If you were to come up with a really outrageous, game-changing idea, one that you're almost certain your colleagues would laugh off the table, would you have the confidence and guts to push ahead with it anyway?

I would love to have been a fly on the wall when someone at Samsung suggested developing the first large-screen smartphone. I can only imagine a company VP insisting that no one would carry a phone with a 5.3-inch screen. In fact, when Samsung introduced the first Galaxy Note a couple of years ago, several critics sloughed it off. One insisted: "You will look stupid talking on it, people will laugh at you, and you'll be unhappy if you buy it." Fortunately, Samsung had the courage and vision to ignore the conventional wisdom and start making what has become a hugely successful line of hybrid phones, the so-called phablet.

Likewise in healthcare, what does it take to get a fresh service, process, or device off the ground, one that could truly improve patient care? Fearlessness is needed, as is a deep knowledge of end users' habits and needs. But there's another ingredient: a subversive mind.

[Should healthcare adopt a more open, web-based approach to health information exchange? Read A Serious Proposal For Healthcare IT Interoperability.]

Like other industries, the healthcare sector has lots of received wisdom, including the notion that the free market system will solve all our problems, including the lack of system interoperability that so many clinicians complain about as they try to exchange patient data across institutions. Despite those complaints, most EHR system vendors continue to hold their proprietary ground.

I hate to recommend more government regulation, but it may be time to declare martial law. The Food and Drug Administration has classified certain types of mobile health apps as medical devices, requiring them to obtain FDA clearance. But the agency has stopped short of classifying EHR systems as medical devices. If the industry can't get its act together, it's time for the feds to step in and insist on uniform data standards, before poor interoperability contributes to a major health incident, such as the outbreak of an infectious disease.

Which brings to mind the Ebola debacle in Dallas. A faulty EHR system may or may not (depending on who you ask) have contributed to the Ebola patient's mishandling at Texas Health Presbyterian Hospital. But an EHR-dependent disaster isn't far off if systems from one hospital to the next can't communicate with one another. FDA oversight of EHR interoperability may not sound like a "subversive" idea, but in today's environment, if it improves data sharing, communications, and patient safety, it fits my definition of the word.

Even more subversive is the notion that data mining can replace large-scale, expensive clinical trials to identify new medical therapies. For decades, medical academics have insisted that treatment protocols be based on double-blind randomized clinical trials, but it's getting harder and harder to find the money to fund these experiments.

As far back as 2004, researchers demonstrated that sifting through large patient databases can help pinpoint effective treatments. For example, when J Kim and associates from the National Cardiovascular Center in Suita, Japan, analyzed the records of 1,100 patients with congestive heart failure -- looking at the drugs they were given as well as their echocardiogram and lab readings -- they identified several specific families of effective drug treatments. Their findings agreed with evidence from several previously conducted large-scale clinical trials, suggesting that it may have been possible to avoid the trials and rely on the data mining results to begin with.

Beth Israel Deaconess Medical Center has taken the mining of patient data to another level, using it to gain deeper insights into disease causation and treatments. In 2012, InformationWeek selected BIDMC as the No. 1 company in the InformationWeek 500 ranking, in large part because of its innovative Clinical Query system, which includes the records of more than 2.2 million patients and 200 million data elements.

Investigators now use Clinical Query 2 to test research hypotheses without the need to recruit patients, said Dr. Griffin M. Weber, director of BIDMC's Biomedical Research Informatics Core, in a recent email exchange. BIDMC and other Harvard-related hospitals are now joining hospitals nationwide to create an even larger database. Said Weber: "These networks give researchers access to data on tens of millions of patients, which is needed to study rare diseases or to look for subpopulations in common diseases that respond to treatments differently."

An outrageous idea? Only to those without the guts and confidence to push ahead.

Considering how prevalent third-party attacks are, we need to ask hard questions about how partners and suppliers are safeguarding systems and data. In the Partners' Role In Perimeter Security report, we'll discuss concrete strategies such as setting standards that third-party providers must meet to keep getting your business, conducting in-depth risk assessments -- and ensuring that your network has controls in place to protect data in case these defenses fail. (Free registration required.)