The previous section of this article covered the history of HLN’s open source offerings. How can it benefit from this far-thinking practice to build a sustainable business?
The obvious place to turn for funding is the Centers for Disease Control, which lies behind many of the contracts signed by public health agencies. One way or another, a public health agency has to step up and pay for development. This practice is called custom-developed code in the open source policy memorandum of the federal Office of Management and Budget (p. 14 of the PDF).
The free rider problem is acute in health care. In particular, the problems faced by a now-defunct organization, Open Health Tools, were covered in another article of mine. I examined why the potential users of the software felt little inclination to pay for its development.
The best hope for sustaining HLN as an open source vendor is the customization model: when an agency needs a new feature or a customized clinical decision support rule, it contracts with HLN to develop it. Naturally, the agency could contract with anyone it wants to upgrade open source software, but HLN would be the first place to look because they are familiar with software they built originally.
Other popular models include offering support as a paid service, and building proprietary tools on top of the basic open source version (“open core”). The temptation to skim off the cream of the product and profit by it is so compelling that one of the most vocal stalwarts of the open source process, MariaDB (based on the popular MySQL database) recently broke radically from its tradition and announced a proprietary license for its primary distinguishing extension.
Support has never scaled as a business model; it’s very labor-intensive. Furthermore, it might have made sense to offer support decades ago when each piece of software posed unique integration problems. But if you create good, modern interfaces–as Arzt claims to do–you use standards that are familiar and require little guidance.
The “open core” model has also proven historically to be a weak business model. Those that use it may stay afloat, but they don’t grow the way popular open source software such as Linux or Python do. The usual explanation for this is that users don’t find the open part of the software useful enough on its own, and don’t want to contribute to it because they feel they are just helping a company build its proprietary business.
Wonks to the Rescue
It may be that Arzt–and others who want to emulate his model in health care–have to foster a policy change in governments. This is certainly starting to happen, as seen in a series of policy announcements by the US government regarding open source software. But this is a long road, and direction could easily be reversed or allowed to falter. We have already seen false starts to open source software in various Latin American governments–the decade of the 2000s saw many flowery promises these, but hardly any follow-through.
I don’t like to be cynical, but hope may lie in the crushing failures of proprietary vendors to produce usable and accurate software for health care settings. The EHR Incentive Programs under Meaningful Use poured about 28 billion dollars into moving clinicians onto electronic records, almost all of it spent on proprietary products (of course, there were also administration costs for things such as Regional Extension Centers), with little to show in quality improvements or data exchange. The government’s open source initiatives, CONNECT and Direct, got lost in the muddle of non-functional proprietary EHRs.
So the health care industry will have to try something radically new, and the institutions willing to be innovate have their fingers on the pulse of cutting-edge trends. This includes open source software. HLN may be able to ride a coming wave.