• Depressive symptoms measured by the Patient Health Questionnaire-9 (PHQ-9) [ Time Frame: 12 months ]
  The Patient Health Questionnaire-9 (PHQ-9) measures symptoms of depression through 9 items, each of which is scored 0 ('not at all') to 3 ('nearly every day'). Low scores are equivalent to less symptoms of depression, the scale range is 0 to 27 (9 items)
• Anxious symptoms measured by the General Anxiety Questionnaire (GAD-7) [ Time Frame: 12 months ]
  The General Anxiety Questionnaire (GAD-7) measures generalized anxiety disorder through 7 items, each of which is scored 0 ('not at all') to 3 ('nearly every day'). Low scores are equivalent to less symptoms of anxiety, the scale range is 0 to 21 (7 items)
• Probability of depression (predictD risk algorithm) [ Time Frame: 12 months ]
• Cost-effectiveness and cost-utility [ Time Frame: 12 months ]

• Depressive symptoms (PHQ-9) [ Time Frame: 12 months ]
• Anxious symptoms (GAD-7) [ Time Frame: 12 months ]
• Probability of depression (predictD risk algorithm) [ Time Frame: 12 months ]
• Cost-effectiveness and cost-utility [ Time Frame: 12 months ]

The main goal is to design, develop and evaluate a personalized intervention to prevent the onset of depression based on Information and Communications Technology (ICTs), risk predictive algorithms and decision support systems (DSS) for patients and general practitioners (GPs). The specific goals are 1) to design and develop a DSS, called e-predictD-DSS, to elaborate personalized plans to prevent depression; 2) to design and develop an ICT solution that integrates the DSS on the web, a mobile application (App), the risk predictive algorithm, different intervention modules and a monitoring-feedback system; 3) to evaluate the usability and adherence of primary care patients and their GPs with the e-predictD intervention; 4) to evaluate the effectiveness of the e-predictD intervention to reduce the incidence of major depression, depression and anxiety symptoms and the probability of major depression next year; 5) to evaluate the cost-effectiveness and cost-utility of the e-predictD intervention to prevent depression.

Methods: This is a randomized controlled trial with allocation by cluster (GPs), simple blind, two parallel arms (e-predictD vs "active m-Health control") and 1 year follow-up including 720 patients (360 in each arm) and 72 GPs (36 in each arm). Patients will be free of major depression at baseline and aged between 18 and 55 years old. Primary outcome will be the incidence of major depression at 12 months measured by CIDI. As secondary outcomes: depressive and anxiety symptomatology measured by PHQ-9 and GAD-7 and the risk probability of depression measured by predictD algorithm, as well as cost-effectiveness and cost-utility. The e-predictD intervention is multi-component and it is based on a DSS that helps the patients to elaborate their own personalized depression prevention plans, which the patient approves, and implements, and the system monitors offering feedback to the patient and to the GPs. It is an e-Health intervention because it is based on a web and m-Health because it is also implemented on the patient's smartphones through an App. In addition, it integrates a risk algorithm of depression, which is already validated (the predictD algorithm). It also includes an initial GP-patient interview and a specific training for the GP. Finally, a map of potentially useful local community resources to prevent depression will be integrated into the DSS.

• Behavioral: e-predictD intervention
  The intervention is based on validated risk algorithms to predict depression and includes: 1) Mobile applications as main user's interface; 2) a DSS that helps patients to develop their own personalized plans to prevent (PPP) depression; 3) eight intervention modules (the core of the system) including activities to prevent depression, to be proposed by the DSS and chosen by the patient. The intervention is biopsychosocial and multi-component, including the following modules: physical exercise, improving sleep, expanding relationships, problem solving, improving communication skills, assertiveness training, making decisions and managing thoughts. Patients will implement the recommendations and the tool will monitor these actions, offering feedback to improve their PPP at 3, 6 and 9 months. The intervention also includes an initial and single 15-minute face-to-face GP-patient interview.
• Other: Brief psychoeducational intervention
  The intervention consists of an App that weekly send brief psychoeducational messages about physical and mental health (depression, anxiety, sleep hygiene, physical activity, etc.)

• Experimental: e-predictD intervention
  In this arm, patients will receive a personalized intervention to prevent depression based on ICTs, risk predictive algorithms and decision support systems (DSS) for patients and General Practitioners (GPs).
  Intervention: Behavioral: e-predictD intervention
• Active Comparator: m-Health control
  In this arm, patients will continue receiving the usual care from their GPs. In addition, they will use an App with the same appearance as the e-predictD App but it will only send weekly messages about physical and mental health management. This intervention is not personalized and does not include GP training and GP-patient interview.
  Intervention: Other: Brief psychoeducational intervention

Inclusion Criteria:

• PHQ-9 <10 at baseline
• Moderate-high risk of depression (predictD risk algorithm score ≥ 10%)

Exclusion Criteria:

• Not have a smartphone and internet for personal use
• Unable to speak Spanish
• Documented terminal illness
• Documented cognitive impairment
• Limiting sensory disorder (e.g. deafness)
• Documented serious mental illness (psychosis, bipolar, addictions, etc.)

• Preventive Services and Health Promotion Research Network
• Institute of Biomedical Research in Málaga (IBIMA)
• Andalusian Regional Ministry of Health
• European Regional Development Fund (FEDER)
• University of Malaga