Copyright  2003 Civic Research Institute, Inc. This article originally appeared in Medical Software Reviews, Vol. 12, No. 4, July/August and is reprinted here with permission of the publisher. All rights reserved. To learn more about Medical Software Reviews, please contact CRI customer service at 609-683-4450 or visit the  Medical Software Reviews page on the CRI website: www.civicresearchinstitute.com.  

Author's  Note: This article is displayed here by special and unusual permission of the publisher.  Please contact them at the sites above for any hard copy reprints or subscriptions.  Please respect the copyright and DO NOT print out or copy this article from the Internet.  Thanks- CJZ

© 2003 CRI Health Care Publications

Medical Software Reviews July/August 2003

CASE IN POINT

INCORPORATING PATIENT-ENTERED DATA IN YOUR EMR, USING INSTANT MEDICAL HISTORY

by Charles J. Zelnick, MD

By: Charles J. Zelnick, M.D.

RATINGS: (1 = poor, 5 = superior)

• Performs stated functions: 5
• Content: 5
• Documentation: 4
• Error handling: 4
• Ease of use: 5
• Value for price: 5

Contact: 

Primetime Medical Software

4840 Forest Drive, #349

Columbia, SC 29206

803.796.7980 voice

928.962.0034 fax

http://www.medicalhistory.com

Email: info@medicalhistory.com

Pentium III 1.0GHz, 133Mhz Bus, 256 MB Random Access Memory, 4 MB Video Memory, 20 GB Hard Drive, CD-ROM Drive, minimum 16X

Windows XP/2000/NT is recommended for local installation. Other operating systems can access Instant Medical History through the web version. Linux and Unix versions must be installed on a server; can utilize a Java version of vendor’s toolkit to develop some forms internally.

HL7 Version 2.3

XML

Approximate Costs:

A subscription to Instant Medical History for office use is $50 monthly. Subscription for both office and home use (patients can access program through the Internet) is $100 monthly. Instant Medical History is frequently bundled with an EMR system.

United Kingdom English and Dutch are available. Version with Spanish-language questions is being beta-tested (output will be in English).

Interfaces: The vendor currently offers interfaces with about 15 well-known EMR systems, including HealthMatics (A4 Health Systems), SOAPware (DOCs), Praxis (Infor-Med); other interfaces are in development.

I first heard about Instant Medical History (IMH) in 1996. I was attending a software beta-testers retreat in the Seattle area, and our ride missed the ferry across the Sound by one minute. So I passed an hour with Allen Wenner, MD, a family physician from South Carolina, eating clam chowder and talking about IMH. He told me that in the early 1990s he realized that in taking patient histories, he was asking the same questions over and over , and he figured he could get a computer to do that job for him. He estimated the project would take about six months of programming effort. Fifteen years later, his six-month project continues—extended and in constant development-in contrast to many other software projects which have come and gone.

At first I was very skeptical. How could a computer replace my excellent history-taking skills, which had been formed by a rigorous education, then honed by 10 years of small-town family practice and 5 years of academic teaching? We were taught in medical school to admire the chairman of the internal medicine department for his ability to ferret out the single important question that would unravel a tangled medical history and make the diagnosis obvious. No computer could duplicate that, I thought. However, curiosity made me agree to take a look at Allen Wenner’s program

Honesty compels me to share that this episode was the beginning of a personal friendship with Dr. Wenner, and a relationship with his company, Primetime Medical Software, as a beta tester. I began by “playing” with IMH in my clinic, when it was still a product that ran on the old DOS operating system. I asked some of my patients to try it out, and immediately began to notice that it changed my practice. Over the years I have made suggestions and provided question sets (such as the obstetric question set) to the Primetime staff, and in the last few years, I have collaborated with the company to integrate my use of Instant Medical History with our EMR, Medicalogic’s Logician. In a previous issue of MSR, I touched on this process (Medical Software Reviews, Vol. 11, No. 3, May/June 2002, pp. 7-10; 15). In this case study, I will provide more details about the theory and practice of using patient-entered data with an EMR, by recounting my experiences in using and integrating these two pieces of software.

