Chart Abstraction

The abstraction process has been a very important and challenging part of our EHR implementation. The data entered at this stage touches on many facets of the application. It is also important to have a proper representation of personnel at the abstraction planning stage to capture the many factors surrounding abstraction process flows that are decided upon at this stage. For instance, what chart data is important for us as a clinic to enter prior to a patient coming in and what data is not? Do we enter social and family history, immunizations, surgical history, chronic disease, labs etc? Do we want to go back 1 year or in some cases 20 years or maybe only a few months? How do we want this information displayed? Should we use only numbers and no characters for lab results (this can be important for reports), should we allow free text or alter the EHR in some areas so that only pre-set drop down menus of responses are available. Which charts are the clinics going to work on first for the abstraction process? How do we incorporate all of our government reporting (GPRA, Diabetic Audits, Immunizations) requirements into our abstraction process? These were just a few of the issues that had to be addressed before beginning this process.


The personnel participating in our initial abstraction planning included from CRIHB; the Integrated Care Coordinator, Technical Systems Coordinator and Health Systems Development Director. NextGen consultant and Project Champions from our MACT Health Board clinics also took part in the planning. We decided that some of our biggest issues were going to be finding the staff to actually manage the process of abstracting the charts. We needed a method that didn’t require in depth knowledge of the EHR and kept the process flowing as expeditiously as possible. Based on an abstraction template that our consultant had built for other organizations we decided to utilize some of the work they had already done and incorporate some of our needs into the existing template. One of the hurdles for us is our need to report on GPRA, Diabetes and Immunizations to the government. This meant that a great deal of historical data was important to us as some reporting requirements force us to go back many years. It also required a larger breadth of information to be entered than might be typical of a CHC. To allow staff to enter data quickly we did not want them to have to move around through 5-10 different templates within the NextGen system so we incorporated all of the necessary fields into one “hybrid” template.

Once we had a template that was acceptable to everyone, we had to decide on the priority of patients to begin abstracting. In our case, we chose the diabetic patients because they are our biggest chronic disease base of patients and are typically seen the most often. We then chose to enter patients that were scheduled for appointments at least a few weeks in advance. Finally, patients that had been seen in the past 6 months were abstracted followed by patients seen in the past year etc. After the initial phase was complete the remaining patients would normally be entered a day or two in advance of scheduled appointments.

The template has become a very important piece of our process. The idea that some data could be extracted by non-clinical people was helpful in finding personnel that had available time to initiate the process. We created sections within the template that were labeled as nurse/provider only which of course were areas of the abstraction process that we agreed should only be entered by a trained clinician. Examples of this would be immunizations, lab results, medications, diagnostics etc.

The information we chose to include in our abstraction process included the following:
Chronic Problems
Past Medical and or Surgical History
Allergies
Diagnostic History
Medications
Family History
Immunizations
Health Maintenance and Disease Management (including Flu, Pneumo, TD etc.)
Alcohol and or Tobacco Use
We added a grid at the bottom that gave all the information at a glance of the most recent dates for exams, labs, etc. to make it easy to see what a patient was missing.

In summary, we have abstracted close to 50% of our diabetic charts. Data elements were determined by many key stakeholders as well as the need to submit government reports. The largest challenge we faced was to instill a sense of urgency to get the data loaded. Although key stakeholders were engaged early in the process, the initial momentum to get the process kicked off started slowly. In order to complete the abstractions in a timely manner, clinic managers need to be prompted and coached to give their staff the time to complete abstracting duties. In our case, the clinic managers were not as proactive as we would have liked.