PRIM-CARE was a pragmatic cluster randomized controlled trial that compared a 12-week nurse care-manager program with treatment as usual for patients with depression [8]. The methods have been described in greater detail in previous publications [8, 10]. In brief, the study was carried out at 23 primary care centers (11 intervention and 12 control) in the regions of Västra Götaland and Dalarna, Sweden, starting in 2014. Centers were eligible for inclusion unless they already had a care manager or the equivalent of a care manager. The centers were located in rural and urban areas of varied socioeconomic status. Figure 1 presents a flow chart of the trial.
Population
Patients at the participating primary care centers who met the eligibility criteria were invited to take part in the study. Patients were included if they were ≥ 18 years, newly diagnosed (< 1 month) with depression (International Classification of Diseases and Other Health Problems, 10th edition: diagnostic codes F32 and F33) [11], and the depression was mild or moderate (< 35 points on the Montgomery-Asberg Depression Rating Scale) [12, 13]. Patients who had recovered from previous episodes of depression could be included in the study. Patients were excluded if they had a current diagnosis of cognitive impairment (e.g., dementia), bipolar disorder, psychosis, or substance use disorder or were unable to speak or read Swedish.
Outcomes
Outcomes in the PRIM-CARE study were measured at baseline, 3, 6, 12, and 24 months.
Depressive symptoms were measured with the Montgomery-Asberg Depression Rating Scale - Self-assessment (MADRS-S) [12, 13]. MADRS-S has nine items and asks about symptoms during the last 3 days. Total scores range from 0 to 54. A total of 0 to 12 points indicates no or very mild symptoms of depression; 13 to 19 points, mild symptoms; 20 to 34 points, moderate symptoms; and 35 to 54 points, severe symptoms.
Fifty percent reduction in depressive symptoms as measured with MADRS-S scores from baseline to each follow-up occasion was also investigated.
Remission from depression was operationalized as a score of < 12 points on MADRS-S, which indicates no or very mild symptoms of depression.
Quality of life was measured with the EuroQol five Dimension 3 L scale, English tariff (EQ-5D-3L) [14, 15]. The EQ-5D-3L is a three-level version of the EQ-5D health index. It measures health-related quality of life and has two sections. In the first section, respondents select one of three levels (no problems, some problems, extreme problems) that best matches their health-related quality of life in the areas of mobility, self-care, everyday activities, pain/discomfort, and anxiety/depression.
Self-efficacy, confidence in care, and quality of care from a patient perspective were assessed at 24 months via a study-specific questionnaire. The English translation of the questionnaire is presented in Supplemental material 1. The 20 items on the questionnaire were inspired by items on patient questionnaires about self-efficacy [16,17,18,19] and an instrument about quality of care from a patient perspective [20]. The study-specific questionnaire was not validated. Response alternatives for the statements about self-efficacy and confidence in care were provided on a 5-level Likert scale that ranged from not at all confident to completely confident. Responses were dichotomized in the analysis (not at all confident to moderately confident versus very confident to completely confident). Responses to the statements about quality of care from a patient perspective were provided on a 5-level Likert scale that ranged from not at all true to completely true. Responses to this item were also dichotomized in the analysis (not at all true to moderately true versus very true to completely true).
Use of antidepressant medication (Y/N) was an outcome in the analyses because a goal of care management was to support patients in adhering to antidepressant treatment [8]. According to national guidelines, once antidepressant use is initiated, patients should continue to use the medication for at least 6 months [21].
Other variables
Data on patient background variables were collected at baseline, including age, sex, working (full-time, other [25–75% of full time]), marital status (cohabiting, single), born outside Nordic country (Y/N), educational level (primary education, secondary education, university or college education), leisure-time physical activity (sedentary Y/N), smoking (Y, sometimes/No), alcohol at least once a week (Y/N), sick leave last year (Y/N), and on sick leave at baseline (Y/N).
Psychotherapy use was also monitored during the study period. Starting at the 3-month follow-up, the patients reported whether they had had contact with a psychologist or psychotherapist since baseline (or the previous follow-up), and if so, how often. At the 3-month follow-up, these self-reported data were complemented with data from electronic patient records. Psychotherapy use (Y/N) was treated as a potential confounder in the analysis.
