Boston - Advanced Markov/DES (2016)

2016/07/20 - 2016/07/20 | Boston, MA Sold Out

This course will be offered in Boston, MA on July 20, 2016.

The course will run from 8:30 AM to about 5:00 PM.

Course Flow

This one-day course extends training into advanced modeling techniques and tools.

We start by building a complex Markov simulation model with mutli-dimensional calculations driving the probability of events. Those events then lead to long-term changes in costs and utilities, mimicking the impact of events on real patients.

The next section focuses on Discrete Event Simulation (DES) models, where Markov transition probabilities are replaced by time-to-event distribution sampling. We start with simple examples with fixed risk, then extend it with more complex time-to-event calculations.

In the next two sections, we use built-in reporting and debugging tools that help you understand, validate and debug your models.

We then introduce more complicated models. First, we work with dynamic cohort models that focus on population dynamics and budget impact. Then we move to parallel trial models where there is interaction among the individual patients within the cohort.

The course finishes with bootstrapping to use real patient data to assign personal characteristics to our model’s simulated patients. We then segregate the output using those characteristics to look for subgroups that might benefit from a different treatment strategy.

Course Agenda

  1. Build/Analyze Markov Simulation Model
    1. Add advanced techniques to a Markov model using patient data to drive model values
    2. Add nested clones and complex variable definitions in the simulation model
    3. Analyze the Markov simulation model and interpret results
  2. Build/Analyze a Discrete Event Simulation Model
    1. Compare a simple Markov model to a simple DES model
    2. Build a Discrete Event Simulation (DES) model
    3. Analyze DES model via Microsimulation
    4. Add complexity for non-fixed probabilities and time horizon
    5. Run sensitivity analysis on DES model
  3. Simulation Time Reporting
    1. Track patients in model
    2. Generate cohort-level details from patient simulation analysis
  4. Debugging models
    1. Output complex calculation trace data to the console
  5. Dynamic Cohort
    1. Discuss when Dynamic Cohort models are needed
    2. Use non-coherent probabilities to set the cohort size
    3. Add to the cohort by cycle using an Entry node
    4. Analyze the model and examine results for the full cohort size
  6. Parallel Trials
    1. Discuss when Parallel Trials models are needed
    2. Use Stop nodes and global trackers to handle resource constraints
    3. Show Stop node and explain about key word which needs to be used.
    4. Analyses the model and examine results
  7. Subgroup Analysis
    1. Heterogeneity using real patient data (bootstrapping)
    2. Draw conclusions for subgroups via simulation result filtering

Prerequisites

Note that experience with TreeAge Pro is required as a prerequisite for the course. You should be comfortable with all topics covered in the Two-Day Healthcare Modeling course.