What's Next?
Preparing for the Future

Future of MT, 1996 / Speech Recognition / Back to Free Stuff

12/97-6/98, Brainstorming group, from Mary:

MT Daily hosted an experimental password-protected message board for about 20 experienced thoughtful online MTs and a few others for the purpose of discussing issues related to MT and generating helpful ideas. We formed seven goals of what we want to do for MTs:

1. Establishing standards of performance.
2. Improving quality of work.
3. Increasing efficiency and pay.
4. Improving business communication.
5. Raising public awareness.
6. Improving training and internships.
7. Encouraging balanced living.

2/97, from Mary: from E-mail to new MT:
Don't worry yourself about the future, you are in a good position now and you are in the "wagon train" of MTs of the future. You have a skill, like knowing Russian, that can be used in a variety of ways in the medical field, but most likely we will be typing narrative for doctors for a long time yet, and I don't think voice will ever become the preferred method, only supplemental. Just remember this, doctors don't have time for it, and hospitals don't have money for all new systems very often, only about every 10 years.

I used Dragon for 6 months and I wrote about it on MT Daily I, you can read it if you want to see how tough it was. I got it to go pretty fast but not as fast as typing with Smartype. It was so inaccurate that the constant correcting with speech hindred the flow of dictation. I had to say, "Scratch that," and then repeat the word, sometimes a few times! If you want to buy my Dragon and try it, you are welcome to. It's probably worth about $400, to a handicapped person with no hands, that is.

Now, the thing to watch is the automatic voice-to-text machines. The doctor puts in the tape and out comes the text, with lots of errors, and that's where we come in. They are only paying MTs $12 an hour to do the keyboard editing of this experimental program in Massachusetts.

So, let's say that transcription starts going that direction. It will probably take at least 10 years for the majority of dictation to go that route, if it does. Look how long it has taken hospitals and clinics to just get their medical records on computers and start dictating instead of handwriting! It's taken the last 10 years for MTs to go from typewriters to word processors.

So it's true that we have to stay flexible, and it's true we need to work together. But I doubt AAMT is going to be of much help, they are not a network of people who can help each other find jobs, like we are online.

Have you joined the student listserv yet? That would be quicker and more efficient than reading You would get about 20 email notes a day from MTs and students, and people feel safe to ask simple questions on there. See this page to subscribe:

I think you've got a great future, and soon you'll be ready to teach other MTs! :) Just keep watching your speed increase and enjoy your kids while they are little.

11/14/96, from Mary:
I am beginning to believe that as the technology becomes more usable and affordable, the standards of quality may change to fit it, much as our expectations changed to perfect spelling when spellcheckers arrived, except that this adjustment will be in the opposite direction, back to allowing for typos (speakos?) for the sake of the overall production efficiency and savings. Thus, there will be two kinds of transcription perhaps, the fast-text (fast-food) variety and the custom high-quality variety. While I shudder at the loss of quality and the danger of major errors of medical meaning, I do appreciate efficiency, and I am sure that while humans make errors, they also have an amazing ability to spot them and understand them. Medical records are so repetitious that any discrepancy is likely to have a corrected citation somewhere else.

This goes to the tension between accuracy and speed, one that we cannot resolve or escape, but only hold in some kind of balance, I believe. MTs probably cover a broad spectrum on this tension, and on different days different balances, depending upon the deadlines and pressures. Perhaps there are at least two kinds of effective MTs, probably more: Practical and Perfectionist. We are going to need to have the competitive edge on both scores to keep MT from going to the machines.

11/03/96, Re Voice-to-Text Machines with MTs as editors:
It will always be humans who dictate or tell the computer what to do. We all know that humans vary in their capacity to remain alert and do what is required correctly and successfully. From time to time because of endless variables which play a role in human capacity/attention, there are deviations from what is required in any given task, including creating patient documentation. It is when the human factor sets in and results in "slips of the tongue" or misstatements of one sort or another that problems ensue.

By the 20th dictation in the ER the other day, the doctor was slurring her speech, barely getting the words out, and if there weren't a human at the other end of the line deciphering what she was desperately trying to say to get her job done already, one can only imagine what the computer printout would look like if the Voice-to-Text process had been involved.

Worse yet is when the doctor is fully alert, functioning perfectly well and up to par, and unwittingly makes a misstatement about a medication dosage or a lab value. From time to time, we have all heard a doctor dictate, for example, as I did just yesterday, Zestril 150 mg when we know full well the doctor meant Zantac 150 mg, given the context of the patient's dictation. We are also familiar with the fact that Zestril is NEVER prescribed in such a high dosage. But what if Voice-to-Text gets ahold of this, and Zestril 150 mg gets into the patient's chart and then that dosage is administered by a well-meaning, unsuspecting, physician's assistant or whomever, and follows the chart instructions? One would imagine that the implications of administering the incorrect dosage of ANY medication may have potentially harmful consequences to patients.

