Leveraging DICE simulation for modelling obesity
Presented by the Economics of Obesity Special Interest Group (EOSIG)
The webinar is taking place at 08:00 AM ET. Please check your timezone online here.
To present a novel approach to building a simulation that provides a realistic transparent framework for modeling obesity with no need for programming or new software.
Many interventions to reduce weight are in development and their benefits should extend beyond weight loss, including lipid and glycemia profile improvements and their downstream consequences. Thus, it is necessary to translate the physiologic gains to tangible clinical outcomes and to examine the potential economic implications in various target populations. A key component of this process is a flexible model that incorporates time-dependence and addresses risk factors as they cluster and change in individual patients. This is challenging because body weight fluctuates, has complex relationships to adverse consequences, and the results of treatment are individual and variable.
DICE (Discretely Integrated Condition-Event) simulation, a modeling technique that conceptualizes a disease process and its management in terms of conditions (bits of information) and events (instants where information changes), was chosen because it supports the development of realistic models in EXCEL® that avoid the over-simplification of Markov models. Management of obesity is simulated in detail, while keeping the model logic transparent. Each simulated person is given a profile comprising demographics, smoking status, clinical characteristics (e.g., full lipid profile, systolic blood pressure, fasting plasma glucose, HbA1c), relevant medical history, and medications. During the simulation, each patient is exposed to risks of cardiovascular events, of developing diabetes, of diabetes-related complications and of death due to related or unrelated causes. Competing risks are implemented by deriving the distributions of event times and sampling a time for each event for each patient. Event times occurring before death lead to processing that event and its consequences. When an event occurs, the management of the patient may change and some risks may be recalculated. Changes in physiologic parameters over time also require recalculating risks. Time spent in the model (survival), quality-adjusted life years (QALY), years spent with diabetes, various clinical events, relevant hospitalizations, use of various treatments and regular monitoring are counted and their associated direct medical costs are estimated. The proportion of patients exceeding user-defined risk thresholds are reported, along with the distribution of ten-year risk of coronary heart disease and cardiovascular disease. The model concept accords closely with reality and handles time dependencies accurately; is flexible; easy to modify with new data or assumptions; and able to examine the influence of assumptions (i.e., “structural” sensitivity analysis).
J. Jaime Caro MDCM FRCPC FACP
Prof Medicine, Epidemiology & Biostatistics, McGill University
Pro in Practice, Health Policy, London School of Economics
Chief Scientist, Evidera
J. Jaime Caro, MDCM, FRCPC, FACP, is Chief Scientist at Evidera where he advances Evidera’s leadership in developing and applying novel techniques in modeling, health economics, comparative effectiveness, epidemiology, and outcomes research. Dr. Caro is also adjunct Professor of Medicine, Epidemiology and Biostatistics at McGill University, and Professor in Practice at the London School of Economics. He also lends his teaching ability to other academic institutions such as Thomas Jefferson University School of Population Health.
Dr. Caro continues to pioneer new methodologies. In an effort to provide an alternative to the well-known cost per QALY technique and avoid many of the latter’s problems, he is working on a broader approach to valuing health benefits. He is also further developing DICE, the unified approach to health economic modeling that he has created. Working with health technology assessment agencies and academic groups, he is formalizing this innovation to enable rapid, standardized and less error-prone development of decision-analytic model. Previously, Dr. Caro adapted an engineering technique – discrete event simulation – to model diseases and their treatment and extended it further to simulate the design of clinical trials and other types of studies, something which has been particularly effective in helping design pragmatic clinical trials. He has also applied the technique to provide comparative effectiveness information in the absence of head-to-head trials in a new method called Simulated Treatment Comparison. In response to increasingly frequent requests by authorities to provide data to support the value of new interventions, Dr. Caro has applied simulation techniques to make post-marketing registries more feasible and efficient – an approach called SAVES.
On behalf of the German health technology assessment agency, he proposed an innovative approach to the assessment of health technologies, involving the efficiency frontier. As part of his work with governments, he has helped the World Bank Institute and the InterAmerican Court for Human Rights address the growing problem of Supreme Courts overriding health care system decisions and ordering them to provide treatments that had been considered unwarranted. After leading the Quality Assurance for Modeling Studies Task Force, jointly sponsored by ISPOR, Academy of Managed Care Pharmacy and the National Pharmaceutical Council, and the ISPOR-SMDM Good Modeling Practices Task Force, endorsed by the Society for Medical Decision Making, Dr. Caro was recently named co-chair of the ISPOR Science and Research Committee and has been awarded the Marilyn Dix Smith Leadership Award.