Implementation and Evaluation of Primary Care Behavioral Health in Sweden (KAIROS)

Implementation and Evaluation of Primary Care Behavioral Health in Sweden - A Cluster Randomized Design for Superiority and Non-superiority Comparisons of Effects on Patient Reach, Availability, Costs, Sick-leave, Model Fidelity, and Patients' Everyday Function, Symptoms, and Experiences

Vastra Gotaland Health Care Region, The Kamprad Family Foundation for Entrepreneurship, Research & Charity, Karolinska Institutet, Capio Group

In this multicenter study, the investigators want to compare treatment outcomes for patients with mental and behavioral health problems in traditional primary care (Care As Usual, CAU) and primary care centres that work according to the Primary Care Behavioral Health (PCBH) model. In addition to this, the investigators want to study organisation-level outcomes, such as access to care, perceived teamwork and work environment. To achieve this, primary care centres that have expressed interest in implementing PCBH will be cluster randomised between implementing directly or waiting for implementation.

The overarching goal of primary care is to offer all patients individualised and context-sensitive healthcare with high access and continuity. One of the reasons primary care struggles with this goal is that a large proportion of patients suffer from mental and behavioural health problems, alone or in combination with one or several chronic illnesses. Despite many patients needing psychosocial interventions, there is a lack of mental health professionals as well as clear pathways for these patients. Primary Care Behavioral Health (PCBH) is an innovative way of organising primary care, where mental health professionals have more yet shorter visits, strive for same-day access, and have an active consulting role in the primary care team. To help patients achieve relevant behavioral changes, so called brief interventions are used, which are based on isolated components from psychological treatments such as Cognitive Behavioral Therapy (CBT) and Acceptance and Commitment Therapy (ACT). Brief interventions usually stretch over 1-4 treatment sessions. Assessments within the model are generally contextual and largely avoid psychiatric diagnostics, instead focusing on the patient's situation and their associated coping strategies - whether they be positive or negative. Although PCBH is gaining in popularity, recent reviews conclude that among the evaluation trials there are very few comparative studies. As such, there is a great need for proper evaluation of a method that is already widely implemented. Data will be collected at 17 primary care centers (PCCs) in Västra Götalandsregionen, Sweden, that have expressed an interest in implementing PCBH. The study is looking at both patient and organizational level variables. The PCCs will be randomized between implementing PCBH immediately (December 2021) or delayed with 5-9 months (implementation in late spring or early autumn of 2022). Outcome data will be collected also from the PCCs waiting for implementation, while they apply Care as Usual (CAU). Collection of patient self-ratings will start before the first implementation of PCBH to ensure functional data collection routines and possibly create a longer baseline, but the main analysis will use data from the period from when PCCs in the Early implementation arm actually have implemented PCBH to a good enough level and until PCCs in the control group start their implementation. Data from all PCCs will be continuously be collected also during the period when the delayed group implements PCBH. For individual patients, the primary end-point for the primary outcome (everyday functioning) will be 12 weeks after the first visit. Patient outcomes are also measured after 1 year. Outcome at the organizational level will be measured in several ways, and here the primary outcome will be two measures of waiting time between identification of patient (i.e. them self-referring for mental or behavioral health issues) and the first visit: (a) actual waiting-time and (b) Outcome at the organizational level will be measured in several ways, and here the primary outcome will be two measures of waiting time between identification of patient (i.e. them self-referring for mental or behavioral health issues) and the first visit: (a) actual waiting-time for each patient identified as relevant for inclusion and (b) Third Next Available Appointment (TNAA) for the PCC as a whole, measured at the same day and time each week. Fidelity will be measured by an expert group as well as using four questionnaires, one for each of mental health professionals, medical doctors, registered nurses and leadership. Implementation data will also be collected from these professional groups/roles at the implementing PCCs, for example with the s-NoMAD questionnaire. During the implementation process, all psychosocial resources (all relevant psychologists, social workers, psychotherapists and so on) will be trained in the PCBH model during 3 days. The psychosocial resources are required to already be equipped with basic training in psychotherapy, for example in CBT or psychodynamic therapies. All training will be followed up with six 4-hour group supervision sessions. PCC leaders, doctors and nurses will be included in the supervision when a need to work on inter-professional issues arises and might also receive some separate, extra supervision for up till 10 h in total if necessary. In addition to this, all PCCs will go through a half-day training of all clinical personnel at the PCC, of which 2 hours is set aside with an implementation group where all professions at the PCC are represented. The overall main research question is: Is PCBH superior to traditional primary health care in yielding the best patient outcomes regarding level of functioning (primary outcome) and symptoms, reaching the most patients, and reducing wait-times and costs? Secondary research questions include: Is (1) true also for the sub-group of complex patients with somatic/mental comorbidities and social difficulties? Is PCBH superior to traditional primary care in reducing objectively measured sick-leave? Is this, and level of functioning and symptoms, moderated by which intervention is given within PCBH? Is this true also for the sub-group of complex patients? Is this subgroup also reached by behavioral health interventions more than in traditional primary care? Does the implementation of PCBH result in lower levels of prescribed pharmacological treatment for depression, anxiety, sleep, and pain? What changes to group and organizational variables like teamwork, knowledge about and respect for other professions, clinical knowledge about behavior health and work environment satisfaction occur after the implementation of PCBH? How does the level of fidelity, acceptability and implementation quality affect patient and implementation outcomes? Do these variables moderate results in all above research questions? Which obstacles and facilitators are observed during implementation of PCBH? What factors influence fidelity and health care personnel satisfaction with PCBH? PCBH has the potential to increase the quality and access of care for many patients with mental and behavioural health problems. This study is the first to step towards answering whether or not the effects of PCBH are large enough to merit large-scale implementation.

