Minimizing Health-Compromising Behaviors via School-Based Programs: An Optimization
Mathematical modeling helps optimize school-based programs to minimize health-compromising behavior among youth.
School health programs are united by their desire to promote health and health-related outcomes among youth. They are also united by the fact that their expected effects are contingent on successful program implementation, which is often impeded by a multitude of real-world barriers. This study develops a mathematical model that calculates the number of lessons that should be taught to minimize substance use in a school environment, given expected levels of predetermined implementation barriers. The findings from this exploratory study support the utility of applying mathematical modeling during the program planning and implementation processes of school health programs.
Dr. Banafsheh Behzad from the College of Business and Dr. Niloofar Bavarian from the college of Health and Human Services along with Sheena Cruz, a graduate student, developed a mathematical model to study health-compromising behaviors via school-based programs. This multidisciplinary study shows the importance of using quantitative approaches in studying health-compromising behaviors such as substance use among youth.
School-based health-promotion programs come in various forms and have a variety of aims. What unites these programs is that their attainment of desired outcomes is often directly associated with implementation levels. One indicator of implementation, which broadly refers to program delivery, is dosage, which refers to the quantity of program delivery. The purpose of this exploratory study was to apply mathematical modeling to one specific, school-based program, Positive Action (PA). PA is a social-emotional and character development (SECD) program in which six lessons are delivered by the classroom teacher. From 2004 to 2010, PA was implemented via a randomized controlled trial in seven Chicago Public Schools in low-income settings. The outcome of interest in this study was substance use, a health-compromising behavior. The goal was to determine if it might be possible to obtain estimates of the specific amounts of implementation needed (i.e., number of lessons for each of the six units needed to be taught) to minimize substance use, given pre-specified and expected levels of various implementation barriers. Implementation is influenced by factors that can occur at various levels (e.g., teacher, school, student). Linear Programming is an optimization technique which uses a mathematical model of linear equations with the objective of planning the best possible allocation of scarce resources, under a set of constraints that serve as barriers to implementation. This study uses linear programming to calculate the optimal levels of program implementation needed to minimize substance use, subject to known levels of implementation barriers (e.g., disruptive behavior, teacher education, teacher attitudes towards character development, school resources, and school safety). The results of this exploratory study demonstrate the feasibility of applying linear programming to school health programs.
Closing Paragraph: This research calculates the optimal numbers of lessons per unit to be taught weekly in a school environment in order to minimize substance use, subject to set values of known implementation barriers. On average, teachers in the PA trial came close to teaching the optimal numbers of lessons from each unit. This may explain, in part, why previous studies from this trial showed an impact on substance use, as compared to control schools. Teachers who taught less than the optimal amount should be consulted to better understand implementation barriers, and how they can be overcome in future studies.