Practice setting. The Cedar Rapids Medical Education Foundation, our family practice residency, is sponsored by two competing community hospitals in the midsize city of Cedar Rapids, Iowa. It consists of 20 residents, 6 physician faculty, a PharmD, a PhD psychologist, and 17 support staff. Our model clinic, the Family Practice Center (FPC), is fully computerized and sees an annual volume of 14,000 to 17,000 patients

How does Instant Medical History work? Essentially, you can think of it as a “compulsive third-year medical student in a box.” Third-year students focus on learning to take histories. Typically they are assigned a patient to interview in the hospital, and so will arrive at the bedside with a clipboard and a long list of questions on a form. Given a chief complaint, they pull out the list of questions for the relevant organ system, and they make sure all are answered. Positive answers may elicit further questions. Once the present illness is covered, the student moves on to ask about past medical history, family and social history, allergies, and finally performs a complete review of systems.

In some ways, ironically, our best histories are performed when we are medical students, because we are given this hour or so to talk to a patient and make sure we understand every aspect of his or her illness. As we progress in our careers, however, the pressures of time and concerns about reimbursement inevitably require us to abbreviate our histories. Dr. Wenner told me he was finally inspired to start work on IMH when he sent a patient to a teaching hospital for a confusing medical problem, and the patient returned and told him about the “great young doctor with the short white coat, who listened to him for an hour and made the correct diagnosis.”

Instant Medical History works like the good student: by simply asking questions—and a lot of them. Any office staffer can initiate the process. The staffer brings the patient to the computer, calls up the first data-entry screen, and gives some basic instructions about the program and how to enter data. tThen the patient is left alone (except for the occasional question) to complete the process, which usually takes between 5 to 20 minutes.

The patient starts, of course, by entering name, gender, birth date, and a chief complaint. The software proceeds very systematically. For example, if the patient enters “headaches” as the chief complaint, the first question might be “Is this your first time at this office?”–followed by; “Is this the first time you’ve seen this doctor for this problem?” Subsequent screens ask “Do you have a headache now?” and “How long have you had this headache?” and the like The questions appear in large, bold fonts, about 16 point size (easily readable by most patients with visual problems) and the patient responds by pressing a keyboard shortcut (e.g. “Y” for YES, “N” for NO) or using a mouse to click on the response buttons.

Only one question is presented at a time, on each screen. That keeps the display visually simple--helping the patient to focus on each query, in a way that paper forms cannot; But the sequencing is also crucial to the operation of the software, because complex branching operations are occurring in the background. For example, if the patient says “Yes” to “Are there any things that make your headache worse?”, the program will branch to this query: “Do foods make your headache worse?”A Yes to that triggers further questions, concerning which foods are involved, and the various branches let the program go into great detail (“chocolate?”, “sour cream or yogurt?”, “tea, coffee, colas?”, “bananas?”, “citrus fruits?”, “onions?”, “lima beans, navy beans, pea pods?”, “pickled or preserved foods?”, “nuts, peanut butter, seeds?”, “vinegar?”, “pizza?”, “pork?”, “herring?”, “avocados?”, “alcohol?”, “aged cheeses?”, “hot fresh breads, donuts, or cakes?”). A No answer will abort or terminate a particular branch, of course. Thus the patient’s simple Yes or No answers allow the program to draw a detailed road map of information, systematically and thoroughly. Similar branching is used to gather details about the patient’s family history (“Who in your family had migraines?”), social history (“How much alcohol do you drink?”), and past medical history (“Have you ever had a head injury?”).

In this way, an entire history is built, one question at a time. The question sets have been developed by experts in each area, or based on published literature, and then field-tested and improved over many years. Patients recognize the thoroughness, of course, and appreciate it. At the end of the questions, the software instructs the patient to call the nurse. The program reorganizes the answers into the familiar narrative history format, with positive answers listed in bold type, followed by light-grayed negative responses. This output can be customized in a variety of ways, including a detailed vs. a bulleted format, according to physician preference. If configured to do so, this instant history will automatically appear on a printer for immediate use in the office. Behind the scenes, the software saves the data in three output files: an “.imh” file (the program’s proprietary format which can be re-imported and processed by the software itself), a “.doc” file, readable with any word processor, and an “emr” formatted file, which can be imported into various EMR programs. The format of that “emr” export file can be set in the “Options” to be compatible with one of several commercial EMRs. Installation is very easy, and periodic updates may be downloaded from the company’s web site. Moreover, a web version is available, allowing the patient to complete the form at home, prior to a visit (I have no personal experience with this option).