Primary care centers are cluster randomized between implementing PCBH directly or implementing later. Patients seeking care at each PCC are not individually randomized, but are given the form of care that is currently offered at their PCC (pre- or post-implementation).

Patients will only be informed that the study is assessing the quality of primary care, and will not know that PCBH has been or will be implemented. Since primary and most secondary outcomes are patient-rated, outcomes assessors are also deemed to be blind.

PCCs randomized to this arm will have a delayed start of their PCBH implementation, waiting between 5-9 months. During this time, the same patient-level and organizational-level data will be collected from these centers while they continue to use traditional primary care / Care As Usual (CAU) .

Focused CBT, Focused ACT, Focused Cognitive Behavioural Therapy, Focused Acceptance and Commitment Therapy

'Brief Interventions' (BI) is a multitude of interventions used in patient visits within PCBH. BI start immediately at the initial consultation, which ends with a personally tailored and evidence-informed plan adjusted to the patient's context. The interventions within BI often have their foundation in CBT, ACT or Motivational Interviewing (MI), however interventions from other schools of therapy can also be used. The common theme is that they are principle-based rather than manual-based and focus on behavioural change in relation to a problem, rather than focusing on a specific diagnosis. Follow-up appointments are scheduled flexibly depending on the patient's perceived need. A BI treatment usually consists of 1-4 appointments with several weeks apart and has an open ending, where the patient easily can schedule a new appointment. Clinicians delivering brief interventions will have had 3 days of training as well as regular supervision.

Care As Usual includes the PCC's current routine care for patients with mental and behavioral problems. This can include a multitude of procedures and treatment, such as pharmacological treatment, supportive care, cognitive behavioural therapy (CBT) and psychodynamic therapy of varying lengths. Interviews with patients as well as medical journals will be used to categorise what type of care each individual patient has received.

The eight items constituting the four domains of Life activities, Cognition, Getting along, and Participation in the 12-item version of the WHO Disability Assessment Schedule 2.0 (WHODAS-12) will be used as primary outcome, since these are condition-independent measures of Behavioral Health relevant everyday functioning. The scale ranges from 0 to 32 points. A lower score means better functioning. As such, a lower score is a better outcome.

