Family physicians’ views on participating in prevention of major depression. The predictD-EVAL qualitative study

Background The predictD intervention, a multicomponent intervention delivered by family physicians (FPs), reduced the incidence of major depression by 21% versus the control group and was cost-effective. A qualitative methodology was proposed to identify the mechanisms of action of these complex interventions. Purpose To seek the opinions of these FPs on the potential successful components of the predictD intervention for the primary prevention of depression in primary care and to identify areas for improvement. Method Qualitative study with FPs who delivered the predictD intervention at 35 urban primary care centres in seven Spanish cities. Face-to-face semi-structured interviews adopting a phenomenological approach. The data was triangulated by three investigators using thematic analysis and respondent validation was carried out. Results Sixty-seven FPs were interviewed and they indicated strategies used to perform the predictD intervention, including specific communication skills such as empathy and the activation of patient resources. They perceived barriers such as lack of time and facilitators such as prior acquaintance with patients. FPs recognized the positive consequences of the intervention for FPs, patients and the doctor-patient relationship. They also identified strategies for future versions and implementations of the predictD intervention. Conclusions The FPs who carried out the predictD intervention identified factors potentially associated with successful prevention using this program and others that could be improved. Their opinions about the predictD intervention will enable development of a more effective and acceptable version and its implementation in different primary health care settings.


Introduction
Interventions to prevent depression are effective but their effect sizes are small to moderate [1][2]. These interventions are primarily psychological or educational and are provided by mental health professionals [2]. Several randomized controlled trials of primary prevention of depression in primary care have been undertaken [3], although in only two the intervention was implemented by family physicians (FPs). In a third trial, (the "CATCH-IT" study) [4] the effectiveness in adolescents of brief counseling was compared with a motivational interview administered by FPs. While both interventions seemed effective, there was no evidence either way for superiority. Our research team carried out the predictD-CCRT study to compare a bio-psycho-social intervention implemented by FPs (the predictD intervention) versus usual care [5]. The predictD intervention is based on a risk algorithm to predict the onset of major depression at 12 months in primary care attendees [6]. This intervention was tailored to each patient based on his/her risk profile for depression (risk factors present) and his/her risk level (likelihood of becoming depressed at 12 months), and it was developed as five a priori active components [5][6][7]: a training workshop for FPs; communicating the level and profile of risk of depression to patients every six months in a 10-to 15-minute interview; constructing a personalized bio-psycho-social intervention to prevent depression; offering a brochure; and activating and empowering patients. This intervention reduced the incidence of major depression at 18 months follow-up by 21% versus the control group (usual care) [7] and it was also costeffective [8]. Qualitative methodology is the best way to elucidate the active ingredients/ mechanisms of action of these complex interventions in order to promote them and to adapt the interventions for application in different settings [9]. Therefore, the aim of our study was to assesss FPs' opinions about the potential successful components of the predictD intervention for the primary prevention of depression in primary care and to identify areas for improvement.

Design
A qualitative study using face to face, semi-structured interviews was undertaken. Our ontological position is relativism (i.e. the view that reality is subjective and may differ from person to person). Our main aim is to understand how FPs have experienced the implementation of the predictD intervention, from their individual´s perspective, taking into account the context in which it had been implemented. The methodological approach used was phenomenology.

Setting
The study population comprised FPs working at 35 urban primary care centres in seven Spanish cities (Barcelona, Bilbao, Zaragoza, Valladolid, Jaén, Granada and Málaga). The Spanish National Health Service provides universal coverage for citizens through a public system financed by taxes and is free at the point of use. Health care services are distributed into Health Areas and Basic Health Zones according to geographical, epidemiological and socioeconomic criteria. Each Health Area covers a population of 200,000-400,000 inhabitants and is composed of several Basic Health Zones, which are the minimum units of health care organization. Basic Health Zones are organized around a primary care centre. Primary care centres serve populations of between 5,000 and 30,000 inhabitants and the primary care teams are composed of FPs, pediatricians, nurses and, in some cases, social workers. They provide a broad range of services, including the treatment of common mental disorders (shared with mental health specialists in severe cases) such as anxiety or depression [10], health promotion and preventive services. In Spain, primary care is the patient's first point of contact with the public health system. Each patient is assigned to only one FP, who functions as a gatekeeper to the wider system. Patients can visit their FP as often as they want without having to pay for consultations, even when they do so for preventive reasons. The average time that a Spanish FP spends per patient is between 8 and 10 minutes.
All the primary care centre staff members, including the GPs, are salaried. GP salaries contain two elements: a larger fixed payment and a smaller incentive, based on elements such as numbers of patients assigned, fulfilment of objectives, patterns of prescription and pay-for performance incentives [11].

Sample selection
For inclusion, the FP had to be one of the 70 FPs who had performed the predictD intervention and completed the follow-up at 18 months.