So, it is the human factor that is invoved --both the doctor's humanness with the potential for slipups or misstatements and the transcriptionist's humanness with the ability to pay attention and remain alert for any discrepancy which needs to be pointed out, questioned, or otherwise flagged for attention/action by the physician that will always remain out of the picture when people introduce the high technology involved in voice-to-text.

Well intentioned marketers of Voice-to-Text technology have never addressed these issues in their presentations, and never will because thay can't. What is unfortunate is that many people in the health care industry will proceed with advancing this technology without taking into account the safety factors and screening necessary to avoid such problems.

We need only one case of dire consequence in such a situation to have the patient suffering from receiving the incorrect medication dosage, or whatever, to have that patient point out to us all that a potentially life-threatening mishap caused by the computer revolution is just not worth it. That individual, I am convinced, would prefer to keep the human element in there at all costs. The voice technology people, however, will not address this issue and will proceed with their income-producing activities and will adress huge markets for implementing their wares and high-level, perhaps well-intentioned administrator will also go ahead and bring this technology to the forefront. The sad part is that unsuspecting patients will have no awareness of the potential, even if unlikely, danger that will lurk in their medical record documentation to their detriment.

11/96, Eric Fishman, M.D., 20th Century Eloquence
I am a practicing Orthopedic Surgeon who employs 4 full time transcriptionists for my group practice in my office, who uses voice recognition software for some of my dictations, and who owns a company which re-sells all of the major voice recognition products. I'm sure that voice recognition is not a subject which is well liked in this newsgroup. However, it is my opinion that for a few adventurous transcriptionists; those who are willing to take on some additional training and/or expense, that the rewards in voice recognition will be well worth the effort. From a transcriptionist's perspective I would state that there are 2 classes of voice recognition programs: those which require the person dictating to make corrections, and those which allow a different person to do the corrections.

Up to this point, all of the programs available were discrete speech, and required the person dictating to make corrections. This is because there was no audio file available with the printed word. These products left the transcriptionist completely out of the loop. Happily, for both you and me, this is changing. There are now two companies producing voice recognition software that will, in my humble opinion, offer the enlightened transcriptionist an opportunity to make substantially more per hour than previously. IBM and Philips Dictation Systems are currently offering continuous speech recognition programs which will not significantly impact the way the physician dictates. The physician will be able to dictate at their usual rate of speech, without the pauses required in discrete speech products such as Dragon Dictate, Kurzweil Voice or IBM VoiceType. The good part, from your perspective, is that there will be a linked audio and text file saved in the computer. The computer text file will be between 95% and 98% accurate, after a little training. Therefore the physician has two choices: Make the corrections themselves (something that most physicians probably will not be particularly interested in doing), or send the linked voice and text file to a transcriptionist. With this linked file, the transcriptionist should be able to proof and make corrections, at up to 200 words per minute. This is clearly substantially quicker than most transcriptionists are currently outputting. Thus, for the brave, if you are interested in potentially almost doubling your income, I have a suggestion: Make an arrangement in which you would provide this hardware / software combination to your clients, and offer them a slight discount if they use it. Thus, even though most of these systems start at around $10,000 - $20,000, even for a simple stand-alone system, I would anticipate a very very rapid return on your investment, and a very competitive position in your local marketplace. We are currently working on a system that will allow transcription companies to download the files from distant sites, make the corrections and then upload them back to the person dictating for printing and signature. Even without this system, however, there are opportunities available. Obviously, as a provider of this software I have a biased opinion. And I will admit, that at some point in time, probably many many years in the future, it is likely that this type of software will displace some transcriptionists. Nonetheless, I believe that at the present time there is an excellent opportunity awaiting those who wish to follow it. If anyone wishes to review some of the information currently available, it can be found at our website. Currently there are continuous-speech programs ONLY for Radiology and Psychiatry; however, I expect this list to expand significantly over the next 3-6 months.

From Mitzi Ponce:
My personal opinion is that we are not going to see a major implementation of voice recognition any time in the near future, at least not on the clinic level. I can briefly list the reasons I belive this is so, and I would be very interested in your feedback.

