Psychiatric Consultation-Liaison Intervention in Primary Care: A Pilot Study

The pilot study on the impact of a Psychiatric Consultation-Liaison Intervention in Primary Care Medical Consultations in French-speaking Switzerland (PLIMeC-P), is a mixed method randomized controlled trial. The investigated health-related intervention is a brief Consultation-Liaison (CL) psychiatry intervention in primary care. Primary Care Physicians (PCPs) have a key role in preventing, detecting, and managing mental disorders. Therefore, the optimization of both the quality of their interventions and their cooperation with psychiatric and psychological care networks are important challenges. It is well demonstrated that multidisciplinary interventions improve the efficacy of mental health care, CL psychiatry being one such type of intervention. Therefore, community CL-psychiatry, for example in private primary care group practices, are rare. The impact of such community, primary care CL-psychiatry interventions, should be investigated. The mixed methods randomized controlled trial PLIMeC study will examine the impact of a CL-psychiatric intervention in primary care settings, for newly reported mental illness, versus a Treatment As Usual (TAU) control arm. The intervention consists of a CL-psychiatric intervention into private medical practices, provided to patients suffering from mental health problems, a group of patients generally under-treated. The CL-psychiatrist will receive and discuss PCPs' referrals for patients with mental distress, who don't have a psychiatric follow-up. After a brief intervention (1-4 sessions), feedback conjoint (PCP-psychiatrist) session will be organized, to complete the intervention and provide proposals. The pilot study (PLIMeC-P) will determine whether the main planned study (PLIMeC) is feasible and practicable. It will be conducted on two sites, the Neuchâtel Psychiatry Centre (CNP) and the North-west Adult Psychiatry Service (SPANO), Department of Psychiatry of CHUV, Yverdon. For the pilot study (PLIMeC-P), 15 eligible participants are expected for each group, 30 participants in total. They will be recruited in three private primary care group practices. After eligibility and randomization 15 participants will be included for the intervention arm (brief CL-psychiatric intervention) and 15 for the control arm (TAU). The expected number of participants for the main trial (PLIMeC) will be estimated through analysis of the initial results of the PLIMeC-P.

1. BACKGROUND AND RATIONALE According to the WHO, mental health is a central component of personal health and well-being. Mental disorders claim a heavy toll on individuals, their families and society, and are a major public health issue. In Switzerland, 18-21% of the population suffers from psychological distress, of which 12.7% have been diagnosed with a psychiatric disorder. Primary care physicians (PCP) are usually the first contact with healthcare for patients suffering from psychological distress or mental disorders. More precisely, a prevalence of 25-60% is reported for mental disorders in primary care, and this is the case worldwide, as well as in Switzerland. More than half of those patients have no specialized psychiatric care and PCPs tend to base their treatment on pharmacotherapy, support therapy, and psycho-education. PCPs are therefore key stakeholders in the detection, prevention, and adjusted treatment management of mental disorders. Therefore, the optimization of both the quality of primary care interventions and cooperation with psychiatric care networks are important challenges. Patients with mental disorders are often considered complex patients and tend to be frequent care users, PCPs feeling lacking competency in dealing with them. Close collaboration between psychiatrists and PCPs, and multidisciplinary interventions can reinforce PCPs' skills and help them gain confidence in managing mental disorders. A wide range of collaboration models between psychiatrists and PCP exist, the Consultation Liaison (CL) interventions being one of them. However various barriers are described regarding such a collaboration, notably because of different treatment paradigms and of lack of accessibility. In Switzerland, CL-psychiatric interventions are mainly used in hospital settings. To introduce a model of care that offers optimal management of mental disorders in primary care, ensuring an adequate continuity and good coordination of care, for patients with complex needs, a multidisciplinary collaboration project between PCP and CL-psychiatrists has been developed in a region of French-speaking Switzerland. A CL-psychiatric consultation within primary care practices has been implemented, to promote the quality of mental health care and facilitate communication between healthcare professionals. The need for further research on the implementation of psychiatric