After having spent > 10 years working on clinical medical device software, I'm reaching the conclusion that open source is likely to be a better approach to design control of medical device software than what opaque companies are currently able to achieve.
Edit Switching Costs and Medical Device Risk Management
One of the main strategies that recurs when technology-based vie for market share is to attempt to
control the platform; that is, creating and controlling a common environment for building additional applications or solutions. Within the proprietary realm, this leads directly to the idea of increasing
switching costs in order to prevent users from being able to switch to a different platform.
From a market efficiency point of view, the problem with increasing switching costs is that it can often lead to tactics that involve arbitrarily wasting the time of other firms or groups operating in the same field. If the switching cost is proportional to how much others may know about a particular platform, then as others learn more about the platform, the switching cost will go down. In order to keep switching costs high, then, the platform proprietor can benefit by coming up with clever means of wasting the time of others; for instance by making vaporware announcements, spreading misinformation, planting red herrings, etc.
In a medical device domain, the question is the effect of these tactics on design control of device(s). Design control is essential for reducing risks associated with operation of the device(s). If the tactics for keeping switching costs high cause such risk management activities to become less effective, then the increased switching costs can directly act against efficacious risk management of the device(s). This would be the case, for instance, when the risks are associated with the interoperability of devices from multiple vendors, and vendor(s) are attempting to increase switching costs in order to keep clinical users locked in to their particular devices or 'ecosystems'.
Edit Why Open Source Medical Devices?
The quality of some of the software that I have worked on (software that is used in MANY clinics throughout the world) is simply horrendous, including lack of effective design documents and technical decisions being made by individuals who are cleary not qualified to be making such decisions. With open source projects, I don't think there is an expectation that the people working on the code are necessarily more qualified; however, if a bad design decision is made, the code base can always be 'branched' in order to circumvent the potential downstream problems.
In addition, just having a larger group of eyeballs, from diverse environments, looking at the code should gradually improve the quality over time. More eyeballs also leads to a greater understanding of the design as it relates to actual clinical use, _not_ filtered through the obfuscated understanding of marketing folks.
Of course, open source would also make it more obvious if certain groups or individuals were attempting to force the design in a particular direction due to hidden motivations (see the section above on 'switching costs').
Edit ITK / VTK
These are libraries for graphics and image processing, as opposed to applications (though there are a number of applications).
Edit Osirix
A Macintosh-based image visualization platform. Haven't used (because I don't have a Mac), but I hear good things.
Edit CERR
A research platform, operating within MATLAB, for various radiotherapy algorithms.
Edit Open Source EMR
This is an interesting field, as the possibilities for applying open source to EMR are fairly large.
Edit References
Information Rules (by Carl Shapiro and Hal Varian). Good introduction to how companies can use increased switching costs to monopolize markets.