SYSTEMS APPROACH TO DISEASE MODELING

Purpose: Current disease models often focus on a narrow class of questions, related to a specific disease stage. Our objective is to develop a systems approach to health economic modeling, capturing all stages of disease evolution. Such an approach will enable (i) evaluation of a wide range of therapeutic options, (ii) optimization of resources allocated to reducing disease burden, and (iii) comprehensive assessment of the downstream effects of clinical decisions.

Methods: We demonstrate the feasibility of the proposed approach by developing a comprehensive representation of colorectal cancer (CRC), which is fully integrated with other diseases captured by the Archimedes Model. The CRC model consists of  (i) a natural history component that tracks development of pre-cancerous lesions, their transformation into cancerous tumors, growth and spread of tumors, cancer symptoms, recurrence, metastasis, and deaths, in the general population as well as in specific subpopulations, such as those with inflammatory bowel diseases or carriers of mismatch repair (MMR) gene mutations; (ii) a healthcare process component that describes the interactions between cancer patients and the healthcare system and the impact of medical interventions, such as behavioral modification, prevention activities, cancer screening, and treatment on CRC outcomes; and (iii) a family component that tracks the development of CRC and related diseases in family members.

Results: We used the proposed simulation framework to address a wide variety of questions related to colorectal cancer, including

• What is the cost effectiveness of hybrid screening strategies (combining stool tests, sigmoidoscopy, and colonoscopy) for CRC as compared with single-modality strategies [1]?
• What is the impact of primary genetic screening for MMR mutations in the general population when combining family history-based risk assessment with genetic testing [2]?
• What are the optimal ages to screen for CRC in elderly patients with chronic illnesses (e.g. those diagnosed with hypertension and/or diabetes) [3]?
• What is the cost effectiveness of adjuvant therapy for stage II colon cancer in patients with pre-existing diabetes [4]?

These modeling studies highlight the fact that interesting and unexpected behaviors can emerge from complex relationships between different stages of the same disease, as well as between different diseases.

Conclusion: A systematic and holistic approach to disease modeling is crucial for accurate health economic assessment of modern medicine.