Waiting time from the patient contacting their PCC with a behavioural health concern to being seen by a physician or psychosocial resource among patients in the study

From documented identification/self-referral to documented first visit at PCC (up to 6 months after identification)

The two domains of Mobility and Self-care will be measured and the whole WHODAS will be used in some secondary analyses, but not for the primary outcome since these domains tend to form two independent factors, many patients will have initial adequate function in these domains and they are not expected to change much. The scale ranges from 0 to 48 points. A lower score means better functioning. As such, a lower score is a better outcome.

Waiting time until next available appointment for a third hypothetical patient that calls in on a given day for every clinician at the PCC, measured on the same time every week (e. g. Mondays at 9 AM), a measure of access to and and availability of care that is independent of patients own ability to show up at an offered time.

The full version of a well-validated scale for measuring depressive symptoms. PHQ-9 total score for the nine items ranges from 0 to 27. A lower score means less depressive symptoms. As such, a lower score is a better outcome.

PHQ-2 consists of two questions from the PHQ-9. This will be administered as part of the PHQ-9 on the above mentioned time points, and will replace the PHQ-9 on the mid-treatment measurements. The scale ranges from 0 to 6 points. A lower score means less depressive symptoms. As such, a lower score is a better outcome.

The full version of a well-validated scale for measuring anxiety symptoms. GAD-7 total score for the seven items ranges from 0 to 21. A lower score means less anxiety symptoms. As such, a lower score is a better outcome.

GAD-2 consists of two questions from the GAD-7. This will be administered as part of the GAD-7 on the above mentioned time points, and will replace the GAD-7 on the mid-treatment measurements. The scale ranges from 0 to 6 points. A lower score means less anxiety symptoms. As such, a lower score is a better outcome.

Social anxiety symptoms are measured with a very short version of a well established patient-rated scale for social phobia. The scale ranges from 0 to 12 points. A lower score means less symptoms of social anxiety. As such, a lower score is a better outcome.

Symptoms of alcohol misuse are measured with a very short version of a well established patient-rated scale for alcohol misuse. The scale ranges from 0 to 12 points. A lower score means less symptoms of alcohol misuse. As such, a lower score is a better outcome.

Symptoms of panic attacks are measured by a very short version of a well established patient-rated scale and has been created by factor analytic item-reduction and sensitivity to change analyses from large datasets from previous trial within the research group. The scale ranges from 0 to 8 points. A lower score means less symptoms of panic disorder. As such, a lower score is a better outcome.

Karolinska Exhaustion Disorder Scale 3-item (KEDS-3) + 2 confirming questions from the Swedish questionnaire 'självskattat utmattningssyndrom' (s-UMS) ['self-rated burnout-syndrome']

Symptoms of stress-related exhaustion are measured with a very short version of a well established patient-rated scale and has been created by an expert group, who chose three representative questions from the original KEDS scale. To aid with diagnostics, two confirmatory questions from the Swedish s-UMS scale were added, asking for duration and presumed cause of symptoms. The scale ranges from 0 to 18 points. A lower score means less symptoms of stress-related exhaustion. As such, a lower score is a better outcome.

Stress symptoms are measured by a very short version of a well established patient-rated scale and has been created by factor analytic item-reduction and sensitivity to change analyses from large datasets from previous trial within the research group. To increase face validity of the very short scale, we also added two more direct questions ("Have you been feeling stressed lately?", "Have you been feeling tense lately?"). The scale ranges from 0 to 16 points. A lower score means less stress symptoms. As such, a lower score is a better outcome.

Symptoms of sleep disorders are measured by a very short version of a well established patient-rated scale and has been created by factor analytic item-reduction and sensitivity to change analyses from large datasets from previous trial within the research group. The scale ranges from 0 to 8 points. A lower score means less sleep problems. As such, a lower score is a better outcome.

