had participated in the ODS study or had a likely or verified psychotic disorder (other than psychotic depression, ICD-10, F2*.**) or organic brain disease were excluded. The control group patients were matched with the intervention group patients by clustering according to the current psychiatric hospital contact (inpatient/outpatient), AUDIT score in four categories (0–7, 8–10, 11–19 and 20–40 points) and BDI score in two categories (17–29 or 30–60 points). Roughly 1650 patients´ files were screened before the amount of patients in the control group was considered to be close enough to the intervention group. The main characteristics of intervention and control group patients at the baseline are presented in Table
1. The characteristics were mainly similar between the groups, only the baseline GAF score (
p = 0.035) and the frequencies of personality disorders as a secondary psychiatric diagnosis (
p = 0.037) were statistically significantly different between the groups. The psychiatric diagnoses were register based and determined by the responsible doctor in previous treatment contacts.
Table 1
The main characteristics of patients at baseline
Number of patients | 242 | 205 | |
Female / Male | 148 (61.2) / 94 (38.8) | 111 (54.1) / 94 (45.9) | 0.135 |
Outpatient / Inpatient | 189 (78.1) / 53 (21.9) | 156 (76.1) / 49 (23.9) | 0.615 |
Register basedf clinical diagnosis (ICD-10): |
Primary psychiatric diagnosis | | | 0.103 |
Depressive disorder (F32.x) | 98 (40.5) | 89 (50.0) | |
Recurrent depressive disorder (F33.x) | 92 (38.0) | 48 (27.0) | |
Bipolar disorder (F31.x) | 19 (7.9) | 17 (9.6) | |
Anxiety disorders (F40–43.x) | 9 (3.7) | 20 (9.6) | |
Other disorders | 18 (7.4) | – | |
Secondary psychiatric diagnosis |
Panic disorder (F41.x) | 26 (10.7) | 16 (7.8) | 0.296 |
Phobic and other anxiety disorders (F40,42,43.x) | 22 (9.1) | 22 (10.8) | 0.550 |
Personality disorder (F6x.x) | 9 (3.7) | 17 (8.4) | 0.037 |
Tertiary psychiatric diagnosis |
Anxiety disorder (F4x.x) | 12 (5.0) | 5 (2.5) | 0.168 |
Personality disorder (F6x.x) | 8 (3.3) | 10 (4.9) | 0.393 |
Alcohol dependence (current) | 47 (19.4) | n/a | |
Use of other substance (last 12 months) | 22 (9.1) | n/a | |
| mean (SD) | mean (SD) | p |
Age (years) | 38.8 (12.2) | 39.5 (12.7) | 0.567 |
Baseline MADRSa | 23.2 (6.7) | n/a | |
Baseline BDIb | 27.9 (7.3) | 28.8 (7.9) | 0.214 |
Baseline AUDITc | 10.7 (9.9) | 9.8 (9.0) | 0.319 |
Baseline GAFd | 45.9 (10.7) | 43.6 (11.7) | 0.035 |
Antidepressive medication was used for 239 (98.8%) patients in the intervention group and for 203 (99.5%) patients in the control group (cumulative doses: 28.0 mg, standard deviation [SD] 20.6 mg vs. 26.7 mg, SD 21.2 mg fluoxetine equivalents [
17,
18] , respectively). Sixty-six (27.3%) patients in the intervention group and 72 (35.1%) patients in the control group were using antipsychotic medication (cumulative doses: median 62.5 mg, interquartile range [IQR] 93.75 vs. 125 mg, IQR 187.5 chlorpromazine equivalents [
19], respectively).
Information on somatic comorbidities was collected during the baseline assessment of the intervention group patients and from the psychiatric patient files of the control group patients. Thirteen patients in the intervention group and one patient in the control group had hypertension (I10). Four patients in the intervention group had diabetes mellitus type 1 (ICD-10, E10). Eight patients in the intervention group and one patient in the control group had diabetes mellitus type 2 (E11). In the intervention group, five patients had metabolic syndrome (E66), three patients had coronary heart disease (I25) and two patients had sequelae of cerebrovascular disease (I69). The patients in either group didn’t have any current infections.
