Page History

Selection of initial model input parameters is an important step during model development. Input parameters are typically chosen to represent characteristics of the entire population: transitional probabilities (initiation, cessation, quitting, switching) and health outcome (mortality and/or morbidity rates). The initial values of these parameters are typically estimated from sources representative of the population, such as complex probability-based surveys or census data.

Transitional Probabilities 

Transition probabilities are typically chosen to represent characteristics of the population of interest (e.g. initiation, cessation, quitting, switching). Transition probabilities determine the rate at which the population changes with respect to time during the period of interest. For example, one might be interested in projecting the health impact of introducing a new tobacco product to the US population for a period of 50 years starting with the year 2000; transition probabilities govern transition between tobacco use behavior as well as health outcome during the projection period (50 years, in this example). The following figure shows different transition for a 2-products model developed by the FDA; each letter represents a transitional probability representing transitions from tobacco use behaviors over time.  

The figure below illustrates transitional probabilities for a two-product model. Transition probabilities are estimated from population surveys and other sources, and they are typically tabulated by sex, age and race. In some models, these are constant during the entire simulation period, but they could also change over time. 

Mortality Rates and Risks

Another type of input parameters are those related to risk; this included mortality rates and risks related to comorbidities. For many models, these are projected over the entire simulation period (e.g., 50 years in the future), and they are estimated from population surveys or other data sources. These parameters are tabulated based on age, sex, race, tobacco use status, and years of tobacco use and/or former use.

Birth Rate

At each step of the simulation period (e.g., each year), individuals may enter the population according to birth rates that are projected across the time horizon of the model. These rates are typically obtained from government census data.

Migration

Similar to birth rates, at each step of the simulation period individuals may enter the population according to migration rates. Base migration rates are generally obtained from national census data and if necessary are projected across the simulation timeframe. Depending on specific modeling assumptions, immigration demographics define the sex, age, race/ethnicity, and prevalence of tobacco product use behavior (never, current, former dual use, polytobacco use) in the immigrating population at a particular period.