Procedure
The interviewers were experts in qualitative interview techniques. Research assistants contacted the FPs who were informed of the date, time and place of the meeting but did not explain the study objective in detail. The FPs were only told that the purpose of the meeting was to seek their views on their participation in the predictD-CCRT study. None of the FPs knew the results of the effectiveness [7] or cost-effectiveness [8] of the predictD intervention when they were interviewed. The interview guide was piloted with one FP. The topic guide used in the interviews is shown in Table 1. The interviews were conducted between March and April 2013 and lasted 20-60 minutes. The interviews were all conducted in a place convenient to the FPs (quiet areas of primary care centres) and were audio-recorded digitally, transcribed and anonymized.

Analysis
We used thematic analysis [12]. Using the transcriptions and to ensure data quality, the information obtained was triangulated by the participation of three analysts (PMP, AF, SCC) from different cities and professional backgrounds with experience in using qualitative research who independently analyzed the interviews [13]. These three researchers became familiar with the interviews by listening, reading and rereading them. Themes were identified and manually coded independently by each of the three researchers involved in the analyses. They then came together to compare and discuss differences in the analyses. Themes were then recoded and classified, identifying common patterns and convergences and divergences in the data through a process of constant comparison. In order to enhancing the quality of this research [13][14], the interviewers and analysts kept a personal research diary alongside the data collection and analysis to record any reactions to events occurring during the research. Respondent validation was conducted, comparing analysts' interpretation of the phenomena with that of those who had participated. Participants were sent a summary of the findings. Forty FPs participated in this validation. Changes suggested by the FPs were incorporated into the final description of the phenomena.

Ethics statement
All participants gave written informed consent and participant anonymity was maintained using personal codes in the transcripts. The study was approved by the ethics committees in each participating Spanish city.

Results
The 70 FPs participating in the intervention arm of the predictD-CCRT study were invited to take part in the predictD-EVAL study. Of these, 67 (95.7%) agreed to participate. The remaining three did not participate because they did not complete the predictD-CCRT study: two due to sick leave and the third due to relocation to another health centre. The characteristics of the FPs are shown in Table 2. The FPs' opinions were described in six broad thematic categories.   The FPs perceived the information on the level and profile of risk of the patients, the brochure/advice and the training as mechanisms by which patients increased their self-knowledge, satisfaction and empowerment, therefore, improving their mental health. Additionally, these components fostered a relationship of trust and mutual understanding between the FPs and the patients and prompted the FPs to pay closer attention to the psychosocial aspects of their patients, favoring the knowledge of aspects of their patients, leading to increased satisfaction in the FPs. In consequence, the FPs spent more time with their patients, put into practice communications skills and problem-solving strategies, activated resources available to each person and involved social workers, while at the same time they tried not to get too involved themselves. All this, from the perspective of the FPs, resulted in the prevention of depression (Fig 1).

Summary
This study reports on the opinions of the FPs concerning facilitators and barriers for the implementation of a personalized program, based on a risk algorithm, to prevent the onset of major depression in primary care. This information can be used to improve the predictD intervention, and to ease its adaptation and implementation to different settings. In general, the FPs had a positive experience with the predictD intervention, as it was easily embedded into their practice. It was perceived useful as a bio-psycho-social approach for improving the emotional health of their patients and their relationship with them, as well as their own satisfaction as a FP. However, they also detected some barriers such as lack of time, and the need for specific training to effectively communicate the risk of developing depression.

Strengths and limitations
To our knowledge, this is the first study that explores the opinions and experiences of FPs who performed a personalized intervention to prevent major depression in primary care. The opinions of several of the FPs have been contrasted (see below) with the data available on the effectiveness and cost-effectiveness of the predictD intervention [7][8]. We emphasize that none of the FPs knew the results of the effectiveness or cost-effectiveness of the predictD intervention when they were interviewed for the predictD-EVAL qualitative study. The research team was multidisciplinary (family physicians, nurses, psychologists, psychiatrists and social workers), the approach was oriented, and the data were analyzed and interpreted from different professional perspectives thereby enriching the discussion and interpretation of the information. In addition, the data were triangulated for the elaboration of categories by three analysts. Respondent validation was conducted as part of a process of error reduction to establish the level of correspondence between researcher and research subjects.
There are also some limitations: First, some of the information derived from the FPs' opinions would have been even more useful if it had been contrasted with the opinions of the patients. Even so, we know that most patients believed that the predictD intervention was acceptable (76%) and that patient adherence to the intervention was also good (90% of the patients attended at least two of the three planned intervention visits) [7]. Second, in some cases one year had elapsed from the time the FPs performed the interventions until they were interviewed. This time period may have been too long, affecting the memory of the FPs. Third, it is possible that the FPs gave more positive views about the intervention in order to satisfy the interviewers who were members of the research team for the study. Fourth, the FPs participating in this study possibly had a greater psychosocial and preventive orientation than the total of FPs because many FPs declined to participate in the trial [7]. Therefore the opinions of the FPs in this study would have limited its generalization. Most physicians tend to use a biomedical rather than a patient-centered communication style [16], and FPs with lower psychosocial orientation scores have poorer doctor-patient relationships [17].