Baseline assessment
For the ODS intervention group, the baseline assessment was based on the Cube Method which is an integrated assessment method for comorbid psychiatric and substance use disorders in clinical settings; the method was developed by Kampman and Lassila in the South Ostrobothnia hospital district [
20,
21]. Systematic assessment helps to target suitable treatment for each patient. The Cube Method was used to decide which patients would be additionally treated with motivational interviews. The baseline psychiatric diagnostic evaluation was performed using the Mini International Neuropsychiatric Interview (MINI) [
22]. Depressive symptoms were rated using the Montgomery–Åsberg Depression Rating Scale (MADRS) [
23] and level of functioning was assessed using the Global Assessment of Functioning scale (GAF) [
24]. All evaluations were conducted by experienced psychiatrists or trained research nurses. The use of psychotropic medications was recorded using a paper-and-pencil diary. The study protocol is described in more detail in the Clinical Trials registry [
13].
For the control group, all data were collected retrospectively from the hospital district patient registers. The diagnoses had been determined by the responsible doctor and were not confirmed using structured diagnostic interviews. GAF scores were estimated according to case notes by a trained research nurse supervised by a research doctor. In cases of insufficient information, scoring was omitted.
The GAF estimates were calibrated by prior rating of five test patients and comparing the possible differences between the raters (one research doctor and two research nurses). After the calibration process, the inter-rater correlations were strong (r > 0.9).
Procedures and follow-up
In the intervention group, BA and antidepressive medication were used for 239 patients. Additionally, motivational interviews (MIs) [
25] were used during the first appointments with patients having alcohol use problems (baseline AUDIT ≥11) to be able to better ensure the patients engagement to the treatment. BA and MIs were implemented by the regional staff (registered psychiatric nurses, psychiatric practical nurses and psychologists) responsible for the patient. The regional staff was systematically trained to use the selected interventions [
7]. The minimum duration of BA was set at four appointments. On request, we received information about the completed BA sessions from the therapists of 54 patients. The median number of sessions was 6.5 (IQR 8.25). The decision to start the medication and the type of medication used was based on clinical evaluation (at baseline and at 6 weeks) by the responsible doctor. If the baseline MADRS score was 20 or more the medication was started and the dose was elevated if necessary, or the type of antidepressive medication was changed (from selective serotonin reuptake inhibitors to serotonin–norepinephrine reuptake inhibitors). The follow-up included appointments with a clinical research nurse at 6, 12 and 24 months and depressive symptoms (MADRS) and the level of functioning (GAF) were checked (among other things, please see
ClinicalTrials.gov Identifier NCT02520271).
All patients in the control group were treated in public psychiatric secondary care in the same organization and catchment area as the intervention group over the same time period. Control patients received TAU according to the protocols of the respective treatment unit and were followed-up according to the case notes at 6, 12 and 24 months by estimating GAF scores and obtaining information about possible alcohol use. In the area, psychiatric secondary care is always supervised by a specialized psychiatrist. The psychiatric TAU usually consists of regular individual visits to the outpatient clinic with varying frequency depending on each patient’s situation. Most often, a nurse, psychologist or social worker is responsible for the individual visits; patients meet a doctor only at the beginning and end of the treatment and whenever the treatment plan is evaluated. The TAU methods varied according to the educational background and therapeutic training of the responsible staff.
For both groups, information about the frequency of outpatient visits, number of hospital days and dropouts was collected from patient registers at 6, 12 and 24 months follow-up points and information on mortality was collected from baseline (October 1, 2009) until March 30, 2016. The register data covered all cases during the whole follow up in both intervention and control groups (n = 242 and n = 205, respectively).
Statistical methods
The power calculations resulted in the following detectable limits for primary outcomes MADRS and GAF. Within intervention group including the current sample size, the paired analyses were able to detect a 1.7 point mean response difference in MADRS with a power of 0.8 and type I error probability of 0.05. Between intervention and control groups including the current number of pairs, the independent analyses were able to detect a 3.6 point mean response difference in GAF with a power of 0.8 and type I error probability of 0.05.
The independent samples t-test was used to compare the age distributions, GAF and AUDIT scores between the intervention and the control group at baseline. Chi-square tests were used to compare the frequencies of gender, inpatients and outpatients, psychiatric diagnoses and the patients hospitalized during the follow-up between the groups. A linear mixed model for repeated measures was used to analyze the changes in MADRS score from baseline to the three follow-up points in the intervention group patients and the changes in GAF score between the intervention and control groups. In the latter model, parameters time, patient group and an interaction term time*patient group were used in analyses. The Mann–Whitney U test was used to analyze antipsychotic medication doses and the number of days spent in hospital. The level of statistical significance was set at < 0.05. All calculations were performed using the software SPSS for Apple Macintosh version 24 (IBM SPSS Statistics, Armonk, NY) and PS: Power and Sample Size Calculation (version 3.1.2, Vanderbilt University, Nashville TN) [
26].