The software can be further configured for the individual office. Options may be checked to allow the following: Patients may skip questions, or trace back to change answers; they may enter further complaints for more history-taking; they may enter their own chief complaint in free text; or may be required only to give short answers to the built-in or custom-made series of question sets. The software may be “locked down” with passwords to keep patients (and their children) from modifying the display, playing with the computer, and, of course, from accessing networks or previous patients’ histories. You may set the locations of output file storage, and most importantly, you may set the detail level for various organ systems.

This last option is not intuitive, but very useful. For example, there are question sets in the software developed by neurologists which are quite detailed and specific about dizziness and vertigo. As a family practitioner, I might want to turn down the “sensitivity” of this set so that every headache patient does not have to answer 40 questions about some occasional dizziness accompanying headaches. But the neurologist can turn up the sensitivity to obtain a very complete, expert dizziness history.

The software’s specificity also includes many clinical scales. These include both psychological and medical scales standard in the profession. For a patient with headache, for example, the software may branch off and perform a “Migraine Disability Assessment Scale.” Because many headaches are associated with psychological symptoms, the software will do brief mental health screening questions, too, and if these are positive, it might perform a Zung Depression Scale, Hamilton Anxiety Scale, and a Holmes-Rahe Stress Inventory.
The software also has some built-in electronic shrewdness: It responds to somatisizing patients (those who answer Yes to every question about symptoms) by ceasing to ask somatic questions, and branching to the psychological scales. The software also has a “stop” button, allowing an office staffer to stop the history-collection process if the patient is taking too long; of course the system preserves the data collected to that point. In my experience, this happens when the patient really likes the process and keeps saying Yes each time the software asks “Do you want to enter another complaint?” at the end of a segment. (Stopping such patients has not caused incomplete histories. In fact, the data collected by that point has always been sufficient to cover the chief complaint and one or two additional concerns.)

The idea of using a computer to collect patient histories actually dates back to 1968. Schumann in 1975 showed that a computer adds 40 percent more information to the clinical interview, and that 22 percent of the time it led to improved patient-physician communication. (J Fam Pract. 1975;2:263-269.) In a recent overview article, “The Patient-Computer Interview, A Neglected Tool That Can Aid the Clinician” Dr. John Bachman reviewed all the published research on patient-entered data (Mayo Clin Proc. 2003;78:67-78). His article enumerates the few drawbacks—and the many benefits—of a computer-aided interview. Of course, a new technology that is truly revolutionary will require a change in habits, and thus can meet resistance.

Indeed, I found resistance close to home. What was most humbling, at first, was the ability of the software to surpass my own history-taking skills. As is shown by the above list of “foods that cause headache”, our brains are not constructed to remember such details well. And of course, I knew in my heart I had been omitting important questions—as shown by history items in letters I received from my consultants, and phone calls from pharmacies about allergies I’d missed. The computer is much better at obtaining a complete and exhaustive history. After using IMH for a short time, I became less satisfied with my unaided histories for such complex problems as dizziness, headache, chest pain, fatigue and tiredness, back pain, etc. On the flip side, the computer cannot discriminate concerning the significance of the response. A patient might indicate that “pizza” will “sometimes” cause “headaches,” but it takes a physician’s review of the overall history to recognize the classic symptom complex for migraine headache. Our minds are designed for pattern matching, and far surpass the computer in that task. To get closer to the ideal goal of a history with zero defects (errors), I need to combine the computer and my brain.

In addition to the computer outstripping me on details, I also had to face sharing control of the patient interview with the program. Sometimes the software even creates “negative feedback” by presenting information that challenges a doctor’s skills or produces an emotional response. One patient complained of “fatigue,” for example; but his IMH showed a sky-high depression score, suicidal feelings, and a gun in the home.. The physician of course saw the need to deal with facts that might not have been revealed without the computer, and indeed with a pattern that a doctor subconsciously might not have wished to know.So in this case, use of the computer interview may have saved a patient’s life.

Because of this natural resistance to change, I found that my physician colleagues, while expressing support for the idea of using IMH in theory, often conveniently “forgot “ to order the interview for their patients—or they simply told us they “didn’t want to think about which patients needed to do IMH.”. Therefore, I trained our office nurses, who were much better at getting the patients “on the machine” before their visits. Ultimately, we put a computer right in the waiting room, and simply had our front desk staffers start the sessions. This was much more successful. Now we just present the history to doctors as part of patient intake. To keep things manageable, we have rolled out the implementation in steps. In my own practice, I use IMH on all new patients, all “physicals” (especially well-child and adolescent visits), all mental health visits, and many new complaints. The software also contains short return-visit question sets but we have not yet used them extensively.

