YUMM (Yale University) Yale University School of Public Health

In the CISNET Multiple Myeloma Incubator Program, the Yale University Modeling Group has constructed and validated a microsimulation model tailored to model the natural history of multiple myeloma (YUMM), from no disease, the pre-malignant condition (i.e., MGUS) development, progression of MGUS to multiple myeloma, and death for the U.S. population aged 40 years or older. The developed YUMM can be tailored to set research priorities and design clinical trials, including assessing sample sizes and power, evaluating the impact of treatment adherence of the participants, and determining the optimal treatment strategies. It can also be used to predict trial results and set policy goals.

Contact: Shi-Yi Wang, MD, PhD shiyi.wang@yale.edu

Overview

The Yale University Natural History of Multiple Myeloma Model (YUMM) models the natural history of multiple myeloma (MM) with a lifetime horizon. It is an individual-based, stage-dependent, state-transition microsimulation model. The parameters vary based on the characteristics of each individual, accounting for risk factors of MGUS/MM development and progression, such as age, race, gender and body mass index (BMI).

The theoretical model builds on the natural history of MM (Figure 1). The YUMM model targets at the U.S. general population aged over 40 years since the prevalence of MGUS for the population aged below 40 years is very small. The model currently integrates three demographic factors, age, sex, and race and simulates the natural history of MM for four U.S. populations, including non-Hispanic white (NHW) men and women, as well as non-Hispanic Black (NHB) men and women

YUMM

Figure 1: YUMM Model Scheme.

The transitional probabilities for disease progression in the YUMM are contingent upon population demographics. Data used to estimate the parameters included the National Health and Nutrition Examination Study (NHANES), Surveillance, Epidemiology, and End Results (SEER) and related software, Veteran Health Administration (VHA), and the Centers for Disease Control and Prevention (CDC) life tables.

The assumptions of the current version of the YUMM included 1) limited demographics, including age, race/ethnicity, and sex; 2) no consideration of smoldering MM (sMM), a more advanced pre-malignant disease than MGUS; 3) not accounting for undiagnosed MGUS/MM or healthcare access; and 4) life expectancy for the entire population is capped at 100 years.

We constructed a stochastic, stage-dependent, state-transition micro-simulation model using US populations 40–80 years of age without MM, stratified by age, sex, and race. For each subpopulation, starting from year 2003 each year individuals turning age 40 entered the model. All individuals in the model were simulated through death or until age 100. We computed the proportion of individuals who transitioned to MGUS, MM, and Death over time.

We validated our simulated MM incidence, age 55-79, using the SEER data each year from year 2013 through 2017. We plotted age-specific MM incidence (Figure 2) to visually present the results in comparison to the data from SEER (2013-2017) data. The simulated results closely approximate the MM incidence in SEER.

MM incidence in the US