Symptoms of health anxiety are measured by a very short version of a well established patient-rated scale and has been created by factor analytic item-reduction and sensitivity to change analyses from large datasets from previous trial within the research group. The scale ranges from 0 to 9 points. A lower score means less symptoms of health anxiety. As such, a lower score is a better outcome.

Symptoms of obsessive compulsive disorder are measured by a very short patient-rated scale and has been created by an expert group. The scale ranges from 0 to 12 points. A lower score means less obsessive/compulsive thoughts and behaviors. As such, a lower score is a better outcome.

Pain is measured on a scale from 0 to 10, where 0 is no pain and 10 is worst pain imaginable. As such, a lower score is a better outcome.

Also, a total index for level of symptoms will be calculated from all symptom items, where different weights will be used to balance the different number of questions for different domains, for example to avoid the larger number of items related to anxiety to overshadow depression ratings. The scale ranges from 0 to 20 points. A lower score means less mental health symptoms overall. As such, a lower score is a better outcome.

Quality of life will be measure be the 12-item Brunnsviken Brief Quality of life (BBQ-12), asking about importance and fulfilment of six areas (e.g. spare time quality, creative work, and friendship). The scale ranges from 0 to 96 points. A lower score means lower quality of life. As such, a higher score is a better outcome.

The ORS is a four-item measure designed to assess areas of life functioning known to change as a result of therapeutic intervention. These include symptom distress, interpersonal well-being, social role, and overall well-being, measured by four visual analogue scales. The scale ranges from 0 to 40 points. A lower score means lower satisfaction with life areas. As such, a higher score is a better outcome.

The Session Rating Scale (SRS) is a four-item visual analogue scale designed to assess key dimensions of effective therapeutic relationships. These include respect and understanding, relevance of the goals and topics, client-practitioner fit and overall alliance. The scale ranges from 0 to 40 points. A lower score means lower ratings of the therapeutic alliance. As such, a higher score is a better outcome.

9-items previously used in PCBH-settings (self-report), and four items from the Client Satisfaction Questionnaire (CSQ). The scale ranges from 0 to 57 points. A lower score means less satisfaction with care. As such, a higher score is a better outcome.

Experienced negative/Adverse Events, where the worst and most probably care-induced event is more thoroughly described and rated on severity and perceived cause

11 items asking patients how often (number of days each week) they engage in behaviors related to maintaining physical and mental well-being, such as physical exercise and having adequate sleep hygiene. The first 10 items are summarized as an average where 0 is minimum and 7 is maximum. Higher scores indicate more well-being behaviors and better outcome.

The same as Primary Outcome, but with a long-term time-frame. The scale ranges from 0 to 32 points. A lower score means better functioning. As such, a lower score is a better outcome.

The Trimbos and iMTA questionnaire on Costs associated with Psychiatric Illness (TIC-P) is a validated patient-reported outcome measure concerning the utilization of medical care and productivity losses. It does not have a minimum or maximum score but collects information that is needed for doing health economic calculations. This includes outcome measures such as amount of days on sick leave, productivity loss when at work, productivity loss at home, and medications prescribed and taken. Patient self-report will be complemented with register-level data on sick leave and medications.

Amount of bookable appointments each week per full time psychosocial resource (personnel in training counted as 75% of full time). An indirect measure of availability to psychosocial resources. Example: A PCC have 2 full time psychosocial resources. During a week, they together have 18 bookable appointments. The "Availability of bookable appointments" of the PCC will then be 18/2 = 9.

