The purpose of this project is to improve the management of suicide and common mental disorder by general practitioners (GPs) in Bulgaria in order to reduce the suicide rate in intervention regions. The study uses a natural experiment design which utilizes a training program aimed at improving the GPs management of suicide risk and detection of common mental disorders. The training program will be offered to four regions (North East, South West, South Central, South East), leaving two regions for control (North Central, North West), in order to evaluate the effect of the intervention.

The purpose of this project is to improve the management of suicide and common mental disorders by general practitioners (GPs) in Bulgaria. The study uses a natural experiment design which utilizes a training program aimed at improving GPs management of suicide risk and detection. The training program will be offered to four regions (North East, South West, South Central, South East), leaving two regions for control (North Central, North West), in order to evaluate the effect of the intervention. Thus, the study's outcome variable will be suicide rate, while region will be study unit.

The project is a collaboration between The National Centre of Public Health and Analysis (NCPHA) and The Norwegian Institute of Public Health. In short, NCPHA is responsible for coordination, management and implementation, while The Norwegian Institute of Public Health (NIPH) will provide advice for highest possible quality and effect of the intervention. The NIPH will also evaluate the training program, which will be used to train the GPs to improve their skills in the management of suicide risk, and the recognition and treatment of anxiety and depression.

The suicide rate is higher in Bulgaria compared to several other European countries (WHO, 2015). It is well known that patients often contact their GP days or weeks prior to suicide (Luoma, Martin, & Pearson, 2002). There is also some evidence suggesting that improved management of suicide risk by GPs may successfully reduce suicide rates in the general population (Mann et al., 2005; Pfaff, Acres, & McKelvey, 2001; Rutz, Vonknorring, & Walinder, 1992).

This project will test if an improvement of skills and knowledge in general practitioners (GPs) may reduce the suicide rate in Bulgaria. There are several studies indicating that this might have positive effect on suicide rates (Hegerl, Althaus, Schmidtke, & Niklewski, 2006; Mann et al., 2005; Nock et al., 2008; Pfaff et al., 2001; Rutz et al., 1992; Szanto, Kalmar, Hendin, Rihmer, & Mann, 2007; Székely et al., 2013; WHO, 2014), however there are also studies showing limited evidence for effective suicide prevention interventions and a need for further investigation (du Roscoat & Beck, 2013; Robinson, Hetrick, & Martin, 2011).

The best method for testing effects is a randomized controlled trial. However, we are unable for practical reasons to randomize, hence the project is set up as a controlled trail, where the north east and northern part of Bulgaria serves as the control group for the intervention group being the remaining other four districts of Bulgaria. The regions are selected by the NCPHA on the basis of practical considerations, not within our control.

The aim is to deliver the intervention to 1650 GPs and 350 social workers and psychologists in these four regions. GPs in the control regions are not supposed to receive the intervention. The purpose of the trial is to explore if this intervention may reduce the suicide rate and the rate of attempted suicides in areas where GPs participate in the training program, compared to the control areas. GPs and health professionals is emphasized because with increased knowledge and skills, these professionals may be able to improve treatment and recognition of anxiety and depression in their patients, and they may also be able to improve their management of suicide risk.

The intervention will be implemented during January 2016 until June 2016, and effects will be evaluated by trajectories of suicide rates for the period 2012-2018. The suicide rate is 10.8 per 100 000 inhabitants per year according to the WHO (WHO, 2015). From this, we would expect 529 suicides in the intervention area and 227 in the control area per year, presuming no effect of the intervention, equal suicide rates in the regions, and a 70/30 distribution of the population between these regions.

Intervention: The intervention consists of two components, Phase I. Distance learning and Phase II. Seminars.

Phase I: Distance learning will be based on a web-based system with login features. The login feature will include reading material and active learning strategies, e.g. videos and tasks. There is evidence to support that active learning strategies is more effective in improving GPs' attitudes, behaviors and skills than passive learning strategies, such as provision of reading material. Some highly cited reviews (Michael, 2006; Prince, 2004; Shellman & Turan, 2006) of active learning strategies (e.g. actors role-playing patients) conclude that there is evidence that active learning strategies have an improved effect on learning compared to passive learning (e.g. lectures). In addition, active learning also receives more positive feedback and higher attendance rates than passive learning approaches. Employing active learning is not necessarily excessively time-consuming (Fenwick, Vassilas, Carter, & Haque, 2004). In fact, evidence suggest active learning is a more time-efficient approach in terms of educational outcomes for participants (Haidet, Morgan, O'Malley, Moran, & Richards, 2004). Video is an effective medium in communicating clinical skills and techniques, compared to written material only. Videos are also included with the purpose of motivating the GPs for staying in the learning program and demonstrating good clinical practice with "fake" patients. There will be features in the website checking that the GPs have actually started the video clips. The main purpose of the written material is to provide practical tools for the GPs in their clinical management. Thus, they are to include practical questions for the GP to ask the patient in the management of suicidal behavior or common mental disorders. In addition, some of the texts will be more oriented towards extending GPs' knowledge about suicide, suicidal behavior and common mental disorders. To increase the likelihood that the intervention will be effective the login feature will include the possibility for GPs to communicate with lecturers and each other by an internal e-mail system. To keep track of website activity the technical subcontractor which is responsible for the website will prepare reports on a weekly basis with information on how frequently the various parts of the website is used. The GPs are incentivized to participate in the intervention by "certificate" and "credits". To receive certificate or credits it is required that the GP has (a) viewed and opened relevant sections on the website, (b) passed a test (e.g. multiple choice), and (c) participated in the seminar.

Phase II. Seminar: The seminar will last for two and a half days. Participation will be rewarded with a course certificate. Successful recruitment to these seminars is also related to the payment of the sub-contractors. There is, as cited above, evidence that practical and active participation in real live sessions are more effective ways of learning than simply reading of a screen. So the seminar of about 18-22 hours is the major part of the intervention. The seminars will disseminate the key learning goals for the seminar.

Learning goals: The purpose of this project is to improve the management of suicide risk, anxiety and depression by GPs. Consequently, the GPs are expected to learn some practical skills, useful techniques, and also change attitudes and reduce stigma, and ultimately change and improve clinical practice.

The main learning goal is to improve management of suicide risk, and also improve recognition and treatment of common mental disorders. The learning goals will be restricted to suicide risk and depression, general anxiety disorder and panic disorder diagnosis. GPs will also learn what not to do in order to decrease potentially harmful practice. The achievement of these goals is based on a combination of activities, including reading material, watching video and attending lectures.

Documentation of effects of the intervention: The effect of the intervention is to be measured by a controlled trail with intervention and control groups. The main outcome of this trail is the development of the suicide rate in the intervention regions compared to the control regions, and this information will be based on registry information of suicide rates. Data will be analyzed on aggregated level by region and time only, and not with any attempt to link suicide or patients to GPs.

Power analysis: The incidence of suicide is about 10 in 100000 in the general population per year. The outcome will be monitored during 24 months follow-up starting just after the intervention has been implemented long enough to expect an effect. The study has 75 percent power to detect a reduction in suicide rate from 20 in 100000 per two years follow-up to 17.5 in 100000, with a 1-sided test (alpha = 0.1). In other words, among the 2000 GPs attending the training, there will be about 800 suicides during the 24 months follow-up period. The trail has 75% power to detect a significant (p<.05, 1-sided) effect if the number of suicides among GPs being trained is reduced from 800 to 700 during the observation period. We will also analyse suicide attempts, which is about four times more common than suicide.

Inclusion Criteria:

• GP, psychologist, social worker in Bulgaria

Exclusion Criteria:

-