interventions in primary care is clearly highlighted. Moreover, the experience of patients, of PCPs and of psychiatrists during such interventions is rarely investigated, whereas emotional experiences seem to present one of the central obstacles to the optimal management of mental disorders in primary care. Our study will investigate the impact of a CL-psychiatric intervention for patients with mental health problems in private primary care group practices, in French-speaking Switzerland, using mixed methods (qualitative and quantitative). In parallel, the lived experience of PCPs and CL-psychiatrists will be investigated during their collaboration, in order to better understand the factors that influence this type of intervention and to better qualify its results. Initially, a pilot study (PLIMeC-P) of the multi-center, multi-method PLIMeC study will be conducted, to determine whether the PLIMeC study is feasible. 2. STUDY OBJECTIVES AND DESIGN 2.1 Hypothesis and primary objective 2.1.1 For the Pilot study (PLIMeC-P) : To determine whether the planned multicentric mixed method design randomized controlled trial is feasible and practicable. • Essential activities in the PLIMeC-P will include testing procedures for assessment methods; conducting initial semi-structured individual interviews to test and adjust the moderator's guide (questions concerning PCPs, CL-psychiatrists and patients lived experience of the intervention); first evaluation and analysis of initial results. These activities will be carried out through ongoing collaboration between the study sites. 2.1.2 For the PLIMeC main study: The primary objective is to compare the efficacy of a CL psychiatry intervention (case discussion, brief intervention, and conjoint feedback session - intervention arm) introduced in a primary care setting, with treatment as usual (TAU - control arm), for patients suffering from mental health problems. More precisely, the investigators will focus on the effect of such an intervention on the psychiatric symptoms and the quality of life of participants. The investigators hypothesize that a significant clinical change will be produced in the intervention arm in comparison with the control arm. The secondary objective is to explore qualitatively the participants' lived experience of the management of their mental health problems in primary care and the effect of a CL psychiatry intervention on this (as well as on doctor-patient relationship), and, in parallel, the PCPs' and psychiatrists' lived experience of their collaboration (barriers and/or facilitators, needs, affects, thoughts etc.). An exploratory objective is to investigate aspects of public health efficiency, in order to answer the question of at what level to best meet psychiatric outpatients and to where best allocate psychiatric resources. The investigators expect a high level of complexity in the basis evaluation of the clinical situations and hypothesize that healthcare professionals' collaboration will ameliorate with the CL psychiatry intervention. 2.2 Endpoints are detailed in the Outcome Measures session. 2.3 Study design The PLIMeC multicenter study is a randomized controlled trial, with a mixed methods design (i.e., using both quantitative and qualitative assessment), including a control arm (CL-psychiatry intervention), and an intervention arm (Treatment as Usual - TAU), in primary care settings in which the intervention of a CL-psychiatrist was implemented. A potential problem associated with the proposed study design could be the specificity of different local primary care settings (different regions, different organization, different clinical experiences of PCPs, different sizes, etc.). To minimize potential bias, first, both arms will be studied in the same primary care settings, with patients naturally seeking help at primary care practices and being randomized. Second, the investigators will use complementary qualitative data to further interpret the quantitative data (employ a mixed-method design). Finally, a variety of private primary care practices, with different demographic variables, will be included in a large region of French-speaking Switzerland for the main study (already 3 different private primary care practices, in different regions of two different cantons for the pilot study). The employment of validated questionnaires, which will be monitored during the pilot study, will also ensure the quality of our data. In addition, the investigators are aware that the sensitivity of PCPs to mental health problems, and their training in these issues, may vary from one practitioner to another. Nevertheless, such heterogeneity characterizes the current state of practice. Consequently, the investigators do not consider this heterogeneity to be a source of bias, as the aim of the PLIMeC study is to estimate the effect of the intervention in the current healthcare