A blueprint for high-volume, high-quality lung cancer screening that is detecting cancer earlier—and helping to save lives
It is oft lamented that America's health care system lags behind other segments of the economy in its use of information technology. As a medical informaticist, Ross D. Martin, M.D., M.H.A., makes it his life's work to erase that deficit. Martin is a postgraduate fellow in medical informatics at the Decision Systems Group, a medical informatics laboratory established in 1979 by Harvard Medical School and Brigham and Women's Hospital in Boston. The group is now part of the Harvard-MIT Division of Health Sciences and Technology.
Martin interrupted his residency in obstetrics and gynecology at Bethesda Hospitals in Cincinnati after two years to pursue a master's in health services administration at Xavier University. After completing his master's residency at PacifiCare of California, Martin worked as an information technology (IT) consultant in managed care. He focused on assessing the feasibility or applicability of new health technologies in managed care.
Martin received his medical degree at the University of Cincinnati. He is editor of Essential Informatics for Medical Groups, a monthly newsletter for physicians and practice managers in small and midsized medical groups. His permanent e-mail address is: firstname.lastname@example.org
Senior Contributing Editor Patrick Mullen discussed the effect of information technology on medicine and managed care with Martin, appropriately enough, in an exchange of E-mail messages.
MANAGED CARE: How much credence to you give the view that physicians are technophobes?
ROSS D. MARTIN, M.D.: Physicians span the spectrum in their comfort with technology, like everyone. There are never-users, first-adopters and everything in between. What perhaps differentiates physicians from everyone else is their relative lack of patience because of the overwhelming time demands placed on them. Physicians demand technology that seamlessly integrates with their workflow. I've never met a physician who feels uncomfortable behind the wheel of a brand new, high-end automobile because he fears the advanced technology it contains. That car operates like every other car, only better and faster. The technology is hidden behind a carefully designed interface.
MC: How well have software designers created elegant and efficient tools for physicians?
MARTIN: Software is becoming increasingly sophisticated and more user-friendly — both in the medical realm and elsewhere. Much of the improvement can be attributed to the general effects of Moore's Law. [Editor's note: Intel Corp. co-founder Gordon Moore observed in 1965 that the capacity of state-of-the-art integrated circuits roughly doubles every year (though this has now slowed to 18 months). If this trend continued, Moore noted, computing power would rise exponentially over relatively brief periods of time.] We have the data-storage capacity and the processing speed to expend on better interfaces. We're learning a great deal about natural-language processing, so our ability to search for the right information based upon simple queries is improving. But medicine is still unique. There are special issues surrounding the physician-patient encounter that need to be considered when designing any system. The medical knowledge domain is far more complex than most other fields. There are still many vocabulary issues (do you call a heart attack a myocardial infarction, MI, or coronary artery occlusion?) and data-structuring issues (what units do you use to define a blood clotting study?) that need to be resolved.
MC: How has clinical care been hurt by the lack of appropriate information technology tools?
MARTIN: During residency, I saw all sorts of examples where patient needs were not met because the right information was unavailable at the right moment. The most common and most frustrating scenario involved calling a physician about a hospitalized patient who was having trouble and hearing, "That's my partner's patient; I don't know anything about her." It's unconscionable for medical groups not to have some system of cross-coverage that includes information access about patients you're supposed to be covering. Other typical examples included lack of access to prior medical records or images, which resulted in wasted time and duplication of effort.
MC: How would you persuade a group of physicians who discount computers as administrative, rather than clinical, tools that it's in their interest to invest in information technology?
MARTIN: They need to look at advances in computer technology as an opportunity rather than a threat. Doctors feel battered by technology because payers were faster to recognize the benefits of using it to improve their bottom lines. Their tools sometimes contain crude and perhaps inappropriate discriminators, but they are getting better all the time.
MC: Can you point to examples of profiling tools that are showing promise?