Firstly, even if the cost is absorbed by the transcription company, clinic, or hospital, I just don't believe most physicians will be interested, at least not the current batch(!). My employer, moved from tapes to a digital dictation system in the past few years. Our experience has been that many physicians are extremely resistant to giving up their hand-held recorders. Many more are suspicious of technology and reluctant to believe that "the reports are going through," even with the ability to use play back to ascertain a report was received. I'm approaching 40 myself and I feel truly comfortable with computer technology, but many of our older physicians are uncomfortable with the technology (no offense meant, and I truly do know that many older physicians embrace technological changes, just, in my opinion, not the majority). Many of our physicians cite convenience as the factor that prevents them from using phone-in systems. Certainly, this is not to say that a great number of our clients have embraced the new technology, because certainly that is so. But, I believe an equal number are resistant. Perhaps as more young physicians, who grew up with computers in the home, come on board, we will note less resistance.

Secondly, we have physicians dictating from everywhere, home, the office, the operating suites, and on and on. I can only see these physicians going to voice recognition if it incorporates the ability to phone in.

Lastly, I just can't imagine physicians being willing, en masse, to adhere to a "forms" style of dictation. As you are well aware, most hospital, clinic, and private physician reports are presented in fairly rigid formats. Patient demographics, major subject headings, operative procedure names, operative indications, dates of services, and so forth, all must be entered at some point. Since I have not personally used either of the two systems you mention, I'm not sure how a physician designates such pieces of information or indicates that he/she is moving from, for instance, the subjective section of a SOAP note to the objective section. My own experience, based on very rigidly formatted inpatient reports I transcribe for a major pediatric hospital, is that physicians, often Residents on-shift for 36 hours or more and zinging through their reports at one or two o'clock in the morning, tend to either skip vital pieces of information altogether or get the whole thing out of order! I have to tell you that my program hates that!

I have one Resident, God love her, who can put 40-50 lines of 65 character/line dictation into a two-minute report. She only stops about every 15 lines or so to take a huge breath and then zings right into her next uninterrupted 15 or so lines of dictation. I've counted her words per minute in comparison to her colleagues in the same specialty group and can tell you that she gets an unbelievable top rate of 300 wpm versus about 120 wpm for her colleagues! She gives no real clues as to what part of her report she is dictating and I can only say that it's a darned good thing she has a thinking human being on the other end of the line! Furthermore, this particular dictator has a distinctively telegraphic style of dictation that employs a great number of abbreviations and tosses English niceties to the wind, say articles, pronouns, etc. I harbor no malice against this dictator (okay, maybe a little), but I cannot imagine a voice recognition program savvy enough to decipher any of what she says. If I lose my concentration, even for a second, everything she says becomes gibberish, as if I were listening to a foreign language.

I can well imagine how much easier my job would be if my task were only to edit and "fancy up" voice-recognized dictation. I would be faster, probably more accurate, and certainly less tired at the end of a day. I just wonder, though, how would you teach those doctors to do it right!

The rest of this webpage was distributed at the AAMT Convention Panel on the Future of MT, led by Joe Weber, August 2, 1996. Contributors are from the ranks of online MTs.
7/19/96, from Mary Morken (webmonitor): The future for MTs is flexibility (ready for anything), fulfillment (of hopes), forms (less boring work and shorthand typing), fairness (negotiating), freedom (choices), fighting (for excellence), fleeing (the sweatshops), forgiving (to keep on helping each other) and facilitating (to benefit from networking).

I hear of MTs filling in forms that are replacing dictation, but the inserts are still being dictated. I think some dictation won't be able to convert to forms, like psychiatry, complex surgeries and consultations. I think more men will join our ranks, and more couples as well. I think the term "medical transcription" as well as the work of MT will survive the changes in medical records processing. I think AAMT will either change radically or wither away. I think established MT leaders will continue to be great assets, but new MT leaders at large will be of great help in this field if they stick close to the interests of most MTs, both the new MTs and the hard-working front-line MTs. I think MTs will improve their use of computer networking, and the "grizzly bears" online will become extinct or civil. I think the Internet will help MTs to hold MT services accountable, improve the performance and expertise of MTs, and facilitate the MT market. I think future word books will come with diskettes, and voice-file compression will shrink transfer time for dictation. I think programs will become more compatible and standardized. I think new jobs will be created for new MTs so they can join the ranks of MT producers; and if this does not happen, the experienced high-production MTs will collapse from exhaustion or grow adhesions to their keyboards!

There are some pretty scary unknown factors that will keep us on our toes, like doctors learning to use computers and voice recognition and the world learning to transcribe. I think American MTs can still lead the world in this market/service in the face of these trends, and that we have such an advantage that we can afford to be generous. Here's the basis of my optimism today: Computers, the law, world health needs and the Internet.