context (real-world data). On the other hand, PCPs working in practices that agreed to take part in the study may be on average more sensitive to education about mental health issues than their colleagues not taking part in the study, thereby constituting a recruitment bias. In order to reduce this bias, with the pilot study (PLIMeC-P) a first analysis will be conducted of the different private primary care practices (3 practices), in different regions, with the participating PCPs. The investigators will identify the level of training and/or awareness of mental health issues among the PCPs included in the study (demographic elements collected before the start of the study) to identify trends, to better calibrate the main study, and carry out sample size calculations. Regarding the qualitative part of the design, a potential problem could be the fact that the local PI supervises the clinical project. To minimize potential bias, reflexivity will be applied during the research process, and the investigators will ensure that the research assistant who will conduct the individual semi-structured interviews will not be directly involved in the clinical project and will have no hierarchical link with the participating PCPs and CL-psychiatrists. Moreover, to ensure the quality of our qualitative data the investigators will use a moderator guide drawn upon three previous qualitative research studies: a) a study on the referral process, as experienced by PCPs; b) a study on psychiatrist/PCP discussion groups, with a sociological perspective (observation in situ); c) a qualitative study on the experience of PCP regarding consultation-liaison psychiatry interventions in primary care. Finally, the interview guide will be monitored during the PLIMeC-P. 2.4 Regarding the study intervention, see Arms and Interventions session. 3. STUDY POPULATION AND STUDY PROCEDURES 3.1 Study population: The study population will be composed of patients who present mental distress in primary care consultations. They will be part of the PCP patient list (new or regular patients). For the PLIMeC-P, two private primary care group practices, in the canton of Neuchâtel (in two distinct regions), and one private primary care practice, in the North Vaud (in Yverdon), will be mobilized, a total of three primary care practices, in three different regions. Participants will be indicated by their PCP working in these practices and will be enrolled by the local PIs or delegates. More precisely, firstly, the patient will be orally asked to consent in order that his/her details are communicated by the PCP to the respective local PI or delegates. Afterwards, the local PI or his/her delegates will hold recruitment, if the inclusion criteria are met and if there are no exclusion criteria, after transmitting the necessary information concerning the study to their patients. Interested patients will be asked to review the information and consent form and sign it if they consent to participate in the study. For Inclusion and Exclusion criteria, see Eligibility session. For the PLIMeC-P, 15 eligible participants are expected for the control arm (TAU) and 15 eligible participants for the intervention arm, with a total number of 30 participants. The investigators will include 10 participants from each of the three sites (5 participants in each arm, per site). In parallel, part of the participating patients (3 from TAU arm and 3 from the intervention arm), who will be chosen for their mixed sociodemographic characteristics, will be asked to participate in individual semi-structured interviews, in order to explore qualitatively their perception regarding usual mental health care and the proposed intervention. Moreover, the participating PCPs and psychiatrists will also be invited to undergo individual semi-structured interviews for the qualitative part of the study, in order to obtain a global understanding of the impact of the intervention. They will decide if they want or not to participate, having all the liberty to refuse the invitation. No health-related data will be collected. No compensation or payment is intended to be given to the participants or the collaborating PCPs. 3.2 Study procedures Demographic data and baseline questionnaires will be addressed to the participants together with the informed consent by email or Post (depending on the participant's choice). Only participants completing the baseline assessment will be included in the study. The baseline questionnaire assessment lasts around 15 minutes. It will include demographic variables and the PHQ (PHQ-9, GAD-7, PHQ-15), WHOQOL-BREF, and INTERMED self-assessment scales. The PHQ is based on five modules covering common types of mental disorders: 1. depressive, 2. anxiety, 3. somatoform, 4. Alcohol, and 5. Eating disorders. The first three modules (PHQ-9, GAD-7, PHQ-15) screen symptom severity with a threshold of 