Copyright 1996 by Craig Feied and Mark Smith
What do we mean when we say "the Internet"? Do we mean the wires that connect one computer to another? The fiber-optic cable that connects local area network hubs from city to city? The satellite linkages that connect hubs from country to country? Do we mean the low-level protocols that allow us to send packets of information across those wires with confidence that they will arrive where we want them to go? Do we mean mid-level protocols that let computer programs send files and messages across those wires, using the low-level protocols? Perhaps we mean the end-user computer programs that give us an easy-to-use interface that lets us get text and pictures from some remote location, across those wires, using those mid-level and low-level protocols?
The fact is, all of these definitions are wrong. No matter which of these definitions we were to adopt, we would be forced to accept the fact that the Internet, as defined, would cease to exist in the near future. The cabling, hardware, and software that we use when we use the Internet are all undergoing rapid change.
No matter how the underlying transport and the top-level interfaces change, the essential power of the ability to connect to other people's information resources using the Internet will remain the same. This is what is essential about the Internet: the ability to give other people information that we possess and the ability to get from other people information that they possess, without regard to the distance that separates us. We use the word connectivity to describe this ability to exchange information at a distance. The Internet provides a previously unavailable level of connectivity to just about everybody who can get access to a computer.
The mere possibility of having simple connectivity is not, by itself, the key to the revolutionary changes that are coming in emergency medicine. For many years I have been able to create a connection between my computer and another physician's computer at some remote site, and have been able to transfer documents and images at high speeds. Unfortunately, only a very small number of people could achieve this. It was expensive, and it required a high degree of specialized knowledge. File formats were nonstandard: I could receive a picture from my colleague in San Francisco, but I might not be able to view it without some complicated conversion.
Several things are at work in today's Internet to raise our level of connectivity. One is the near-ubiquity of connectivity. I can walk up to any internet-connected computer anywhere and can exchange material with any other similarly connected computer anywhere in the world. Attaining Internet connectivity is so easy that the number of connected computers is increasing according to a logarithmic function. Connecting an existing computer requires nothing but a $99 modem and some free software, and asks nothing of the user but a willingness to learn something new.
Another important factor raising our overall level of connectivity is the ease of use of Internet connectivity. Making a computer connection is easier than making a telephone voice connection. To connect to Moscow via a voice telephone line you need to know your long-distance company prefix, plus the country code and the city code, the telephone number and the extension you want, and a 14-digit billing number. Once connected, you need to speak a foreign language to get a message to your intended correspondent. Oh, yes, and you pay by the second. To connect to a site in Moscow via the Internet, you merely click on an icon for the person or site to which you wish to be connected, and the connection is all but instantaneous.
Another important factor encouraging an increase in the general level of computer-based connectivity is the fact that there are no 'message units' for computer connections: you pay a base rate for unmetered access and don't pay extra for individual connection transactions. Once connected to the Internet, there is no deterrent to regular heavy use.
Today, this combination of near-ubiquitous, fast, easy, inexpensive connectivity means that when I want information about the latest experimental cancer treatment regimens, or when I want to know this week's CDC recommendations for immunizations in travelers to Timbuktu, I can get the information in a matter of seconds, no matter where I am. Tomorrow, ubiquitous connectivity will mean that I can get specific medical records for an individual patient just as easily.
What is cyberspace? Where is it to be found? What does it mean to say that something exists or occurs in cyberspace? When I have a conversation face to face with another person, it is easy to define precisely where that conversation is taking place: in my office, in Washington DC, in the United States, on Earth, in the Milky Way galaxy. If I pick up the telephone and call my friend in his office, it's much more difficult to say where the conversation is taking place. Is it in my ear? In the handset of my telephone? In the wires between my office and his? Perhaps it's in the satellite that handles transatlantic telephone traffic. The question becomes even more interesting if the conversation is handled digitally, because the millions of bits of data that describe my voice may travel by a variety of different paths before being re-assembled at the other end. The answer, of course, is that the conversation is not occuring in any of these places. It is useful to consider that the conversation occurs in cyberspace, an electronically defined nondimensional space that connects two or more real dimensional spaces.