1. Computerized records are going to be required more and more (by insurance, Medicaid, HMOs, JCAHO, for faster billing and transfer of records, etc.), so doctors who used to write by hand will have to dictate.

2. Legal pressures for malpractice protection and compliance with regulations in documention are going to keep growing, at least in America, and detailed records, narrative records, will still be required, especially for psychiatry and complex surgeries.

3. The doctors of the world will want to dictate rather than write by hand as soon as they have the computers and the staff, and eventually they will be able to.

4. Most important and within our control, American MTs are becoming an even more valuable resource by the research, technology and networking tools available to them online, and they can organize for more efficient training, production and job placement.

On an individual level, the future of every MT will depend upon continuing to learn and network to be well prepared for the unknown future, whether it is much better or much worse than the present.

From what I hear, there are maybe 150 MT trainees overseas now, in India and Philippines only, doing American dictation. They say it is saving money because they say 60% of their expense is MT salaries, so if they can save money there they can afford the equipment and the American personnel to teach, supervise and proofread the work (100% of it!). One of the online MTs is in Philippines teaching right now, and is going to write some about her experiences. These are very recent experiments, so the long-term viability of this cannot yet be seen.

This will definitely keep us on our toes and needing each other so we can give the finest service possible, but keep in mind there is still a shortage of MTs in America, and the field is growing as everything must be typed more and more for JCAHO and Medicaid requirements, and nurses and therapists all want in on the ease of dictating instead of writing in the charts.

Then there is voice recognition (soon to have continuous speech capability but no one knows how fast it will be able to go yet) and more objective forms for reporting, where the doctors answer questions which can be done quickly, rather than dictating narration. The big advantage of this is that the doctor is prompted for all necessary information so it is more complete, and then statistics can be compiled more easily. Emergency rooms, radiology and pathology are the first departments going that direction, so be sure to train broadly! We gotta be flexible! On the other hand, it takes time for people to switch to new systems, so I think it will be years before transcriptionists are an endangered species, if ever.

I hear of phlebotomists having to find new jobs because noninvasive ways of testing blood are coming in the next 10 years. I talked to a microbiologist yesterday who is retraining for MT because so much is automated in the lab people are not being replaced when they quit. It takes a long time for new technologies to be purchased by us all, so I agree that MT as we now know it is going to be around for a long time in many places, and more of the health personnel will begin to dictate. Changing to editors will be a pay cut for those of us who are independent. I hear of pay like $12 an hour for proofreading. We got it good!

There are definitely changes afoot, cutbacks and such, computerizing medical records the main one that affects us. While some dictation may be done directly to computers by doctors, dictation is still on the rise because nurses and therapist and others are getting in on it, so the market is projected to still have a 15% increase in transcription jobs until year 2000. We will have to watch voice recognition development--but so far it is not fast or accurate enough for doctors to learn it. I think transcription will be around for a long time, even if some doctors go to voice and even with cut backs. Some MT work is going overseas, and that is another consideration--cheaper labor there. It does mean you need to have every advantage for doing excellent transcription so you can compete, and then to be ready to be flexible to use these skills in whatever new ways are needed if transcription were to go out...

No one knows what technology will be like 10 years from now, or what world trade will be like. I think even 3-5 years will see some pretty big changes in technology. Let's make a list of equipment for MTs to face the future with confidence:

1. Settled personal beliefs about ultimate questions of life.
2. Healthy, stable lifestyle and structures of life.
3. Participation in networking with peers who face the same future.
4. Well-developed skills in medical terminology, language, computer and relating to people.
5. Willingness to use skills in new ways.
6. Willingness to use old skills or learn new skills as they are needed.

From Judy in Portland:
7. Sense of humor.
8. Develop attitude that change equals opportunity, even when that change was not by your choice.
9. Save money for lean times.
10. Downsize expenditures BEFORE your income is downsized.
11. Develop lots of interests besides work - they may become your occupation of the future.
12. Don't take your client's or boss' BUSINESS decisions PERSONALLY. I have to keep telling myself this one. The clinic I have been working for, first inhouse and then as an IC, since 1989 merged into a big corporation. After telling us they would keep us because they liked us so well, they're now talking about going to a big service so that all the clinics have the same transcription company. It's hard not to feel rejected even though intellectually I know it's not my fault.

