Models
Modelling is a term referring to analytic techniques applied for simulation of events when direct data are not available. Its objective is to present the investigated problem in a simplified way.
The most widely applied models include:
- decision trees,
- Markov models,
- DES models.
Calculations are performed using the cohort method or Monte Carlo simulations. In addition, the Bayesian methods may be applied in modelling.
In creation of our models we abide by two principles: simplicity - to make the model understandable - and complexity - to include the most important elements of the health problem in the model.
Each model is a unique project created by a team of experienced analysts. The first step is to collect data concerning current approaches to the investigated problem, i.e. search medical databases in order to identify models already created. On the basis of the decision problem analysis and specialist consultations we determine possible pathways of management of a patient with a specific health problem, possible events generated by the disease entity and the related states, in which patients may find themselves.
Our models are developed in the form of interactive, user-friendly software. In most cases various parameter values may be entered and thus numerous possible scenarios may be tested.
Example
A specific disease entity significantly affects the patients? survival. A study for a new agent performed in a 6-month time horizon suggests reduced mortality and decreased incidence of adverse events as compared to currently used medications. However, the effect on patients? survival beyond this time horizon remains unknown. Effects of further use of the new agent may be estimated by means of modelling.