10. The WHOQOL is a quality-of-life instrument containing 26 items, each is rated on a scale of 1 to 5. The score is then transformed linearly to a 0-100 scale with a cut-off \> 60. The INTERMED self-assessment is an instrument to determine the relationship between biopsychosocial complexity and healthcare and social cost; and containing 20 questions, each is rated on a scale of 0 to 3 with a maximal score of 60 and a threshold of 21. Only the subscale regarding healthcare system complexity (Questions 13,14,15,16,20) will be included in the baseline questionnaire. After the final eligibility check and completeness of the baseline questionnaire, participants will be attributed a study code. The key linking the participant's name and the study identifier will be stored at each site. The enrolled participants will be then randomly assigned by the local PIs or delegated collaborators, to the control or intervention arm. They will complete follow-up assessments for both arms, at 3, 6, and 12 months (T1, T2, and T3) post baseline. The follow-up questionnaire will include the Patient Health Questionnaire Somatic, Anxiety, and Depressive Symptom Scales (PHQ-15, GAD-7, PHQ-9) and WHOQOL-BREF self-assessment scales and the questions 16 and 20 of the INTERMED self-assessment scale. Similar procedures will apply. Depending on the participant's choice, they will receive either an email or a paper questionnaire by Post at 3, 6, and 12 months. Regarding qualitative assessment, for the PLIMeC-P, 6 participants of both arms (control and intervention arm, 3 and 3 participants), will be invited to participate in one-to-one semi-structured interviews. In parallel, PCPs and CL-psychiatrists will also be asked to participate in individual semi-structured interviews. Interviews will be conducted by phone or face-to-face, depending on participants' preference, and they will be audio-recorded. A qualitative assessment will be conducted by a research collaborator. 4. STATISTICS AND METHODOLOGY To answer the research question of the PLIMeC-P (Pilot study), aiming to determine the feasibility and practicability of the main PLIMeC study, in the initial phase, the statistical procedures will be limited to questions of practicability and quality. For the main PLIMeC study, the investigators will use a mixed-method approach, combining various data collection and analysis methods (qualitative and quantitative). Mixed methods research links elements from both qualitative and quantitative paradigms to produce converging findings in the context of complex research questions, such as the psychological impact of an exceptional situation, as investigated in our study. This combination can provide a better understanding of such research issues than either approach alone. On the one hand, quantitative methods will be used to obtain a breadth of understanding of the phenomenon of interest, and on the other hand, qualitative methods will be used to explore and obtain a depth of this understanding. More specifically, in our study the combination of methods is used for complementarity, using qualitative data to examine how the participants, patients, and healthcare providers, experience this specific CL-psychiatric intervention and quantitative methods to examine the impact of such an intervention on patients. A fixed mixed methods design will be used; quantitative measurements will be firstly carried out and afterward qualitative ones. In data analysis, the priority among the two methods will be equal, and the nature of integration of partial, qualitative findings expands on the first-phase quantitative results, at the same time helping their interpretation through data convergence. The potential added value of mixed methods includes producing greater insight than a single method could, improving quantitative data interpretation, and enhancing the external validity of these findings. 4.1. Statistical analysis plan: Statistical procedures will be limited to practicability and quality issues in the initial phase. The first statistical and qualitative analyses performed in this pilot study (PLIMeC-P) will also be used to obtain a more accurate estimate of expected observable differences between groups, help in defining the number of participants and clusters (i.e. centres) required for the main study and in adjusting the moderator's guide. 