Question & Answer with Jeff Goldsmith, President of Health Futures, Inc.
Barry Jacobs: What inspired you to create Health Futures?
Jeff
Goldsmith: Health Futures was a hobby that turned into a business. I've
done strategic planning in health care for 20 years. I'm a technology junkie,
and I'm fascinated by biomedical science and IT. These are the principal drivers
of change in the health system. Fundamentally, health care is about seeking
knowledge. It's the most complex knowledge domain in the entire society. If
you combine the NIH budget with the pharma and biotech industries, you've got
$50 billion in R&D spent per year. That doesn't count the software development
activity that's taking place in clinical medicine, which could be another billion
or billion-and-a-half dollars. There's nothing remotely corresponding to this
level of new knowledge creation anywhere else in society.
BJ: What are your present and long-term views on clinical information technology?
JG: What we're seeing now is an exciting evolution in clinical information technology. Most of what's taken place in the past decade has focused on what I would call "digitizing the patient's chart." That is, creating an accessible digital version of the patient's chart, which is essentially a record of what's been done to the patient in the past, and their own health history. In other words, it looks backward at what has happened to the patient — medications taken, health history, etc. To get that information accessible at the point of care has been a tremendous challenge. What I see happening now is, technology is going to enable the clinical team to turn around 180 degrees. Instead of looking backward at what has happened to the patient in the past, clinical information technology will help the clinical team look forward from where the patient is right now to identify the optimal way of resolving the patient's problem. This is something that people are calling the "electronic medical record," (EMR) but that doesn't really capture what they do. They're going to function more like a navigational system in an airliner or, to use Harris Berman's marvelous analogy, a "NeverLost" in an automobile. Hertz's global positioning tool locates you where you are and gives you a path to get to where you want to go. That's where clinical information technology is going, and it's tremendously exciting.
BJ: How do you automate the clinical process?
JG: Many of the vendors are trying to build robust clinical decision-support tools into the so-called electronic medical record. When you fuse Clinical Decision Support Systems (CDSS) with the EMR, you change the way the record is used. That's because the system frames the decisions that the clinical team might make based on where the patient is at the moment. So what CDSS adds to the mix is the ability to make decisions based on the current available state of medical knowledge. Vendors will embed care pathways and guidelines into the decision support software, so that when new information is made available about the patient based on lab tests or diagnostic procedures or a post-surgical evaluation, the technology will be intelligent enough to say, "OK, here's where the patient is; this is where they need to go to achieve optimal clinical results."
BJ: What will happen to the transferability of medical records between facilities during this process?
JG: HL7 successfully standardized clinical messaging inside institutions. But standardization in the structure of the records themselves has not been achieved. Nor is there sufficient standardization to transmit between institutions in some of the key software. For example, there is no controlled medical vocabulary. This is where there are going to be compatibility and interoperability problems. I've just finished writing a paper with David Blumenthal of Harvard and Wes Rishel, who founded HL7, on the need for further federal action to encourage standardization of clinical systems so that they do in fact interoperate. It's not necessarily an area where the marketplace will create standards without a push.
BJ: What's your view on "plug and play" and the standards involved?
JG: DICOM is an extraordinary success story. The radiology profession and the relatively small number of manufacturers of radiology equipment were able to agree on a standard for digital imaging files. The government wasn't involved in this, to my knowledge. It was a voluntary process, and the fact that all of these different hardware devices — hardware manufactured by a lot of different manufacturers — are generating information that can be moved through PAC systems or through broadband connections literally anywhere in the world has revolutionized radiology. It has abolished many of the geographic constraints on radiologic practice. It has made a 24-hour radiology practice possible. Similar standards for information are evolving in the clinical laboratory arena. That's an area where you haven't seen the same degree of standardization. There's tremendous leverage in getting systems to generate compatible messages and to be able to move information across the system easily. These examples, and the success of HL7, should encourage vendors and users to get together and collaborate to make this possible with all clinical information.
BJ: How far away do you think we are from a digital, paperless health care system in the clinic? And what do you think some of the obstacles are? Why hasn't it happened sooner?
JG: The technology is in our grasp today; a few institutions are already paperless. Two of the Mayo Clinic sites have made clinical activity paperless. The University of Illinois at Chicago Circle Medical Center is a completely paperless environment. Institutions have already achieved it. Huge changes in how clinicians interact and how medical decisions are made have occurred. The physicians practicing in those settings can't imagine what it would be like to go back to the old way of doing things. What stands between us and a paperless environment is easily $100 billion in investment, and a tremendously painful, literally, order-by-order re-examination of how clinical workflow operates. It would be idiotic for an institution to simply go in and digitize all their existing order sets. What many vendors are saying — and I'm hearing this particularly loudly from Cerner and Siemens — is you need to rethink the clinical workflow based on the notion that you can make clinical information available instantaneously; you can substitute machine intelligence for human intelligence, and machine memory for human memory. The challenge is to rethink how care decisions are made and to redesign how the clinical team operates. Simply to digitize what we have now would be a colossal waste of money.
BJ: How do you think wireless access will further change this environment?
JG: It's really going to help the physician, because physicians are in motion. To put, essentially, a physician's desktop in a portable device, and to connect physicians in real-time to the flow of new information about their patients would help ease the logistical burdens that physicians have today in managing their practices. The problem is the devices themselves: Is the software robust enough to handle applications this complicated? There are huge data security issues with wireless transmission that we're seeing in Wi-Fi because people can simply pick up the transmissions, and you've obviously got the HIPAA issues about protecting the security of clinical messages sent through a wireless mode. The hospital itself is a laughably hostile environment for wireless. You've got an almost lethal amount of RF signals in the air at a typical hospital; you've got all of the structural steel; you've got lead-lined rooms. It's going to be a tremendous challenge to bring wireless into the hospital, and I still think that's a ways off. The key thing for wireless is that many physicians want to be able to practice anywhere, and we really ought to enable them to do that. They shouldn't have to be at a desk in order to practice medicine.
BJ: What about voice recognition? Where do you think that's going?
JG: The real problem there isn't voice recognition or speech recognition. The problem is in the next step — making the information that emerges from a speech-recognition engine digitally actionable. In other words, what does the information mean? It's that step that's holding back the ability to eliminate typing or point-and-click or the "pick" lists tool set that people use to make clinical decisions. I'm going to defer to my colleagues at Gartner on when this is going to happen. But the real constraint is the issue of natural-language processing and applying intelligence to the stream of words that comes out of speech recognition. We're still a long way off from achieving it. Having said that, there are many companies out there creating intelligent voice-response tools for health care institutions. People wrote off the telephone when we had this tremendous interest in the Internet. But voice communications will be an important channel of connection. We're going to be surprised at how powerful a tool sophisticated voice-response technology will be in filling in some of the gaps in continuity of care.
BJ: Tell me about your new book.
JG: The book is Digital Medicine: A Guide for Healthcare Leaders. It explains how the major streams of innovation in information technology in health care are going to converge over the next decade in an intelligent clinical management system accessible to caregivers, patients, and their families, and how that emerging intelligent clinical system is going to be used, and the changes that it's going to make.