From Melinda Taylor
13. Personal service: Watch for ways to "go the extra mile" and personalize service for clients. Rather than demanding that doctors learn to dictate "the right way" and adapt to our methods, be looking for ways to make it easier for them to get through what is, to them, a low priority task - dictating patient records (and yes, I know the records are extremely important; I'm just talking from the doctor's point of view, as compared to some of their other responsibilities). Just as we, as patients, think more (not less) of the physician who makes the extra effort for us - be it through answering a late night phone call personally or making a house call - so has it been my experience that clients treat us well when we go out of our way for them. In doing this, I've received nothing but the highest respect and appreciation in return. The few extra minutes I spend now and then have been well worth the effort in terms of client and/or service loyalty, and, of course, financial security.

6/96, from Scott D. Silverman, 76226,, SkyBail Computing:
Doctors doing their own medical records are picking from lists, checking boxes, etc. While voice recognition is slow these days (70-90 wpm), because they build most of the chart without dictating, they can get done in about the same amount of time as just dictating to tape. Medical records/administrators like it because (1) chart is available immediately, (2) as opposed to a typed word processing document, this chart is actually a data file tracking 300+ separate fields of information, (3) this data file model for chart building allows outcome analysis, statisitical studies, etc. It also allows billing data to be generated immediately as physician clicks off what tests/procedures were done, (4) reduced/eliminated transcription costs.

One way or another, typing will lose money value as a skill. By moving now to add to your skill list you can get yourself ahead of the pack and I would guess command a higher value for your time within a year or two as more industries/operations look for experts on how to apply voice to what was once done with typing.

6/23/96: From Peg Brundige:
Yesterday at the NET meeting in Anaheim there was an interesting bit of discussion on expanding our field to include coding and medical billling. With all the changes that are taking place now, they were saying that anything that you could include in your service might help to get or keep a client. It was discussed that for a coder or biller to make the transition to our field might be quite difficult, but for us to transition to include those fields in what we do probably would not be that difficult. I think this might especially help the "small to middle-sized" services that are having trouble competing with some of the larger services that seem to be taking over the accounts after some of the mergers.
From Arleen:
I read your diary on DragonDictate with great interest. I think the amount of time it took for you to "train" it and the extra proofreading it required (dropping endings, etc.) exemplifies how far in the future voice recognition replacing MTs really is. Friends of mine in IBM have told me the system they're working on is accurate only if one speaks very slowly. I can't imagine hospitals getting MDs to take the time to "train" these programs or slow down enough. (They can't even get them to punch in patient numbers half the time!) Most of my doctors speak in speeds of fast, very fast to speed of light.
7/96, From Mary:
I've been typing the daily rounds of some attending physicians in a very large eastern hospital and am wondering if other hospitals are having this kind of chart entry dictated and transcribed. Is anyone else doing daily rounds notes?
7/96, from Mitzi: I am involved in a pilot project of daily hospital rounds. Prompted by concerns over accurate billing code levels and legible chart notes, not only for in-house use but also for review by outside agencies, this large midwestern pediatric hospital initiated a pilot program whereby daily chart notes are transcribed by an outside agency. Currently the project involves two services, Hematology & Oncology and General Pediatrics, and involves three types of notes: Initial Care Notes, Subsequent Care Notes, and Procedure Notes (the first two are the familiar Admitting Note and Progress Note respectively). Following patient contact, physicians call a dictation system and record their notes. Typically, the physician provides all statistical information on the patient, including date of birth and medical record number; however, our macro-driven method of entering notes builds a concurrent DOS delimited text file from which we import data into a patient database for our own use. Thus, the patient statistics are always available for subsequent dictations. (Please note that this DOS file is also provided to the hospital for their use as they move forward toward computerized charts.)

Most often the reports follow the S-O-A-P note format; but in the case of new admissions, the notes also include a Chief Complaint and History of the Present Illness, as well as past personal and family medical histories. In addition, all notes are quite likely to include a section of laboratory results and/or direction to residents from the attending who dictates the note. Finally, all notes include check boxes, filled-in by the macro, describing the level of billing for each phase of the in-hospital visit. Typically, while termed 24-hour turnaround, the functional turnaround time is about 12 hours and physicians have the option of designating stat reports for immediate turnaround in emergent situations. All reports are returned via e-mail. The hospital uses its own medical records staff to print and distribute the chart notes.

If the pilot program is successful, it is hoped that all inpatient services will begin to use the system, including physicians with privileges rather than only staff physicians as is now the case. I am working closely with the physician heading up the pilot program and we have developed a visually pleasing chart note that generally is one page in length. The product provided to the hospital includes the chart note and the DOS file; we retain rights over the macro-driven facilities and other intellectual properties used to deliver the product. This process could easily be used in other institutions.