MARTIN: The ones I'm familiar with are not very good examples. Second hand, I know of a system in Utah that looked at urologists and infection rates. The profiles were so good and so discriminating that they told the developer to take a hike. He had seen through the emperor's new clothes! The audience wasn't ready to hear the message, so they killed the messenger. Physicians need to become leaders in this area and take advantage of these tools to better their positions. I suppose you could look at it as a technological arms race, but a better approach is to view it as an opportunity for collaboration instead of retaliation. Tools for credentialing and profiling physicians could be so much better with more comprehensive data.
MC: What kinds of data, and from where? How do you allay concerns of physicians who feel they've been burned by crude profiling tools?
MARTIN: A typical complaint that physicians have about any profiling system is that it does not adequately account for special situations and severity. "Well, sure, I have a 45-percent C-section rate, but my patients are high risk." The more specific data you have that can discriminate and measure severity, the better your results. The computerized patient record, when fully implemented and given time to become information-rich, can be a source for these data. Computerized order-entry systems in hospitals can be another. What percentage of the time did Dr. X fully comply with this clinical guideline for dosing gentamycin? You can find the answers to these questions. Physician buy-in is terribly challenging, but can be improved if physicians are part of the process of deciding what types of profiling should be done. Also, if the tools are used in a way that isn't censuring, but for discovering and working through problems, that will go a long way in making the process more acceptable. Just make sure, if you are going to use measures to effect change, that you've got enough numbers to prove a trend that's more than just a random occurrence. In one case in New York City, an insurer is profiling physicians on customer satisfaction, utilization management, and HEDIS measures, but isn't acting on the results yet. For now, they're just sharing the information with providers. At some point, once they're confident that they have sufficient data to prove a trend, they may check in with some physicians to inquire about why their numbers are so far off the norm.
MC: How does information technology change the nature of being a physician?
MARTIN: Physicians need to see their role in medicine as information brokers, where clinical information is the currency of exchange. Once that happens, every application of information technology will make them more successful — both in terms of quality of patient care and their financial bottom line. Brigham and Women's Hospital has demonstrated the quality and monetary value of information technology a dozen times over. Its online pharmacy-alert system has saved the hospital millions of dollars, avoided hundreds of potential adverse drug interactions, and has almost certainly saved lives.
MC: The electronic patient record has been just around the corner for a long time. How long will it be before medical records are as routinely electronic as bank records?
MARTIN: It seemed to IT specialists in other industries that all medicine needed was a few good programmers, and we'd be dispensing health care from 24-hour terminals in malls. It quickly became clear that medical knowledge is different. Not only is it extremely vast, but it is constantly evolving. Which raises questions for the electronic medical records: How "granular" should the record be? The more detail you structure, the more valuable the data becomes. That structure — or granularity — comes at a high price, but that can be overcome.
MC: How good a model is the paper medical record for an electronic medical record? Is the challenge of digital medical records mainly one of moving existing information forms into a new medium? Or does the nature of the paper chart need to be re-imagined?
MARTIN: You can publish a document on the web by simply transforming a paper document into an electronic one. The result may have some utility, but it often fails to take advantage of the medium. Effective web publishing takes special thought and design work. Similarly, you could use an exact model of the paper record and make it electronic, but it wouldn't be worth much. If we had never created a paper medical record and were designing an electronic one from scratch, I doubt it would look anything like the paper record of today. Our reality is that we have a world of paper records and a world of doctors who are used to them. The reason graphical user interfaces — GUIs — like the Windows and the Macintosh operating systems, work is because they made computers more usable by incorporating the familiar. Need to throw something away? Put it in the trash can. Need to jot down a message? Type it on a sticky note. The most effective implementations of electronic medical records are the ones that reflect the current system of paper records while incorporating much of the functionality that is made possible by converting to an electronic format.
MC: What's holding up the transition to electronic medical records?
MARTIN: One of the barriers to true synergy in this industry is the lack of standardization. It's not that we don't have standards. The great thing about standards in medicine is that there are so many to choose from! Because there are not just three or four dominant players in medicine, achieving consensus is difficult. When the music industry realized the value of digital recording, it didn't take long to come up with a standard for compact discs, because all of the decision makers fit around one table.