But what of the responses of patients? Here, the news was better from the start. When I started using Instant Medical History, I was surprised by how readily patients accepted the technology. Only a few patients are unwilling or unable to use the software. Even many elderly patients seem comfortable with this, as just one more of the technological miracles they’ve seen in their lives. Often, family members are available to help the youngest, oldest or sickest patients with their history.

Beyond this, though, I have found important affirmative help from the system. Sometimes this comes in the form of unexpected information. By offering this computer interview, have learned about previously-unsuspected low literacy in some patients. But often, the help is more broadly significant. In our practice, the nurse often gives a printout to the patient to review before I enter the room. Patients are very interested in their data. They check it for accuracy, and they often comment on it as soon as we get started—especially he psychological scales. One patient, for example, had been to the ER for multiple tests and workups for chest pain and shortness of breath. As I walked in the door, she said, “Doctor, I’ve been worried about my heart--but look, I score over 400 on the stress scale here. Do you think this could all be caused by nerves and stress?” I responded that it might be, but that we would check everything out carefully. Once I finished her exam, she was very ready to discuss my diagnosis of panic attacks,. This type of interaction totally changes the dynamic of the office visit, and is a particular advantage when identifying problems such as alcoholism, drug abuse, domestic abuse and depression. The patient sees that his or her own data, in an “objective” computer, is suggesting the diagnosis--not the doctor. This is a valuable effect. It helps the patient view the doctor as a partner in helping to rule diagnoses in or out.

Conclusion.

The inference is clear. Combining patient-entered data with training and experience is synergistic, in producing a complete, meaningful history. There are many additional benefits. Because IMH produces a beautifully formatted history, any transcription costs can be immediately cut in half. Another advantage is that the documentation, being more complete, provides opportunities to code at an appropriately higher level for work done, and may help protect the physician in an audit situation.

I know that with IMH I can practice better medicine. With the software, I begin the clinical visit with most of the history already obtained, which frees me from the need to record a host of details (often irrelevant) in the patient’s presence. Studies show the average physician interrupts the patient after only 24 seconds. I walk in the door, put my laptop down, and say, “Thanks for entering your history. I’ve read through it, but I’d like to hear what this means to you in your own words.” I then let my patients talk without interruption for several minutes. This helps me overcome the natural tendency in an EMR-equipped practice for the computer to get between us. Patients love this, and I gain deeper insight into their hopes, fears, worries, and the meanings surrounding their illness. Indeed, this is the only situation I know of where a computer helps build trust and therapeutic alliance.

Our most successful workflow change involved the first obstetric intake visit, which we scheduled with 45 minutes of physician time, to allow for intensive data-collecting. We serve a disadvantaged population, however, and were experiencing a 40 percent no-show rate, resulting in much lost productivity. So we added a “registration pre-visit,” run by our office manager. She obtains demographics, has the patient watch a prenatal care video, and confirms vitamin intake At the end of the visit, the patient completes an IMH session. Result: The no-show rate dropped from 40 percent to 15 percent for the following physician visit. We believe this is because the patients, by entering their histories, become more active participants in their care. These results were replicated at our Title X OB clinic. That office, which uses paper and does not have an EMR, was already doing nurse intake visits with an identical no-show rate. However, adding the IMH session produced the same improvement in attendance at the first physician visit.

As I detailed in my previous MSR article, we installed our EMR, "Logician”, in 1999. As all EMR users do after “going live,” we experienced major changes in our office workflows. Our physicians began entering much more data into the electronic record, which caused a loss of productivity and a slowdown in patient flow.

As a trial, I began to electronically “cut and paste” IMH outputs into the History section of the EMR. This saved me considerable time, compared to my colleagues, in doing my records. I trained our nurses to cut and paste for us, and we built them a small program to reduce the process to a few clicks. This workflow required only minimal time for training and software development. The pasted narrative was also quite easy to edit for accuracy. But there was one big drawback.. The history was just pasted in as one large block of text; it was too laborious to snip it apart and paste each section (such as Family History) into the appropriate data entry boxes. So while I had a complete, detailed history, the information wasn’t structured within the EMR’s data organization, and thus was not optimally usable within the EMR system (i.e., for reports and queries).