The concept of cyberspace has special utility when applied to the incredible web of near-universal digital connectivity that we now know as the Internet. For one thing, we can transfer much more than just conversation. We transfer documents, photographs, programs, and datasets from place to place as easily as we can send our voices by telephone. For another thing, unlike voice connectivity, Internet connectivity frees us from the tyranny of concurrency: there is no need for concurrent involvement of both parties to an information transfer. If we have information to share, we can make it available in such a way that access to the information is at the discretion of the user. The transfer can take place asynchronously with no active involvement on the part of the provider. Equally important, the Internet allows us to 'publish' our information so that it 'exists' in multiple cyberplaces simultaneously. Because of this, the Internet also allows us to create virtual information sets: we can consolidate separate information sets so they become available through a single cyberspace location that may or may not have a physical counterpart. Many cyberspace 'sites' contain nothing more than 'pointers' to information that actually resides in fragmented form in multiple separate and physically distant locations.
The end result of all of these factors is a new reality: we have come to think of composite objects that exist only in cyberspace as if they had a physical reality, we think of sites accessible via the Internet as if they were actual physical locations, and when we exchange information with a site, we think of ourselves as 'going to' that site in cyberspace.
Of course, connectivity is not new, and cyberspace is not new -- physicians have enjoyed a high degree of connectivity for years, and many medical transactions already take place in cyberspace. Every time you pick up the telephone and hear from a patient's personal physician that an electrocardiogram was normal in the office last week, you are practicing a little bit of medicine in cyberspace. Medicine will be qualitatively different, however, when we routinely connect via the Internet in order to look at the old electrocardiogram ourselves.
An important factor that is missing in today's Internet is dependability -- our ability to count on being connected to others in our specialty. Significant benefits that seem futuristic today will seem quite natural when we take connectivity for granted.
Another thing that will change is the type of information that we can exchange with those to whom we're electronically connected: instead of just hearing our colleagues read off a cardiologist's interpretation of a patient's cardiac catheterization, we'll be able to view an actual movie clip containing that catheterization, and we'll be able to play it forwards and backwards, zooming in and out, changing contrast and viewing angle as we please. We can look at it in positive rather than in negative format, if we wish -- after all, we're not radiologists, our workplace is not dark, and most of the things we are used to looking at are not negatives of the real world.
The questions of accessibility and of control are inextricably linked. In the past, we had to rely on others to provide an interpretation of a test, and our practice was based on test interpretations that often were performed by a consultant who had never seen the patient. This will not last.
Many clinicians still in practice have never learned to read an ECG; when they were in training ECG's were mystical tracings obtained and interpreted only by specialists. When I was a student, chest x-rays were regarded in much the same way. During my residency, chest x-rays became accessible and understandable by the non-radiologist, but C-T scans were arcane and were interpretable only by a high priest. Today, CT scans are understandable, but MRI scans can only be interpreted by highly skilled professionals who never see the patient. The world continues to change: today we read our own x-rays, we read our own ECG's, we read our own CT scans, we perform and interpret our own ultrasound examinations, and for the majority of cases, we put more faith in our clinically-guided interpretation than we do in the off-line interpretation of a non-clinician who is unfamiliar with the patient.
Tomorrow we will take a new step: we will read our chest x-rays, our CT scans and our nuclear medicine scans in connected consultation with the radiologist, who may be hundreds of miles away ("I'm concerned that this area looks a little funny -- what do you think?"). We will interpret our complicated ECG's in connected consultation with the cardiologist -- and it will all happen in real-time.
The distributed medical record is already here -- doctors just haven't recognized it yet. No patient has all of his or her medical records in one place. A typical patient may have bits and pieces of medical information scattered here and there in medical offices, laboratories, and hospitals across the country. Even if all of a patient's records are at one hospital, parts of the complete record are found in the Medical Records office, other parts in Radiology, some in Ultrasound, another portion in a billing chart in the billing office, and other parts in many other places about the hospital. Even today, significant portions of a typical medical record do not exist in hardcopy. Any billing question, for example, will send a billing representative scurrying for the nearest terminal.
Many issues of security, privacy, and reliability are raised whenever electronic medical records are discussed, but these issues are neither new nor unique to electronically represented portions of a medical record. Sensitive financial data are routinely shared across portions of the Internet, even though the very survival of entire countries depends upon the integrity and security of this data. Although there have been breaches of security and losses related to electronic banking, it is a matter of record that opportunities for financial loss are much greater when manual systems are used to manage and transfer large sums. So we shall find it with medical records. All of the real security issues are soluble, and many of the issues that are hotly discussed today are nothing more than expressions of discomfort with something new.