MC: Who needs to be at the table in health care?
MARTIN: First you should ask, "Who's got a table that big?" HL7, for Health Level 7, is the primary messaging standard in health care, and is the name of the group developing the standard — is working hard at implementing its next version. They gave a nice demo at the Health Information and Management Systems Society conference in Atlanta back in February. They used XML (eXtensible Markup Language) as a format for transferring messages among different vendors' information systems — 3M, Epic Systems, Oceania and MedicaLogic — to show that clinical messages could be seamlessly transferred among systems using this standard. Since money makes the world go 'round, or at least travels around the world fastest, we seem to have figured out how to exchange payment information — with administrative details tagging along — much more easily than clinical information. The HL7 demo showed that it is possible to exchange payment, clinical, and administrative information from disparate sources. Another barrier is privacy and confidentiality. I sat next to a woman on a plane yesterday whose family was registered in a lung-cancer database at the Mayo Clinic. She was reluctant to include her own information, despite their assurances that it would be secure, because she feared the information might someday be used against her.
MC: How valid are worries like hers about the confidentiality of electronic patient records vs. paper records? Is the problem really more cultural (that is, fear of the new) than technical?
MARTIN: Electronic records are potentially far more secure than paper records because access can be controlled at any number of levels. Not only can you decide who can see the record, but also what part of the record that person can see. Access to records can be monitored more completely. The paper record is available to anyone with a clipboard who looks as if he knows what he's doing. Of course, any time you have something of value, there is a worry that someone will try to get at it, especially on a wholesale level, and access thousands of records instead of just one.
MC: And once thousands of records have been collected, there will be a temptation to sell some of that information. A few months ago, South Carolina and other states tried selling databases of digital driver's-license photos to a company that clears check-cashing transactions, until public uproar stopped them. Biomedical researchers would have interest in tapping medical databases. Should they be able?
MARTIN: Absolutely they could use this information — those data have potentially enormous value. Some models of the electronic medical record and the clinical-data repository recognize this value and, appropriately, assign ultimate ownership of the data to individual physicians. These physicians can reap financial benefits by conscientiously maintaining their medical records. While the physician gains some benefit from keeping detailed records, much of the payback comes further downstream. I don't see anything wrong with compensating physicians for the extra time it takes to capture the type of information that is valuable is on the aggregate, not the individual, level. These data can be scrubbed of individual identifiers without compromising a person's privacy. Right now, there is reason to be concerned about someone having access to your medical information. Once we as a society agree we can't discriminate based on genetic information or medical history, we'll be less concerned about who could try to hack in and get information about us. Any fairly motivated individual could get hold of my credit report in about 10 minutes, but I don't lose sleep over it. Somehow we as a society have decided that we value the ability to get money out of an ATM at 2 a.m. more highly than our right to absolute privacy. Once we all recognize the incredible value of having large collections of health care data — for research, for individual care, for understanding best practices and outcomes — we will see tremendous movement toward electronic medical records.
MC: Is it your sense that physicians who are reluctant to use computers feel that software tools serve the interests of payers, rather than the interests of physicians and patients?
MARTIN: Physicians in general are an unhappy lot because their roles as kings of their own domains have rapidly eroded. They are being questioned on all sides — by payers, regulators, patients, credentialing organizations. It's easy to put the onus of responsibility on the nasty managed care company rather than look in the mirror. In this respect, some physicians see technology as just one more tool for others to use to badger them.
MC: With some justification, wouldn't you agree?
MARTIN: Of course there's justification for the attitude, but perhaps not the response, which has been fairly reactionary among physicians as a whole. Instead of recognizing IT as an opportunity, most physicians have perceived it as a threat and have either stuck their heads in the sand or have used obstructionist tactics. Both are short-sighted. At the same time, early purveyors of information technology didn't view the physician as an important customer, but someone they needed to bully into submission and compliance. The role of the physician is clearly changing, appropriately, from that of dictatorial autocrat to team leader. It is still important to fit the technology to the customer, not the other way around.
