The models used by investigators in the Cetner for Health Decision Science combine information about the natural history of a disease and intervention efficacy with other relevant demographic and epidemiological characteristics of a population. The disease-based mathematical models that are developed by many of the Center for Health Decision Science investigators are often "biologically-based," taking into account both epidemiological observations as well as the mechanics of the disease process as it unfolds. Others are mainly empirically based, drawing on evidence obtained from clinical trials or observational databases. Most models, however, combine biological and empirical evidence. Examples of models developed by Center for Health Decision Science investigators include:

• Tuberculosis Models. A model of drug-susceptible and drug-resistant tuberculosis strains allows investigators to compare different strategies for applying standardized treatment regimens under the WHO-recommended DOTS approach, or alternative approaches based on individualized or standardized regimens for multi-drug resistant cases, with or without drug sensitivity tests.

• HIV and AIDS Models. Center researchers are part of a multidisciplinary group of investigators that has addressed critical HIV/AIDS-related clinical and policy questions using mathematical modeling techniques. The CEPAC (Cost-effectiveness of Preventing AIDS Complications) team has developed a computer-based, state-transition model of HIV disease progression and treatment. The model is a first-order Monte Carlo simulation in which disease progression in an individual patient is characterized as a sequence of monthly transitions between health states. In the U.S., results have informed national HIV clinical guidelines and policy decisions, such as clinical guidelines for the use of genotypic resistance tests to guide therapy, coverage rules set by AIDS Drug Assistance Programs, and guidelines for HIV testing. Research with French investigators has shown differences in the spectrum and risk of specific opportunistic infections, highlighting the need for country-specific guidelines and treatment policies.
  

• Hepatitis B and C Models. A dynamic population model of hepatitis B and hepatitis C allows analysts to forecast health and economic consequences in the U.S. The models of hepatitis developed by center investigators have been used to address a wide range of policy questions in different parts of the world. For example, a cohort simulation model was used to evaluate the cost-effectiveness of HBV vaccination strategies targeting high-risk persons in the U.S., while a dynamic population model has been used to project the overall impact of various HBV vaccination policies, and to quantify positive externalities due to the current policies in the U.S. A different cohort model was used to assess the impact of HBV infant vaccination in the Gambia and extended methods to add consideration of affordability to cost-effectiveness analyses. Analysts have assessed the cost-effectiveness of treatment for HCV in the U.S., and have developed a framework to formally consider the value of new technological innovation in future prevention and treatment options.
  

• HPV and Cervical Cancer Models. A series of models of HPV and cervical carcinogenesis have been developed by center investigators to address questions of screening policy, and to inform implementation of the recently-approved HPV vaccine. For example, these models have been used to assess new technology including screening diagnostics, evaluate national and regional screening policies, consider the adaptation of screening programs in the context of vaccination, inform clinical trial design, and develop clinical guidelines for special populations. Models vary from simple state transition models evaluated as cohort simulations to stochastic models capable of capturing complex heterogeneities at the individual and population level. Also under development are dynamic transmission models capable of assessing the indirect effects of vaccination programs, such as herd immunity. Novel calibration techniques allow analysts to parameterize complex models, adapt biological natural history models to different epidemiological settings, utilize region-specific epidemiologic data to reflect important variations and patterns of disease. CHDS investigators are conducting analyses tailored to more than 18 countries including South Africa, Kenya, Mozambique, Haiti, Peru, Thailand, India, China, Brazil, Vietnam, and Costa Rica, as well as the U.S., several European countries, and Hong Kong.
  

• Lung Cancer Policy Model. CHDS investigators have developed a biologically-based simulation model of lung cancer that is being used to evaluate the cost-effectiveness of screening of high-risk individuals, such as long-time smokers, with computed tomography (CT). Screening with conventional chest x-rays has not been recommended in clinical practice guidelines, but CT offers a more sensitive test for tumors at an earlier stage of growth, possibly permitting earlier and more effective intervention. These analyses of alternative screening strategies will be extended to evaluate new targeted treatments for lung cancer, including minimally invasive ablation therapies and pharmaceuticals.
  

• Cardiovascular Disease Models. Cardiovascular disease (CVD) has long been the major cause of death in the U.S., and it has now become a major public health problem in developing countries such as India. Aggressive strategies to control CVD through the use of basic drugs that have long been known to be effective in preventing CVD, such as aspirin, diuretics, beta-blockers, and statins, are now being considered even in poor countries. PHDS investigators have adapted a computer model previously used in the U.S. to evaluate CVD prevention strategies in developing countries. Meanwhile, models to evaluate contemporary technologies – drugs, devices, and diagnostic technologies -- to diagnose and treat CVD in the U.S. are also being developed and used by CHDS investigators.
  

• Maternal Mortality Models. A series of models have been developed to evaluate the health and economic impact of implementing a broad array of health interventions and strategies to reduce maternal mortality and morbidity in low-resource settings. The model simulates the clinical course of pregnancy and the risks and consequences of major obstetric complications (e.g., eclampsia, obstructed labor, postpartum hemorrhage, sepsis, sexually transmitted infections) and unsafe abortion, and can be used to assess the impact of individual interventions, packages of services, and complex strategies. Analyses have addressed the progress achieved in reducing maternal mortality over the last 20 years in Mexico, and recommending effective and cost-effective interventions specifically tailored to circumstances in Mexico for further improvement in maternal health and achievement of Millennium Development Goal 5. Investigators are currently adapting the model to other settings with high maternal mortality, including India, Nigeria, and Haiti.
  

• Women’s Health Package of Services Models. An analytic framework has been developed for packaging multiple interventions during a single point of contact, explicitly taking into account a budget and scarce human resources, constraints acknowledged as significant obstacles for provision of health services in poor countries. CHDS investigators developed an integer programming model to maximize health gains associated with interventions for multiple diseases and provide a simplified example of packaging health services for women during a single lifetime cervical cancer screening visit. These methods can enhance a decision maker’s ability to simultaneously consider costs, benefits, and important nonmonetary constraints. The model is currently being developed to address different strategies for safe motherhood taking into account monetary, human resource (e.g., skilled birth attendants), and infrastructure/facility constraints.
  

• Breast Cancer Models. As part of the NCI-sponsored Cancer Intervention and Surveillance Modeling Network (CISNET) consortium, CHDS researchers developed a discrete event microsimulation model capable of replicating and predicting the epidemiology of breast cancer in the U.S. female population from 1975 to the present. The model accounts for time trends in the incidence of breast cancer and the dissemination of both screening mammography and adjuvant treatment. The end result was a tool that could be used to address policy level questions about the prevention and control of breast cancer. In addition, the model has the ability to conduct individual level counterfactual experiments, i.e., “what if” analyses, that could not be performed in the real world. CHDS investigators are also using the model to provide information about the benefits and risks of screening mammography usage to aid women, clinicians, and policy makers and others who are making decisions about using mammography.