Caused by habits such as poor diets, lack of physical activity practice or smoking, non-communicable diseases were elected by the World Health Organization as one of the greatest challenges of the twenty-first century, despite a lot of information produced in social media focused on preventing this type of disease. This paper presents the Pompilos Model, which aims at improving computer-aided social support by suggesting beneficial health resources and revealing what inuences other people's health, so to foster better health behaviors in social relations. In order to evaluate the model's feasibility, we performed a random experiment during one month and half with two groups to assess the influence of messages related to the prevention of chronic diseases. Those messages presented information on a healthier diet, the practice of physical activities, and ways to lose weight, from monitored Twitter profiles on the habits of health assistant web application's users. So it would be possible to manage food intake, the practice of physical activities, and weight control. Messages related to the prevention of chronic diseases, such as a healthier diet, the practice of physical activities, and weight loss from monitored Twitter profiles were directed to an intervention group as a way to re-engage users in their care activities. With this information, we found a correlation between message reading and the access to the application history feature among intervention users.