Randomization
A total of 23 primary care centers were included and divided into two strata, rural (n = 12) and urban (n = 11) [8]. The centers in each stratum were then divided into six blocks of two centers each. One center in each block was randomized to the intervention group (n = 11) and one to the control group (n = 12) (Fig. 1) [8]. Randomization was performed by a statistician who was not involved in the study and who was blinded to the identity of the primary care centers. Cluster randomization was chosen to avoid treatment contamination between the intervention and control groups [22], which can occur if patients are individually randomized. This design is common in studies of organizational changes [23, 24].
Intervention
One nurse at each intervention center worked approximately 20 to 25% of full-time as care manager. Care managers received a total of 5 days of training in delivering the intervention according to the protocol [8]. General practitioners and managers at the intervention centers also received 2 days of training. Every other month during the study period, personnel from the region and study personnel held in-person support meetings for care managers. A research assistant from the study group visited the intervention centers weekly to provide additional support for personnel and monitor adherence to the protocol.
The intervention was 12 weeks long. The care manager first met the patient for a one-hour meeting to develop an individual care plan. The care manager followed up the care plan with a phone call at 2, 4, 6, 8, and 12 weeks. The purpose of these calls was to monitor depressive symptoms with a rating scale (MADRS-S), encourage behavioral activation, and support adherence to medication and recovery. Care managers stayed in regular contact with the patient’s general practitioner, therapist, and other care providers to inform them about changes that required attention (e.g., potential side effects of medication, new or worsening symptoms, or failure to respond to treatment). Patients could also call the care manager as needed. At the end of the intervention, the care manager and patient developed an individual relapse prevention plan.
Control
Patients at the control primary care centers received usual care, which did not include the intervention (i.e., a care manager). The Swedish National Board of Health and Welfare provides guidelines for treating depression and anxiety in primary care [21]. In brief, these guidelines state that patients should receive a rapid follow-up appointment after diagnosis with depression, as well as ongoing easy access to the general practitioner. Recommended treatment for depression includes cognitive behavioral or interpersonal therapy and/or antidepressant medication [21]. A research assistant from the study group visited control centers regularly to monitor adherence to the protocol and collect data about routine care from electronic patient records.
Enrollment of patients, diagnostic procedure, and data collection
During patients’ primary care center appointment for depressive symptoms, general practitioners invited the patients to take part in the study. Patients received verbal and written information about the study and provided written informed consent to participate. They received an appointment with the care manager (intervention centers) or a study nurse (control centers), who used the depression module of PRIME–MD [25] to confirm the general practitioner’s diagnosis of mild to moderate depression. At the intervention centers, care managers collected baseline data, and at control centers, a study nurse collected them. Follow-up data were gathered with questionnaires sent via regular mail 6, 12, and 24 months after the start of the study.
Statistical analyses
Continuous variables, such as age and symptom severity, were analyzed with the independent sample t-test or Mann-Whitney U test. Categorical variables were described as numbers and percentages and analyzed with Pearson’s chi-square test. Mean intra-individual change in depressive symptoms and quality of life in the intervention and control group were compared with mixed model analysis with repeated measures, adjusted for the type of primary care center (sparse vs medium-high patient inclusion rate) and the patient’s age, sex, educational level, use of antidepressant medication, and variable-related scores at baseline [26]. Statistical significance was set at P < 0.05. No multiple adjustments were considered. Because of sparse data from some of the primary care centers, we did not adjust for the cluster randomization.
Power calculations are shown in detail in the article that describes the 3- and 6-month results of PRIM-CARE [8]. The statistical analyses were carried out with SPSS, version 25, and SAS, version 9.4.
Ethical considerations
The Regional Ethical Review Board in Gothenburg approved the study on 2 January 2014 (Dnr 903–13) and the 24-month follow up on 27 July 2018 (T598–18). Complementary approval for the Dalarna portion of the study was received from the Regional Ethical Review Board in Gothenburg on 7 January 2015 (T975–14). The study was carried out in accordance with the World Medical Association Declaration of Helsinki [27].
The heads of primary care in Region Västra Götaland and Dalarna County provided written permission to conduct the study, as did the manager of each participating primary care center. Participating personnel provided verbal informed consent. After receiving verbal and written information about the study and before they were included in the study, participating patients provided written informed consent.
The study was registered on 2 February 2015 at ClinicalTrials.gov, identifier NCT02378272.