11 ways to digitize health care of the future - Next!

Physician Executive, July-August, 2003 by Joe Flower

Grand Amsterdam Hotel, Netherlands-The new and old clash and energize each other here in this vast ancient congeries of buildings. It was a convent in the 15th century, a hostel for great lords in the 16th, seat of the admiralty of one of the world's great sea powers in the 17th and 18th century, seat of the city government in the 19th and 20th, now in the 21st, a hostel for great lords of commerce.

I am touring the elegantly paneled council room where I am to speak to a meeting of health care distributors, looking around at the thin, modernist chandeliers and the medieval saints painted behind the dais.

The sense of ancient and modern is sharpened by the fact that I am not actually in Amsterdam yet. At the moment, in fact, I am sitting on my back porch with my wireless laptop. I typed in www.thegrand.nl into a browser. I am peaking around the meeting room remotely, using the Web site's "panocam."

Step back for a second and think about that. A 600year-old Dutch establishment markets itself, in moveable pictures and text, through a global electronic network, to a guy on his back porch in a seaside town in northern California.

No one would have imagined any part of such a scene 10 years ago. Digitize commerce and put it on the Web and things change.

Back to the future

Step forward another 10 years. Imagine a time when we have completely digitized the inner workings of health care. What changes? What could we do that we cannot do now? What could we do that we don't even imagine now?

I can think of 11 ways that health care will be able to change when digitized medical records, images, and lab results are standard and all medical devices and personnel are connected--with proper safeguards for privacy and accountability. There will be many more ideas that have not yet even popped into anyone's imagination.

I. Massive database

Digitize all medical records. Use standard codes for each field, such as disease category, age, complications, drug used. Allow them to be accessed online (with special codes that only allow the patients' own clinicians to see personal identifiers). Presto! You have created a massive, distributed, n-dimensional database of real-world cases that can easily be queried. You can design and run your own instant, customizable, retrospective clinical studies. How will my patient do on this drug at this dose? Find me 10,000 patients exactly like mine and find out how they did.

2. Just-in-time information

Digitize the physician's information gathering (picture the doc at the bedside tapping radio buttons on a wireless FDA and scrawling a few notes). Connect the physician's PDA, in real time, to a database of the peer-reviewed medical literature armed with pattern-recognition software. The result is an instant, omnipresent digital adviser. The doc taps in the patient information: "female 73" and gives a low LDL cholesterol level. The system recognizes part of a pattern and asks for the blood albumin level and gives the pattern: "Low LDL + female + >70 + low blood albumin = high risk sudden death." It then offers the appropriate peer-reviewed literature and recommended therapies. It gives the physician the information she needs, when she needs it and leaves the judgment up to her.

3. Virtual physician extenders

Connect the same database of the medical literature to decision-tree software and you have a powerful medical decision-making tool. Ask a few basic questions: Fever? Tenderness below the right first rib? Pupils unusually dilated? Move on to clinically more complex questions: Spleen enlarged? PSA antigen present? Answer questions in any order, skip the ones you can't answer now. At any stage, you can ask the software for a set of possible diagnoses, with probabilities assigned to each: 73 percent chance it's a cold; 26 percent chance it's flu, 1 percent SARS and here are the tests or questions that would differentiate further. A nurse or a minimally trained receptionist could use such software to set up the case for the doctor. In remote places where doctors are rare, the software can actually take the place of much of what a doctor does and quickly discover whether a patient must be evacuated to a distant hospital. One such software product already exists, the Problem-Knowledge Coupler (www.pkc.com).

4. Shifts in clinical practice

Connect doctors and their shared experience together through such devices as our first three examples and through Web sites, email and other online resources, and medicine becomes both more collegial and more evidence-based. Physicians also become quicker to adapt to new information and find it harder to stick to outmoded ways when new practices have proven more effective. Medical practice will become more dynamic. The feedback loop between research and actual practice will become much smaller, tighter and faster. The possibility that good research will be lost in the noise of publications will be reduced.

5. Virtual patients

Model a human being in software, from the eyes and toes right down to the intracellular biochemistry. (The Oak Ridge National Laboratory is working with a consortium of government agencies to build just such a model). Now make the model customizable to resemble your particular patient, from the age, gender and medical history down to the peculiarities of the patient's DNA. Now you have your patient instantiated in software, in your computer. Try drugs on the virtual patient before you try them on the real one. Rehearse difficult surgeries on the 3-D visual model. Run different scenarios for the immune response or fracture repairs before you put the real patient at risk.