As medical practices move from the paper world to the electronic world, many find it helpful to point toward some tangible benefit to convince their physicians to cooperate. In our practice, the benefit was improved patient care made possible by our electronic health record (EHR) system's reporting capabilities. It has turned our physicians into EHR believers.

In the beginning

In May 2001, our EHR went live, beginning with just a few providers. For five months, we added two or three providers to the system every two weeks until all of our faculty members, residents and midlevel providers were using the EHR. We were on our way.

Our clinic staff was thrilled. Medical records were no longer misplaced, physician notes and prescriptions were legible, and office processes were more efficient.

Our physicians, however, were not impressed, despite the time and effort we spent installing the EHR system. They complained that office notes took just as long or in some cases longer than in the pre-EHR days. The residents vowed never to join a practice with an EHR.

The physicians' attitudes improved a bit when we added a messaging system in November 2001. It allowed them to answer messages from anywhere they had access to the EHR and made information more accessible because they did not have to be in their office to look at paper charts. They no longer had to hunt down their nurse or play phone tag to exchange information. The physicians' acceptance improved again in 2002 when the addition of document scanning turned our office into a paperless operation.

Finally, we saw our first glimpse of excitement and enthusiasm from the physicians in January 2003 when we began using the EHR to look at populations of patients. It had taken two years, but we had gone from residents and faculty who vowed never to practice with an EHR again to physicians who could not envision functioning without it.

Capturing the enthusiasm

In 2003, we began using software called Crystal Reports to harvest the patient and disease population information our EHR was collecting. We found that this software provides report-writing capability for any SQL-based (structured query language) database and is therefore compatible with many EHR systems. It allows the report writer to tie templates together; manipulate records; query specific data fields; and format, formulate and sort data.

We found a number of powerful motivations for using our EHR for reporting. Information about patient populations that was previously almost impossible to obtain was now readily accessible. For example, when a medication was recalled in the past, we did not have the ability to look through 10,000 charts to determine which patients were affected. Now we can look through those 10,000 charts quickly. (In fact, we did this after a recent drug recall, generating a report and sending letters to all affected patients within one day.)

Initially, we wrote a number of meaningless reports as we learned the system's intricacies. Gradually, our reports became useful. We found them particularly valuable in following health maintenance and in disease populations.

We now use the reports as part of our quality improvement model, which consists of the following steps:

1. Establish benchmarks,

2. Select champions,

3. Set goals for improvement,

4. Design and run reports,

5. Educate and communicate results to providers, staff and patients,

6. Develop incentives,

7. Monitor results.

Destination No. 1: Diabetes

Our practice serves more than 500 patients with diabetes. As providers of care for many in the inner city, we serve an often challenging patient population. Some of our patients cannot afford medications. Many have difficulty finding transportation to get to their doctor's appointments. Most have multiple co-morbid illnesses. Our diabetes population seemed to be a prime target for using the EHR to improve patient compliance and outcomes.

We were fortunate in that we had local and national benchmarks to use for our quality improvement model. Our Department of Family and Community Medicine's Clinical Division had been monitoring key measures in diabetes for several years. We therefore had previous data on our own practice and that of the other residency practices. In addition, we used Health Plan Employer Data and Information Set (HEDIS) information for comparison with Medic-aid and commercial populations.

Next, we decided that we would use a physician and a staff champion for each project. The physician champion was important for physician buy-in, and the staff champion provided continuity, as most providers in our residency clinic are part-time. We also established a steering team made up of faculty, residents and staff.

We set goals incrementally along the way, using specific, measurable, time-limited benchmarks. We made the goals realistic but challenging. Our ultimate destination in diabetes care was and is to meet the levels set by the Diabetes Physician Recognition Program (http://www.ncqa.org/dprp/).

We were prepared for our patients' A1C and LDL levels to remain the same because most of diabetes care depends on patient adherence. We anticipated that we would see improvement in the proportion of patients being tested. We thought this would be a good start. With nursing staff monitoring the reports and contacting patients with telephone calls and EHR-generated letters, it did improve.

However, things that required physician participation, such as foot exams, were more challenging. Our initial reports showed that although many physicians reported doing foot exams, only a few were documenting them. Of those physicians who were documenting them, many were using free text to type this into the record because they either ignored the check boxes or did not know where to find them. Our report-writing software cannot query free text; it relies on the use of check boxes, pick-lists and other standardized text.

As in many situations, the answer was communication. We provided regular communication at department meetings about the project, its goals and our results. The information included our practice's compliance rates, outcomes measures and patient satisfaction with diabetes care. Physicians were given their own diabetes registries, compliance rates and outcomes quarterly. Their rates were compared to those of the entire practice and the Diabetes Physician Recognition Program criteria.

Incentives were established for both patients and providers. Patients whose current A1C rate was under 8 percent received a pin recognizing this accomplishment. When providers' diabetes patient panels had more than 50 percent with a documented foot exam, they were awarded a pair of socks or a foot-shaped key chain. The incentive for providers was surprisingly effective. It set up a friendly rivalry that quickly accelerated our rates to an acceptable level.

Approaching our goal

We saw significant improvement in all of our measures, most of which now exceed HEDIS averages for both Medicaid and commercial populations. We are rapidly approaching our ultimate goal, which is to exceed the Diabetic Physician Recognition Program's benchmarks.

We are particularly gratified by the huge improvement in monitoring LDL, which has risen from 32 percent in 2001 to 91 percent in 2005. Similarly, compliance with foot exams has improved from 1 percent in 2001 to 80 percent in 2005.

The ability to regularly monitor patient compliance and results is essential. We run reports monthly and share compliance rates with the providers and staff at department meetings. The reports are also given to nursing staff, who contact patients by phone and mail to schedule appointments and testing. This two-tiered approach is quite effective.

Provider-specific reports are compiled quarterly and given to faculty, residents and midlevel providers. Physicians receive individual diabetes patient panel reports comparing their patients to the entire practice. In addition, they are given the raw data and nursing staff support to work individually on improving patient compliance. Again, we found that a healthy sense of competition stimulated significant improvement in compliance.

The next step

While EHR reporting has helped us improve our performance on key diabetes measures, we are planning further assessments of health outcomes.

First, surveys are being distributed to our patients with diabetes. These surveys encompass many parameters, including the patients' subjective opinion on how they feel their diabetes is managed and how they rate their health. Because these surveys have also been entered into our EHR, we can easily track and extract information. As we improve individual medical parameters over time, we can survey patients again and compare their health outcome opinions.

Second, we are following patients over time and tracking diabetes-related hospitalizations. We will be able to see if there is a correlation between improved parameters such as glycemic control and a decrease in glycemic-related hospitalizations.

These two methods should give us both objective and subjective data. The objective data is useful for our medical community and educational advancement. The subjective data is useful for our patients. If we can show other patients that by improving certain medical parameters they will feel better, we will have a greater likelihood of influencing our patients' attitudes about their disease and their need to be a part of its management.

Our journey has illustrated that an EHR system can be used for more than just generating patient notes. Because of its reporting capabilities, it has made our practice more efficient and has greatly improved patient care. In the process, something remarkable and unexpected has happened. All providers, even the physicians, have embraced the EHR and the enhancements it can bring to patient care.