In fact, he says a recent Beth Israel random phone survey of 2,000 adults in eastern Massachusetts revealed 19 percent of respondents would change to an e-enabled doctor if they had the choice. “That's a major forcing mechanism,” says Halamka.

Harris offers a more bottom-line example of how these PHRs are a boon to hospitals. He estimates that 2,000 of the 8,000 patients a day the hospital treats have no electronic record with the Clinic in any form. Should the technologies become as popular as their advocates hope, Harris says even those patients whose physicians do not use EMRs could still have critical data stored on a Google or Microsoft-type PHR and available to hospital staff on their first encounter.

“Even a physician who's still using a pen and paper in their office sends their patient to a retail pharmacy or commercial lab or hospital where that information is available in electronic form,” Harris says. “So I can know their basics; their allergies, their medication list, and diagnostic testing, which is an enormous step forward for us when we're seeing one of those 2,000 patients.”

The possibilities of patient-controlled PHRs have also reached hospital directors. Weider, who also thinks the Dossia effort led by some large employers (including Wal-Mart) could be a major player in PHRs, says the Ministry board recently asked him if patients' records could be easily portable, even if they switched physicians with the new technologies.

“I said that, for the first time in my life, I have optimism about that,” Weider says. “We've been stuck with a lack of standards and no one organization big enough to drive it, but now I see these three and, in the next few years, there's going to be a major sea change.”

Uncertainties and imperatives

Even the most optimistic backer of the Google and Microsoft model concedes that sea change will be slow, however. Near-term, both concepts have to work through the pilot stage. Harris, for instance, says the Google Health UI had to be tweaked early on in the pilot to make it more user-friendly; and Dunbrack says the HealthVault setup is also quite complex.

In fact, in order for a prospective user to see how it works, they are first asked to set up an account — which might be off-putting for consumers wary of Microsoft's no-holds-barred competitive philosophy. In fact, even a technology veteran like Stofko still has some “show me” attitude toward Microsoft, which has seen another central Web-wide repository idea (its Passport identity architecture) fall flat.

“Over the years, we've seen Microsoft in and out of healthcare,” he says. “I think before, the perception was they created a healthcare division — at least this was my perception — because a hospital is typically in the top five employers in any city, and that's a huge number of individual desktop licenses. I think they're past that, and with the people they've brought in from the healthcare industry, they've put some meat and teeth behind it. Now, they have to prove the longevity of their commitment, so I'm not past that hurdle, but I am past the, ‘We're just in it for software seats’ skepticism.”

There are perhaps two factors that might determine the ultimate success of the two platforms. The first is patient acceptance. Dunbrack, for instance, says that in the last survey Health Industry Insights did on PHRs, more than half the respondents had no idea what the concept was. Some patients might also hesitate to post data online, as early critics of the technologies warn neither Google nor Microsoft is subject to HIPAA regulations. Both Halamka and Dunbrack, however, say both companies are well aware of the commercial ramifications of a data breach or even a whiff of unbidden commercialization in either platform, and are sanguine, because safeguards are in place.

“Obviously, Google and Microsoft are selling trust,” Halamka says. “That's really all they're selling in their PHRs. If they abrogate that, they will have lost their business model.”

The technological factor that might affect adoption is the platforms' ability to make all the data posted relevant to everyone who looks at it, without making it too simplistic for clinicians or forcing patients to learn ICD coding. Both Halamka and Dunbrack see great opportunities for developers of translation middleware and natural language processing applications. Halamka and Harris say early ambiguities around terminologies will also be offset by the platforms' abilities to receive dynamically “pushed” data from hospital systems, and that there will be enough overlap so clinician and layman will find the platforms useful.

“It's not true for everything, but for basics like medications coming from a pharmacy and lab results it's already in the coded formats I need,” Harris says. “What's not going to be there Day One, and may not be there for a long time, is structured notes and other things where you would have to have some adoptive structured approach to documentation, which is a challenge I don't think we're quite ready for.”

Harris also says unstructured data supplied by patients will be valuable. “I still want to see it, I just won't post it. I'll do that interpretation and put it in my EMR, which is precisely what's happening in the office today.”