For several years our residency program had been considering the purchase of an electronic medicalrecord (EMR). We had seen demonstrations of the products and were aware that a few practices in our area were using full EMRs. We knew of their many advantages, as delineated in several recent articles.1-3
Our faculty had 3 major concerns about the purchase of an EMR. The first was price. The cost of implementation cited in the literature varies widely; one source estimates $15,000 per full-time physician.1 There is disagreement as to whether the operating expenses of paperless EMR systems are less than traditional paper systems.4 The savings in dictation and filing are often offset by fees for service agreements and technical support. Thus, it seemed unlikely that an EMR would significantly decrease our operating costs in the immediate future.
Our second concern was the potential for physician dissatisfaction and disruption of clinical flow. At least 10 residency programs had purchased commercial EMRs and discontinued using them.5-7 Our clinic was running smoothly, so we believed that the advantages of a full EMR would not compensate for the inconveniences and frustrations that seem to accompany a commercial product. Our third concern was the apparent lack of a dominant EMR vendor. A recent survey of the industry revealed tremendous turnover,8 and a survey of family medicine residencies reported that no vendor had more than 25% of the market.5 Consequently, we feared investing in a product when its vendor might go out of business.
Our foremost goal was quality improvement (QI). This should include electronic reminders for due prevention items, the ability to display our completion rates for key prevention items without the time and expense of pulling charts, and the ability to check on critical combinations of diagnoses and medications (eg, congestive heart failure and ß-blocker usage). Second, we wanted to improve the legibility and accessibility of key parts of a patient’s chart, particularly medications and chronic diagnoses. The ability to access the full chart electronically and to change our current dictation of daily SOAP Notes (SpeechStudio; Portland, Ore) were less important to us.
Development
Since we did not believe a commercial program would meet our goals, we decided to create our own partial or miniature electronic medical record (mini EMR). Several reports in the literature have described the value of mini EMRs.9,10 One of the authors with previous programming experience (R.D.C.) began writing the first version in May 1999. We found a formulary database, Multum (Multum Information Services, Denver, Colo) from which we could import generic and trade medication names and categories. We also created a list of 700 primary care International Classification of Diseases–9th revision (ICD-9) codes common in our practice.
Current Use
Starting in May 2000, all of our 6.5 full-time equivalent physicians began using the mini EMR. Our patients’ demographic data were initially imported from our billing program into the mini EMR from a delimited text file. This same method is used to update phone numbers monthly. We had traditionally placed a preprinted sheet of paper for notes and orders on the front of each patient’s chart at each visit. This sheet was replaced with a printout from the mini EMR that included current ICD-9 codes, medications, and reminders for age- and sex-appropriate due prevention items. Front sheets are batch-printed each morning, then placed in the patient’s chart where it remains until the next visit, to be replaced by the most current printed version. When dictating the visit, the physician also updates the mini EMR entry for that patient on the computer. It takes approximately 30 seconds to call up a patient record and enter or change several diagnoses or medications or to add prevention item dates. This is not additional time, since most physicians would otherwise have to update the problem and medication lists in the paper chart. However, it does require that the physician be at a computer terminal. Physicians or nurses also update the mini EMR as data from Papanicolaou tests, laboratory values, and so forth, become available.
Microsoft Access has proved to be very stable, and we have not experienced any system crashes or lockups. Security issues are addressed in 3 ways. First, Access allows group and individual log-in names and passwords for both individual and networked computers. Second, it can generate an audit trail for any changes made to the database. Finally, the database should be used on a standalone network or behind a firewall.
Six months after full implementation of the mini EMR, more than 75% of our patients older than 50 years (N=1912) had been entered into the mini EMR. Acceptance has been very high, even by care providers who were admittedly computer phobic. Since we can easily query the mini EMR and determine our adherence rates for prevention items and medication usage, we are planning a number of QI projects, including improving mammography rates, using β-blockers in patients with congestive heart failure, and lowering low-density lipoprotein levels.