Health professionals for decades have been admonished to provide only treatments that are "standard of care." In fact, countless medical malpractice suits have been filed against clinicians who for one reason or another deviate from that standard.

The expression itself implies that a one-size-fits-all approach is the best way to manage patient ailments. Medical textbooks have encouraged this mentality, suggesting that there's one standard healthy human body, with organs that are more or less the same size and glands that more or less secrete the same quantities of hormones and enzymes. But if you were to observe a few thousand autopsies, you'd discover there's a lot more individuality out there than some authorities care to admit.

Clinical analytics specialists are now joining forces with forward-thinking clinicians to explore these differences in the hope of coming up with more individualized treatment protocols. Two innovators come immediately to mind:

C. William Hanson III, MD, chief medical information officer at the University of Pennsylvania Health System, recently told me about the hospital's efforts to customize patient care. Like most healthcare facilities, Penn has its share of readmitted patients. And since Medicare is now penalizing hospitals for preventable readmissions, hospitals have a strong incentive to find ways to fix the problem.

[ Today's mobile devices have transformed medical care in unprecedented ways. Find out how clinicians are using these tools in the InformationWeek Healthcare Webcast: The Mobile Point of Care: Making the Right Choices ]

In an effort to predict patients' risk of readmission within 30 days of discharge, Penn recently reviewed data from more than two years of inpatient admissions. It examined the relationship between the number of prior inpatient admissions and ED visits and patients' history of being readmitted within 30 days of discharge. That analysis let Penn develop a prediction rule that was sufficiently sensitive and specific to meet its needs.

The analysis indicated that the greatest risk existed for patients who had several previous admissions to the hospital system in the 12 months preceding their current inpatient admission. So Penn's prediction rule inserts a red "readmission risk flag" in providers' patient lists, alerting them that specific patients have a greater than average chance of finding their way back into the hospital within the next 30 days unless something is done to prevent it.

When clinicians see that red flag pop up in the EHR, an interdisciplinary team goes into action to "optimize the patient's discharge," according to documentation Hanson provided InformationWeek Healthcare.

If, for example, a patient lacks a primary doctor at the time of discharge, the hospital will help to arrange for one. If someone is identified as being at high risk, they might send him home with home monitoring equipment that can be checked electronically by their homecare nurses.

Archimedes is also doing its part to optimize patient care. The health modeling organization has collected massive amounts of data from clinical trials, physician interviews, physiologic research, and epidemiological studies, and used that data to simulate patient populations and recommend customized interventions.

As Archimedes explains in its literature, "The Model can create simulated populations that match real populations at a high level of detail. If person-specific data are available, the Model can create copies of real people one by one, matching specific individuals on more than 40 clinically relevant variables."

The vendor's work with Kaiser Permanente recently demonstrated that this isn't just pie-in-the-sky technology.

Physicians at Kaiser Permanente suspected that a drug cocktail consisting of aspirin, lovastatin, and lisinopril (ALL) would benefit patients with diabetes and coronary heart disease (CAD) and lower their costs. But they had no proof. They turned to Archimedes, which simulated a patient population over the age of 55 with either diabetes or CAD in Kaiser Permanente's California region, using data from the health system on drug costs, lab tests, office visits, and admissions for MI and stroke.

Then they simulated two scenarios, one in which patients received the drug cocktail, and one in which they got conventional care, and they calculated clinical outcomes 30 years out. The clinical analytics engine predicted the ALL protocol would reduce MIs and strokes by about 71% and reduce diabetes-related costs by about $500 per person per year.

Kaiser Permanente ran with the projections and put the ALL protocol into use in 2004. By 2007, its admissions for MIs and strokes decreased by 60% in "low intensity" patients, that is, the group that was less than 50% compliant with the drug regimen. The complications were virtually eliminated in those who were more compliant. Estimated savings were impressive as well: "net savings (after paying for the costs of drugs) of approximately $38 million a year, or approximately $350 per person."

Innovations like these suggest we need to look at patient care in a new light. Put another way: "Clinical analytics eats standard of care for lunch."

Not every application is ready for the cloud, but two case studies featured in the new, all-digital issue of InformationWeek Healthcare offer some insights into what does work. Also in this issue: Keeping patient data secure isn't all that hard. But proposed new regulations could make it a lot harder. Download it now. (Free with registration.)