Fortunately, the Logician EMR has a built in HL-7 import interface which may be used to import lab values and narrative information. We contracted with Primetime Medical Software’s engineers to build an interface that would export data in HL-7, and then worked with engineers from GE Medical Systems to be sure this was compatible. The IMH interface was built very quickly, but important design questions took over a year to complete and test.

The main design question we faced was the issue of entering unvalidated data into the EMR. For example, what if the patient says “Yes” to the questions about impotence, yet when the physician reviews the data, the patient says, “No doctor, I’ve got no problems in that department!”. A patient answering 30 to 100 questions will almost inevitably make an error. We certainly don’t want incorrect data about male sexual function entered in the EMR flowsheet for GU problems. We therefore developed a strategy of importing the data into “parallel” terms, which are then validated by the physician. For example, this patient might already have data in the Logician EMR system, and the data field involved might be “ROS-GU”—review of systems, genito-urinary. Within the IMH system, the patient’s complaint of impotence would then map to a term called “DEP ROS GU” (Data entered by the patient, review of systems, genitourinary). In the data display the doctor sees, this DEP ROS GU item shows up in a box next to the ”ROS GU” area. With a few clicks the doctor quickly and easily compares the patient-entered data with the previous EMR information, combining and correcting as necessary, to update today’s record. We made a decision early in the design process to import only positive findings, which keeps this work at a minimum. Pertinent negatives are preserved in the EMR as a separate imported narrative that is signed after review.

This flexible strategy is especially useful in dealing with past medical history, family, and social history in established patients, who often have important data from past visits already stored in the appropriate data areas. I often combine old data, new patient-entered data, and my own new information, just as in the example given, to rapidly and efficiently update a chart. A further benefit of this strategy is that it protects the physician from assertions that the patient’s questionnaire responses were simply copied into the chart and never reviewed with the patient for accuracy. By combining the “DEP” data into standard terms, the physician is signing off on its validity.

I have been testing the interface forms and workflows for the past three months in my office. The combination of structured patient-entered data and the EMR allows me to produce clear legible documents in one to two minutes, while simultaneously talking to the patient and corroborating the data. A preoperative physical is an excellent example of this workflow. The patients typically bring a hospital form that is supposed to be filled out by hand for same-day surgery. I have them do the “anesthesia” question set on the IMH system, import it into Logician, and quickly validate the history, while combining it with important existing data in the EMR. Once my exam is done, a few mouse clicks complete the physical documentation. Check- offs in the problem list and one line typed in the assessment area finish the process. I hit Print, and a complete document is soon ready for the patient to carry to the preoperative area. The process is much faster—and the result more legible—than I could ever accomplish by hand.

SUMMARY:

Instant Medical History is software that looks simple, but in fact is revolutionary. The expertise contained in its branching algorithms is like having hundreds of specialists collecting your histories for you.

Indeed, some such system is really a necessary complement to an EMR system. As noted above, converting to an EMR system will typically cause an initialdrop in productivity, due to the physician and staff having to enter more data into the new software. IMH solves much of this problem, since patients enter their own data. Moreover, IMH provides a higher quality history, better documentation, and improved doctor-patient communication. When added to a structured EMR, the speed and effectiveness of the clinical encounter improves. In effect, time is leveraged to improve patient education and compliance, or to see more patients.

Early in this article I noted my personal acquaintance with the designer of the IMH system, and my experience helping to improve it. But while I believe this particular product is a good one, the point made here is of more general importance. Software that takes patient histories effectively is valuable, and if it can interface with an EMR system, can be a huge help. As a user of such software with an EMR, I feel like a carpenter, who having used a hammer all his life, suddenly acquires the power and efficiency of an automatic nail gun. While I may still use the hammer for fine work or tight spots, routine use of the nail gun builds things faster, and lets me spend more time thinking about the design of the house. In fact, I look forward to the day when most patients will enter their histories over the Internet (an option which IMH already provides, as noted above)before their initial office visits. That will help us better ,match schedules with appropriate times and resources for patient needs. All in all, a program like IMH enhances the practice of medicine. It doesn’t “take over” the doctor’s job; instead, it helps the doctor—with patient as ally—to concentrate on doing that job more effectively. I’d never go back to the old way.

Charles Zelnick MD teaches and practices at the Cedar Rapids, IA, Family Practice Residency. He can be reached at czelnick@juno.com.