Number of available time slots for appointments in the coming 4 weeks divided by total number of time slots in the caregiver's calendar, an indirect measure of availability

Percentage of patients presenting with a new primary reason for visit or more than 6 months since last visit

The number of patients seen per full time equivalent mental health professional for at least one visit in relation to all patients listed at the PCC

The number of patients seen per full time equivalent mental health professional for at least one visit in relation to all patients with an ICD-10 diagnosis of F10-99, R40-45 or Z55-65

To measure PCBH fidelity and competence among psychosocial resources This questionnaire, self-rated by psychosocial resources, is not yet validated and has been created by the research group, as previously existing questionnaires have been difficult to apply to the Swedish healthcare system. The questionnaires will be validated parallel to this study. The scales include subscales to measure several specific areas of interest, such as collaboration between psychosocial resources and medical personnel. Summarized as percentage of perfect Fidelity, where higher scores means better Fidelity.

To measure PCBH fidelity and competence among leaders of PCC:s This questionnaire, self-rated by leaders, is not yet validated and has been created by the research group, as previously existing questionnaires have been difficult to apply to the Swedish healthcare system. The questionnaires will be validated parallel to this study. The scales include subscales to measure several specific areas of interest, such as collaboration between psychosocial resources and medical personnel. Summarized as percentage of perfect Fidelity, where higher scores means better Fidelity.

To measure PCBH fidelity and competence among nurses This questionnaire, self-rated by nurses, is not yet validated and has been created by the research group, as previously existing questionnaires have been difficult to apply to the Swedish healthcare system. The questionnaires will be validated parallel to this study. The scales include subscales to measure several specific areas of interest, such as collaboration between psychosocial resources and medical personnel. Summarized as percentage of perfect Fidelity, where higher scores means better Fidelity.

To measure PCBH fidelity and competence among General Practitioners This questionnaire, self-rated by General Practitioners, is not yet validated and has been created by the research group, as previously existing questionnaires have been difficult to apply to the Swedish healthcare system. The questionnaires will be validated parallel to this study. The scales include subscales to measure several specific areas of interest, such as collaboration between psychosocial resources and medical personnel. Summarized as percentage of perfect Fidelity, where higher scores means better Fidelity.

The AIM-IAM-FIM measurement with the three subscales Acceptability of Intervention Measure (AIM), Intervention Appropriateness Measure (IAM) and Feasibility of Intervention Measure (FIM) to measure implementation constructs acceptability, appropriateness and feasibility of PCBH implementation. Each subscale includes 4 items rated 1-5 each subscale have thus a min of 4 and max of 20, while the total scale has a min of 12 and max of 60. Higher scores indicate better outcome.

The original British instrument the Normalization Process Theory Measure (NoMAD) is based on the four core constructs of the Normalization Process Theory: Coherence, Cognitive Participation, Collective Action, and Reflexive Monitoring. They represent ways of thinking about implementation and are focused on how interventions can become part of everyday practice. The s-NoMAD is the validated Swedish version of the NoMAD. It includes 20 questions about the intervention, corresponding to the four constructs of, with Coherence and Cognitive Participation having four items each (each scale with a min of 0 and a max of 16), seven items for Collective Action (with a min of 0 and a max of 28), and five items for Reflexive Monitoring (with a min of 0 and a max of 20). Higher scores means better outcome

Subscales Meaning of Work, Work Life Conflict, Quality of Work, Work Engagement, Quantitative Demands and Sense of Community at Work from the Copenhagen Psychosocial Questionnaire III (COPSOQ-III)

A measure of work strain. Total score summarized as an average of all items, presented as a percentage form 0% to 100% where higher scores means less strain/stress and more satisfaction/health and thus represents a better outcome.

Inclusion Criteria: All patients from age 18 who seek care at the PCC, who are deemed to be suitable for Behavioural Health interventions and booked to the mental health professionals at the PCC, according to screening methods and/or clinical assessments made by health care personnel at the PCC, will be included. This broad criteria reflects the naturalistic setting where decisions of clinicians, rather than highly standardized criteria, are the basis for inclusion. Exclusion Criteria: Does not speak Swedish well enough to fill out questionnaires. Is in need of emergency type care, like with suicidal ideation or behaviours, ongoing psychosis or mania.