Modeling Diet Decisions for People Living with HIV in Consideration of Age, Weight, Height, and Gender Constraints

Abstract

People living with HIV as well as AIDS patients, who do not receive proper and timely medical treatment, are open targets for all kinds of other infections owing mainly to their relatively weak immune systems. We emphasizes upon the fact that, in most (if not all) such cases, poor nutrition intensifies the progression of the disease and that achieving basic nutritional recommendations is important at all stages of the disease. This paper aims to develop a cost-effectiveness computing model (mathematical model) in diet decisions for people living with HIV in consideration of age, weight, height, and gender constraints. The consideration of these factors tends to avoid undertaking/overtaking of the nutrients which may lead to more serious problems. This model combines multiple linear regression model and linear programming model. The multiple linear regression model predicts the nutrient requirements in the human body of the factors age, weight, height, and gender. The multiple linear regression model gives out the maximum allowable amount of nutrients (upper bound) and minimum amount of nutrients required (lower bound). These results are used to restrict some constraints in the linear programming model, while others are restricted to the maximum allowable amount of foods. From the linear programming model adequate amount of foods that achieve the nutrients recommended are computed. The linear programming problem formulated is solved by the two phase simplex method in MATLAB. Results show that multiple linear regression predicted values are close enough to the actual recommended dietary/daily intake values. The optimal nutrients are reached at much less cost when the multiple linear regression predicted values are used as nutrient recommendations to restrict the constraints in linear programming model compared to when actual recommended dietary/daily intake values are used. Since our model gives adequate amount of foods at much less cost than when the actual values are used then this justifies that our goal has been successful reached. The mathematical model developed could potentially be extended to different groups of people who must manage their diets and therefore promises to have a wider applicability.