As soon as physicians get used to the Internet as a means of being connected, we'll accept it just as we've accepted fax machines. What happens today when a colleague calls to request an old ECG on a patient who is in a distant emergency department with chest pain? You obtain the old ECG and you fax it through. You know this is a minor technical breach of patient confidentiality, but you also know that it is in the best interests of the patient to have an old ECG available for comparison. A few months from today, in some hospitals it will no longer be necessary to call for the old record nor to fiddle with a fax machine. It will be possible to call up old ECG's on the computer screen, to put a copy in a public directory under a numeric label, and to tell your colleague where to point a viewer. In a matter of seconds, the two of you will have created a distributed medical record. In 1996 your colleague will probably print out a local hard copy of the ECG to stay with the printed chart, but before the year 2000 the concept of a 'hard copy' will seem silly -- after all, the cardiologist on call will be looking at the ECG remotely, and the CCU will want an electronic copy too. In a few years most of the patient's chart will not exist in hard copy, unless you choose to print out an extracted subset for a particular purpose.
Imagine, for a moment, how it might work today. A patient will be referred by his primary physician to a neurology center, where he will have an EEG, some blood work, and an MRI scan. At the time of the visit, the neurology center will give the patient a card with their internet address and with a special patient identification number (PIN). The patient will give a copy of this card to his primary physician, and will keep the card in his wallet, alongside a similar card given him by his primary physician, together with one given him by the pharmacy where he fills his prescriptions and another one from the hospital in Chicago where he had a cardiac catheterization on a trip two years ago.
Suppose that a week later the patient travels to Washington DC on business, and develops another terrible headache. He comes to the emergency room, and he presents all of his medical cards with their internet addresses and their PINs. The emergency physician enters the internet addresses and PINs to gain access to the records kept online by each of the patient's outside medical care providers: the past medical history, prior ECG's the medication list, the cath report and a video of the cath, the MRI scan and interpretation, and all the rest. In fact, to a user of the emergency department record in Washington DC, it is not obvious which parts of the record reside where. The whole thing will seem so natural that nobody will think twice about looking at a sonogram from St. Louis alongside an MRI from Juneau. The era of the cyber-medical-record will have truly arrived.
Another way in which the Internet is changing society, and emergency medicine along with it, is in the development of cyberspace communities -- people with shared interests who live in different locations all around the world, but who are in daily communication just as though they all lived in the same small town. The concept of peer review takes on a whole new meaning when it is carried out on a regular daily basis in a secure forum with open discussion among many interested members of the same profession.
Examples of the impact of this sort of community abound: a query recently went out on a popular emergency medicine mailing list: the author was seeing a large number of patients with symptoms of influenza, and wondered where else it was cropping up, and how other clinicians were dealing with it. Pooled information from many different practitioners gave a clinical picture of the current epidemic that was very different from the official reports published in the Mortality & Morbidity Weekly Report. The use of recently-introduced antiviral agents was discussed and reviewed online in this context, with the result that a number of physicians altered their routine ptactice. The publication of an interesting article in a prominent journal often leads to a lively online discussion among a number of interested and knowledgeable people, providing a review of the material from several different viewpoints that is equally useful for those who are less knowledgeable in the particular field under discussion.
An outgrowth of specialized cyberspace communities are specialized registries of clinical information. Such specialized registries have proven their worth in the field of cancer therapy, where constantly updated pooled clinical data from many practitioners and many clinical trials helps us to decide on the optimal therapy for an individual patient. Of special interest to emergency medicine, collaborative registries have been and are are being founded for trauma, for complications of AIDS, for thrombolytic regimens as used in the transcatheter lysis of deep vein thrombosis, for tracking infectious epidemics, and for several other types of clinical data.
With ubiquitous connectivity we gain access to a mind-numbing variety of information resources. Fortunately, information technologies are already available that will allow us to develop 'smart' programs that can search multiple resources to find the answers to specific questions. The first of these 'information agents' have just begun to appear on the Internet.
Information agents will be an important aid to clinical decision making, but a much more fundamental change is coming, again riding the coattails of ubiquitous connectivity. This fundamental change is based upon the concept of object orientation, a concept that will forever change the way we practice. Ubiquitous connectivity coupled with object orientation will add 'intelligence' to what are now inanimate objects. This intelligence will reside in a portion of cyberspace that is associated with the physical expression of an object or a class of objects. We refer to the things an intelligent object knows how to do as its 'methods'.
