And with the passage of new federal legislation, there will be an additional 32 million newly insured patients by 2014, adding further pressure on providers. Meanwhile, EMR systems are being promoted by federal stimulus incentives, and considerable research shows that physician productivity is reduced significantly by EMRs.
EMRs reduce physician productivity because they force providers to use drop-down menus, or free-text typing — both of which are slower than dictating. A more recent approach, voice recognition software on a physician’s computer, solves one problem but creates another: Physicians are allowed to continue dictating, but then need to spend considerable time correcting the transcribed file, slowing them further.
My experience trying to adopt EMR
The challenge is to successfully adopt an EMR system that does not impede a physician’s natural workflow. A well-designed EMR starts with an efficient physician-computer usability model. Yet often this part of an EMR system is poorly designed, creating frustration and mistakes.
As a physician, my experience with EMRs began in the 1990s when I was management information systems director at our hospital. We tried to deploy an EMR in the critical care unit. We hired all the nurses and other people required to operate it. It was turned on at a cost of over $2 million. Within a week it was turned off. Why? Because it destroyed the physicians’ and nurses’ productivity. They simply revolted and would not use it.
That experience showed me that there is a lot more to EMRs than getting a good computer system. People have to use the system and their behavior must be taken into account.
Our clinic’s goals for an EMR
Since 1999, I have worked with The Surgical Clinic, a Nashville-based group. In 2004 we had 16 surgeons at five locations. And we chart conflict. We needed charts for clinical visits. We also needed charts for billing. But a paper chart can only be in one place; therefore, the physician’s need for the chart always trumped the billing office’s need. As a result, billings and income were delayed.
At the time, we had paper charts at all five office locations. Frequently, bills were delayed because the physical chart was not in the billing office. In addition, payors were requiring more documentation, adding to the stress. To cope, we hired a large medical records staff. They were constantly moving around paper charts. We decided to fix the chart-chasing problem by going from paper to a computer-based system.
After a review of the available electronic systems, we settled on Allscripts’ TouchChart product. We set up the electronic record to replicate the look and feel of our own paper charts.
Graduated rollout of the EMR to all five clinics
In our deployment, we initially worked with a few of our physicians. This helped us work out the bugs in the system. Many times, hospitals don’t have the luxury of turning on one unit at a time, but for us it was important to have certain doctors champion the EMR. To minimize the loss of productivity, we scanned the charts the day before they would be used in the clinic and eliminated the paper record. This turned out to be a stroke of accidental genius because physicians were thus forced to use the electronic charts.
Eventually, all the physicians began using the EMR as designed. At that time, we were using a commercial transcription service. The doctor would key in the patient’s social security number and then dictate over the telephone. The transcription would come back to us via fax. It would then have to be electronically filed with the chart by a staff member.
Working this way, we experienced no increase in physician productivity, but we could send out bills quicker. We saw a 30 percent decrease in the time-to-billing. We also solved the chart access problem. It was easy for anyone to access the chart in any clinic or billing office. Physicians could now log into the chart from home. But we inadvertently converted our paper chasers and chart filers into electronic document chasers and filers and we saw no staffing efficiencies.
New problem arises from first attempt at EMR
We solved one problem and created another: mistaken identity. We had records that were incorrectly filed. We found that there were three places the electronic identification could be entered in error:
- The physician enters the chart number incorrectly.
- The transcriptionist mistypes the identification number.
- The document filer, viewing on a small screen, routes the document incorrectly.
Our medical records staff was now chasing misfiled documents and we did not decrease the number of people required to manage files. As a result, we began to explore how we could increase productivity while solving the mistaken identity problem.
Looking for a better way to manage records without people
Our epiphany came when we realized we needed to eliminate as much human interaction with the entire process as possible. You can’t apply a computer to a bad process and expect a good outcome. We refined our goals to enable our EMR system to automatically file records without human interaction, and to tag documents with an identifier that allowed the file to be appropriately placed in the correct section of the chart.
We began to look for something to make the process smoother and less error-prone. We selected Entrada to improve our processes and facilitate our use of the EMR. Entrada replaces medical transcription with a web-based technology platform that lays a foundation for an EMR. Through an HL7 interface to our GE Centricity scheduler and billing system, Entrada generates a “job list” for each physician, based on that physician’s workflow preference. The doctor then dictates relevant portions of the patient encounter, which is already stamped with the appropriate demographic identifiers. That way, the doctor does not have to enter the identification — it can’t be incorrectly entered. The dictated voice files go straight into a voice recognition engine; they are digitally captured and ready for correction by an editor. The final document is automatically routed to the EMR into the appropriate location within the file.
The advantage of the system is that it matches the physician’s natural workflow. A physician can easily dictate, which remains the preference for most doctors and certainly for most specialists. Special macros can be designed based on each individual doctor’s preferences. We have found the record to be very accurate.
Costs reduced by $250,000
The changes we made reduced our transcription cost by 30 percent. We also reduced our medical records department by four people. Our group, which has grown to 25 physicians, now saves nearly $250,000 a year by automating our workflow. And each physician has gained almost two hours a week in added productivity.
To conclude, our surgical group found an efficient data entry method that we could use for physicians without altering their normal work habits. We reduced errors by reducing human interaction with the EMR system and saved a lot of money.