Comparison with existing literature
Previous acquaintance with patients and the doctor-patient relationship, the establishment of a basic psychotherapeutic relationship (not a psychiatric interview or formal psychotherapeutic interventions) [18], a family-oriented practice [19][20], social prescribing and community referral by FPs [21], and management of physical problems are all components of care or strategies commonly used by FPs in their clinical practice. The fact that FPs consider some aspects or components of the predictD intervention to be similar to what they already do in their routine practice favors their acceptance and implies a greater perception of self-efficacy in the role of the FP as facilitator in the predictD intervention and suggests a greater likelihood that they will be confident applying the predictD intervention. These are also strategies requested by primary care patients for prevention and the promotion of healthy habits [22].
In general, the FPs were satisfied with the training workshop for the predictD intervention; however, some pointed out the need for more extensive and continuous training. This might have been because a number of FPs had some difficulty in implementing the predictD intervention in the first consultations. The reduction in the onset of major depression attributable to the predictD intervention seemed to increase over time, which might be due to a doseresponse effect of the intervention or simply a need for time and the accumulation of intervention visits to create the changes needed to prevent depression [7]. This would be in accordance with the opinion of the FPs regarding how their skills continued to improve as they had more visits-interventions with the patients [9]. Short-term training (less than 10 hours) is as successful as longer training Interventions for providers to promote a patient-centered approach in clinical consultations [23]. Perhaps an increase in training hours in the predictD intervention may not be necessary, but a review to improve the efficiency of training in those 10-15 hours is indicated or adding a booster after a given period.
FPs emphasized that the predictD intervention contributed to improving the doctor-patient relationship and patient confidence. Patients who gain a greater insight into their symptoms and feel understood by their physicians may be less anxious and depressive, have more confidence in their physician's abilities, and be more trusting of their physician [24]. Furthermore, sharing of power and responsibility, the use of empathy, and treating the patient as a person all seem to be important FP communication strategies which help address barriers to completion of preventive services by patients [25].
Lack of time is a barrier widely reported by FPs, especially when undertaking preventive activities [17]. Some FPs considered that the predictD intervention generated a higher time and workload and others felt the opposite. In any case, objective data showed that the patients who received the predictD intervention, during the 18 months of follow-up, made fewer visits to the primary care centre than those in the control group, and this happened despite the patients in the intervention group having performed the three specific and mandatory visits (one every six months) of the predictD intervention [8]. In observational studies, the association between depressive and anxiety disorders and frequent primary care attendance has also been described [26][27][28].
Some FPs suggested that patients' knowledge of their level and profile of risk of depression can cause alarm, fear and even the onset of depression or anxiety. However, the predictD intervention reduced the incidence of both depression and anxiety [7]. Moreover, in a previous study, primary care patients were asked if they would like to know their level and profile of risk of depression, and they generally displayed a positive attitude and were willing to be informed of their risk of depression [29]. In addition, FPs reported no adverse effects associated with the predictD intervention, and only three patients (0.0018%) in the intervention group contacted researchers, during the 18 months of follow-up, to complain about relationships with their FPs and to request a change of FP [7].
A number of FPs expressed some difficulty in getting their patients to understand their profile and level of risk for depression. That also happens, for example, in cardiovascular risk management [30]. How risk is communicated influences prevention [31], so improving clinicians' skills to communicate it efficiently is a challenge that needs further evaluation.
The brochure delivered to patients was positively valued by FPs. The FPs may feel more secure in being able to provide information in a standardized way. However, it would have been useful to have collected the patients' opinions in order to contrast this with the positive opinion of the FPs.

Conclusion
The FPs suggested ways to adapt the predictD intervention to other settings and make its implementation more efficient. Some proposed prioritizing the 'constructing a tailored biopsycho-family-social intervention by FPs to prevent depression' component of the predictD intervention in patients at higher risk of depression. This is relevant to whether or the predictD intervention was more clinically and cost-effective in high or low-moderate risk patients. Other FPs suggested involving other primary care professionals (e.g. nurses) and minimizing the time to obtain information on the level and profile of the risk of depression. Our research team launched a free website (http://www.predictplusprevent.com/index.php?idioma=en) that anyone can access to calculate their risk of depression, anxiety or hazardous/harmful alcohol consumption. The website also provides information and resources on how to reduce risk. Some FPs, in addition to stepped care depending on the level of risk, also suggested specific interventions for patients with particular risk factors for depression (e.g. for sedentary or insomnia patients).
The predictD intervention requires little adaptation because it is based on usual components of primary care that could converge in the prevention of depression, and there now are validated risk algorithms in several European countries [32], the United States [33][34][35], Australia [36] and Latin America [37].
If depression prevention programs were massively scalable (e.g. through ICTs) and/or implemented in large populations (primary care, schools, and workplaces), they would have a considerable impact on patient health and quality of life and cost reduction, even were their effectiveness relatively low [1][2][3]. In this sense, a paradigm shift is needed [38]. This study provides key information on which to build an improved predictD intervention, enhancing factors considered useful by FPs and minimizing identified barriers. However, in order to achieve a more effective and acceptable intervention, the opinions of the primary care patients who received the predictD intervention are needed.
Supporting information S1 File. Interview guide used in the study in Spanish. (PDF)