Advanced Decision Modeling
[ EL006 ]

This week-long, project-based course aims to provide students with an understanding of advanced methods used in decision-analytic modeling and cost-effectiveness analyses. These include topics like the latest methods for calibration and validation, quantifying uncertainty, and consideration of heterogeneity of patient benefits and equity issues. The course combines lectures and readings to give theoretical foundation and perspectives with in depth project work and presentations to give practical concrete understanding in a way that furthers students’ specific research goals.

Course Structure (note: the online version of the course may differ somewhat):
Each day will begin with a lecture by Professor Goldhaber-Fiebert on an advanced methods topic. After the lecture, lab sessions will commence with students working on their projects as Professor Goldhaber-Fiebert circulates through the room and students assist each other in a collaborative environment. Most days Professor Goldhaber-Fiebert will also give an afternoon lecture. In addition, at the end of days 2, 3, and 4, Professor Goldhaber-Fiebert will give an additional, shorter, informal lecture (i.e., "a chalk talk") on a methods topic tailored to specific issues that are arising within students’ projects. Additionally, throughout the week, Professor Goldhaber-Fiebert will have one-on-one meetings with students about their projects.

• Understand how empirical data can be used to indirectly estimate parameters and their uncertainty for decision analytic simulation models (calibration);

• Be able to differentiate model calibration and model validation and to determine which situations it is appropriate to use each;

• Understand the importance of correlations in the uncertainty distributions of model parameters for probabilistic sensitivity analyses (PSAs) and

• Be able to identify methods to capture correlations in uncertainty for use in PSAs;

• Understand how patient heterogeneity and behavioral responses can alter the results of cost-effectiveness analyses;

• Be able to design models that capture patient heterogeneity and behavior;

• Be able to describe Markov Decision Processes and other operations research techniques and to identify features of decision problems that make require these techniques;

• Be able to describe key features and properties of compartmental dynamic transmission models for infectious diseases (i.e., SIR models) and how results of cost-effectiveness analyses of infectious disease control may depend on the types of models used;

• Understand the use of Value of Information analyses in relation to standard cost-effectiveness analyses;

• Be able to conduct an Expected Value of Perfect Information analysis;

• Be able to design and implement a model based cost-effectiveness analysis that appropriately incorporates advanced decision modeling techniques.

• Advanced Topics in Medical Decision Making (EL004) or an equivalent advanced course in decision modeling (e.g. Harvard T.H. Chan School of Public Health RDS288 or RDS285)

• Using R for Decision Modeling, Simulation, and Health Technology Assessment (EL005) or an equivalent particularly for those students planning to implement their project in R.

Assignment(s), Attendance

T.B.A.