Why Open Source?
From Prosit
This page is approved PROSIT Content and should not be edited between releases.
Contents |
[edit] Why Open Source? An attempt to give an explanation
In this editorial, I would like to give a reasoning for why we use Open Source for developing disease models. Please feel free to comment!
Best regards
Wendelin Schramm
[edit] Why Open Source?
Disease Modelling is a standard technique for decision analysis and health economics. The mathematical approaches are sound and guidelines exist for developing and publishing good models [1].
End of message.
[edit] But the impact of disease modelling is still low…
This is a fact.
Why is this?
Through many years of practical experience, I conclude that there are two main reasons:
1. Disease modelling is complex. This means the users or customers of disease models must believe in the credibility of a piece of software developed by third parties, with a mathematical background far away from common knowledge and understanding. It is a classical "black-box" problem.
2. Disease Modelling is subject to interests. Today most of the leading disease models are developed for or by the pharmaceutical industry. Which has lead to important quality problems as these models are used e.g. for pricing and reimbursing of pharmaceutical products [2].
The result: People do not believe in models.
The PROSIT community wants to overcome these two main obstacles.
Ad 1.) We cannot reduce the complexity of modelling as such and we cannot eradicate the black-box issue. Modelling simply remains complex. But, we can make sure that experts from all over develop models together. Everybody has an insight and can participate and nothing can be held back to the experts nor to public. If one believes that a mistake has been made, he or she can simply post his or her remarks and the community will go for it. Controversies and problems are laid open. Thus any user can check the situation for a specific model.
Another aspect: we are using spreadsheets for modelling. Admitted, thus we cannot implement sophisticated simulation techniques such as discrete event simulation. But our approach adds to the credibility as in our spreadsheet cells we cannot hide anything. A hidden calibration factor? A tuned parameter distribution? No way, there are no tricks or cheats, because everything would come out. Even if we used white letters in front of a white background.
Ad 2.) We all here have interests when we sacrifice time and when we share knowledge in this Open Source community. But our interests are related to making good models and not in achieving desired results by means of modelling. It is a simple but important change in paradigms.
Of course, there are also major cons like the slow speed of development, the constant lack of resources and many other issues. But nothing can stand against the two big pros:
Transparency and Credibility achieved by using the methods and the software of the Open Source movement.
[edit] Literature References
[1] Drummond MF, Jefferson TO. Guidelines for Authors and Peer Reviewers of Economic Submissions to the BMJ: The BMJ Economic Evaluation Working Party. BMJ1996; 313: 275-283
[2] Sanchez LA. Evaluating the Quality of Published Pharmacoeconomic Evaluations. Hosp Pharm 1995; 30: 146-142