The World Health Organization defines obesity as any abnormal or excessive fat distribution that presents a health risk. There are several ways to categorize someone as obese; one of the most common method is using a Body Mass Index (BMI). According to WHO, someone is overweight and obese when the BMI is equal to or more than 25. In 2016, more than 1.9 billion people 18 years and older were overweight, and over 650 million of them were obese; out of the world’s population, there were 13% obese adults aged 18 years and over. In Indonesia, in 2018, the average number of adults ages 18 years and over who are obese is 21.8%, with the highest obesity rate being recorded in North Sulawesi, at 30.2%, and the lowest obesity rate being recorded in East Nusa Tenggara, at 10.3%. This result tends to increase compared to previous years. Many programs have been done worldwide to overcome the problem of obesity, such as promoting physical activity programs to bring more active people to a healthier world, regular health check-ups, and motivating individuals to go on a healthy diet by reducing sugar consumption and eating more vegetables and fruits.

In this study, we present a mathematical model that describes how obesity spreads among the human population, considering human awareness levels to describe the difference in lifestyle of humans, in which the transition between this group depends on the media campaign from the authority about the importance of healthy lifestyles and persuasive capability of individuals who quit obesity. The model is constructed as a set of four-dimensional nonlinear ordinary differential equations. The transmission diagram of the model is shown in Fig 1.

Figure 1. Transmission diagram for an obesity model with human awareness level.

Possible equilibrium points are investigated regarding their existence and local stability criteria. We obtained a basic reproduction number (R0) of the model from the next-generation matrix approach. The results of the analysis show that the obesity-free equilibrium is locally asymptotically stable if R0 is less than one and unstable otherwise. A transcritical bifurcation when R0=1 was investigated using the Castillo-Song bifurcation theorem. We conducted an elasticity analysis of R0 numerically to show the robustness of R0 to the value of the change of parameters in the model, and we found that the social contact rate is the most influential parameter in determining the magnitude of R0, followed by a healthy lifestyle campaign from the government. Finally, we conducted a short discussion to understand the possible scenarios in the field obtained numerically based on our analytical results.