The appearance of intelligent objects will change our life in profound ways, and medicine will be no exception. For example, today a medication is just a powder in a bottle. In a typical emergency department, that medication may be prescribed hundreds of times a year. Each time it is prescribed, a clinician must take active steps to seek out and acquire information about the correct dosing and administration of the drug, and those steps must be repeated by every clinician every time the drug is given. Before an unfamiliar drug is administered in the Emergency Department, similar information has been looked up by a student, an intern, a senior resident, an attending, a nurse, and a pharmacist. Each of these clinicians seeks and finds information about the drug from a different source: subsets of information about the medication exist in a variety of locations, and must be actively sought with 'secret' knowledge of where to look and what questions to ask. The hospital pharmacy guidelines tell the nurse how to mix it and with what other infusates it is compatible. The resident's survival guide tells how much to prescribe. To find out whether it can be given to a pregnant patient, a consultant usually is called by telephone. The dosing for patients in renal failure is found in the Physician's Desk Reference (but don't forget to look in the supplements in case there are any changes). Any interactions with other drugs the patient may be taking will be caught only by the pharmacist, and only if the interactions are well-known and well-documented.
Each clinician believes his or her portion of the information is complete and correct, and after a number of tedious look-ups, each clinician believes he or she 'knows' all important information about the drug, at which point the information becomes frozen in memory. After this point any changes in the published recommendations for that medication will not be noticed. If memory fails, an incorrect dose can be (and often is) given.
Looking back from the future, this present state of practice will seem incredible, and even barbaric. To begin with, in the future, all of the information for a drug will be integrated together and uniformly available everywhere: no drug will be marketed without a centrally-administered data resource that provides a publically available constantly up-to-date version of all known information related to that drug. What is more interesting, though, is that it will not be physicians, nurses, and pharmacists who use the data resource. In the future, the data resource will principally be used by the drug itself. This is not a typographical error, nor is it a fanciful dream. In the future, a medication and its cyberspace data-self will be tightly linked: an object and its methods. A medication will 'know' to whom it is being administered, and each medication will 'know' its own list of indications and contraindications, interactions and side effects.
Even today, many emergency departments keep all medications in a locked, computer-controlled pharmacy supply cart. In order to obtain medication, you must request it from the computer for a specific patient. Today the computer knows a lot about the patient's billing charges, but nothing about the clinical conditions for which the drug is being ordered. In the future, when the cyberspace representation of a medication is 'drag-n-dropped' on a particular patient, the medication object will examine its new environment (the patient data record) and will send system messages containing the correct dose for that patient, the necessary adjustments in any other medications for compatibility, the method of mixing, which infusate to piggyback it into, and so forth. If a medication knows it shouldn't be given in pregnancy, it won't allow itself to be given to a pregnant patient unless you explicitly force the issue with an override. If the medication can't tell from its environment whether the patient might be pregnant, then if there is already urine in the lab, it will order and check a pregnancy test. If there is no urine available, it will ask for some.
Medications will not be the only 'objects' with 'methods' in the emergency department of the future. Lab tests will have 'methods', and will order companion tests if needed for meaningful interpretation. An order for a CSF glucose, for example, might look for a recent serum glucose, and if none was found, might proceed to order one. It is entirely possible that an ECG might read itself, compare itself to its younger self, order a set of cardiac enzymes, request the pharmacy cart to release a thrombolytic agent, and put in a page for the CCU resident, notifying the clinician of these actions as they are performed.
The advent of object orientation and of ubiquitous connectivity among all components of our local environment will bring changes that may leave very few aspects of modern medicine untouched.
So far as decision assistance is concerned, it seems apparent that our practice will be changed in one more small, but still important way. Because it is reliable and can be incrementally improved until we are satisfied, it is inevitable that the computer will become a 'trusted advisor'. The concept of 'trusted advisor' is a very important one. As an example, if a medical student tells you that a patient's vital signs are stable, you take the chart and look at the vital signs yourself. If a nurse tells you that the vital signs are stable, you enquire "what arethe most recent vital signs?" If your trusted colleague tells you "the patient's vital signs have been normal," you accept that information. The reason you accept one but not the others is not that you suspect the student of lying, or the nurse of failing to notice an abnormal vital sign. The reason one statement is more credible is because you trust your colleague to have applied the same perceptual filters to the data that you yourself would apply. The vital signs are normal according to criteria that you yourself would apply, taking into account the clinical setting. The statement "the vital signs are normal" becomes a "trusted observation" from a 'trusted advisor'.
The computer can become a 'trusted advisor' because it can be relied upon to apply criteria that you yourself would accept, and can inform you of exceptions that you would want to know about. When this comes to pass, physicians can be freed from much of the busywork of modern medicine. If you can accept it when the computer says "all the labs are back, and they are all normal, " you will be free from the need to examine each and every lab result from every patient. If you can trust the computer not to let an intern discharge a patient whose presenting complaints and vital signs are inconsistent with the discharge diagnosis, you will be free from the need to hover over the intern and all of the intern's patients.