4.2 Regarding the main PLIMeC study: 4.2.1 Quantitative analysis plan: The investigators shall consider a total of 4 primary outcomes, consisting in continuous scores measured on the PHQ-9, GAD-7, PHQ-15, and WHOQOL-BREF (psychological domain). The longitudinal follow-up scores will be modeled using a linear mixed-effect regression. The level of statistical significance in this study will be fixed at 5%. However, since a total of 4 primary outcomes will be investigated, the statistical significance of the intervention effects in each outcome at a given timepoint will be corrected for multiple testing using the Holm-Bonferroni procedure. Sample size calculations for the main phase will be precised on the basis of results from the pilot study. Regarding each of the 3 subscales of the PRIME-MD, a difference of more than 5 points will be deemed as clinically relevant. Regarding the WHOQOL-BREF, it reports a standard deviation of 2.8 points for the typical inter-individual variability in scores on the psychological domain. Without further knowledge at this point, the investigators shall postulate a moderate effect of the intervention on each outcome e.g. at T3, with a standardized mean difference D=0.5 according to Cohen's definition of effect sizes (0.2=weak, 0.5=moderate, 0.8=large). Following calculations in Vierron and Giraudeau for multicenter randomized trials and allowing for a family-wise error rate of 5% while assuming an intraclass correlation of 0.1, a minimum of 81 individuals should be included per arm (162 total) in order to reach 80% power. This sample size should be further inflated to account for dropouts. As mentioned above, in all cases, this sample size will be re-evaluated following the pilot study results and, if necessary, modified by means of a protocol amendment before the start of the main study, and results from the pilot study will be used to estimate the dropout rate on each occasion. For the exploratory objective, the investigators will investigate the healthcare system complexity of clinical situations, in the basis evaluation (T0), measured by the INTERMED Self-Assessment Version scale (Questions 13,14,15,16, 20) and the question 24 of the WHOQOL-Bref questionnaire, and will compare the evolution of the collaboration of healthcare professionals for participants of the intervention arm (CL psychiatry intervention) with them of the control arm (TAU), the moment they are enrolled by their PCP in the (T0), in 3 months (T1) and 12 months (T3) later; measured by the INTERMED Self-Assessment Version scale (Questions 16 and 20 and the question 24 of the WHOQOL-Bref questionnaire in T0, T1 and T3. Firstly, data will be screened for missing cases and outliers using descriptive statistics and plots. Afterwards, descriptive statistics (i.e., means and standard deviations) will be used to report the levels of healthcare system complexity of clinical situations at T0, T1 (3 months), and at T3 (12 months). Descriptive statistics will also be used to report frequencies of healthcare system complexity of clinical situations at T0, T1, and T3 following screening guidelines as proposed by the authors of the scales. 4.2.2 Qualitative analysis plan: The secondary objective is to explore qualitatively the lived experience of participating healthcare staff (PCPs and psychiatrists) and patients of this multidisciplinary intervention (collaboration), including their impact on the management of mental disorders, complexity, barriers, and encountered difficulties for the first, and whereby such an intervention has an effect on the doctor-patient relationship and how it responds to the mental health care needs for the latter. Qualitative semi-structured interviews will be conducted by local co-investigators in each center (CNP and SPANO). The interviews will be audiotaped, transcribed, coded, and analyzed using ATLAS.ti software. The analysis of the coded interviews will be based on thematic content analysis and will be carried out by the local co-instigators and the Sponsor-PI, initially in parallel, without one seeing the categories created by the other from the same raw data. These categories will then be compared to establish the extent of possible overlap in the categories. Thus, these different sets of categories will then be combined or reorganized. Once the new categories have been established, a clarity check will be conducted by external stakeholders (primary care physicians and psychiatrists) and their feedback will be incorporated for further revision and refinement of the categories. Another objective is to deepen our understanding of the complexity of the investigated healthcare systems and to better interpret the evolution of the collaboration of healthcare professionals (both for the intervention and the control arm), the investigators will also explore qualitatively: i) the patients' mental health care needs and the PCPs' mental health service needs; ii) to what extent they are satisfied or not by the intervention; iii) interaction between complexity and these needs, as also how complexity is addressed. A combination of quantitative and qualitative data will be performed, qualitative findings expanding quantitative results and at the same time helping their interpretation through data convergence. 5. Handling of missing data and drop-outs: During the study, every effort will be made to gather all the required data and guarantee good quality of data.