MC: How have managed care plans helped or hurt the adoption of information technology in health care?
MARTIN: They've certainly been drivers in the application of information technology, because they recognized the value early on. The way they've hurt progress is by creating adversarial relationships with physicians and by failing to adopt standards. You can create the best system in the world, but if it is incompatible with other systems and can't share information, all it will do is make my world more complex.
MC: How can physicians and health plans use the Internet to be more productive?
MARTIN: Excuse the shameless plug, but that's what my newsletter — Essential Informatics for Medical Groups — is all about. I'm fond of M.D. Consult, developed by three major medical publishers and which puts searchable full-text journal articles and complete textbooks on the Web. It's available by subscription, and can also be licensed by organizations. Applications like this, which provide the right information at the right time, are perfect for the Internet.
MC: The National Institutes of Health is talking about establishing a single electronic repository for peer-reviewed biomedical research papers. NIH may include some form of peer review, potentially rendering peer reviewed journals redundant. What are the implications of such a database for the dissemination of knowledge?
MARTIN: Having an electronic repository of all full-text journal articles that can be accessed directly from a Medline search will be a great thing. M.D. Consult is one of several efforts that do this on a limited level today. Right now, when I want to get an article for research purposes, I still have to head over to the medical library, hope the journal is on the shelf, then spend 10 cents a page making a copy. I'd rather pay the publisher a dime a page for royalties and get an electronic, searchable, readable document in my computer, instead of paying Xerox or Canon for paper and ink. The money/rights exchange issue — what's known as digital-rights management (DRM) — hasn't been resolved. Once we've perfected the ability to transmit a document with ownership rules embedded in it — how many times I can look at it, when, if ever, it expires, how many copies I can make, etc. — the electronic commerce part of information exchange will make incredible advances. I don't think peer review will go away. If anything, the ubiquity of information makes branding even more important. I want to know that the New England Journal of Medicine has reviewed this information and thinks it's worth reading. As for a time line, the enabling technologies are not going to be the drivers as much as are people's and publishers' acceptance of the DRM concept. Right now, publishers charge what I consider to be unreasonable rates for electronic copies of articles. They know that one electronic copy can turn into 2,000 copies as easily as it can turn into two. Once it becomes relatively cheap and convenient to pay the piper — the publisher — for just the information I need at the moment, that's when the change will happen.
MC: What other useful tools are available today?
MARTIN: A big growth area for health care is in use of intranets — where organizational data are made available to individuals within the organization. We've learned from the Internet that access to information is only valuable if that information can be appropriately filtered and presented. When I look for "headache" on the Internet, I don't want the web site of some teenage grunge band in Seattle. But that's what I might get in the first dozen hits. Intranets use the same technology but keep the information focused at the appropriate level. E-mail is an excellent addition to the doctor's bag. We're doing this interview as an E-mail exchange. So even though I'm out of town right now, I can continue this dialogue, taking advantage of a spare moment that I didn't know I would have. No more phone tag! There are some special considerations that apply to the transmission of specific medical information over E-mail, but those issues are simple to address.
MC: How did you come to be interested in medical informatics?
MARTIN: I started off as a computer science and business major in college, seeing myself as a liaison between business and technology. I eventually got sidetracked into political science and medicine. I had always thought that I would eventually migrate over to the business-and-technology end of medicine — maybe 10 years into practice. I took the opportunity afforded by a flexible job as a house physician to get a master's in health services administration. After that, I started doing consulting work in managed care. There came a point when I had to decide whether to return to residency and become board certified. I decided instead to get more training in medical informatics, which is what I'm doing now at Harvard/MIT. I saw that I could better serve patients as a champion for applying information technology to medicine.
MC: Thank you.
A blueprint for high-volume, high-quality lung cancer screening that is detecting cancer earlier—and helping to save lives
Multiple Sclerosis: New Perspectives on the Patient Journey–2019 Update
Summary of an Actuarial Analysis and Report