By Stephen Rothemich, MD, MS, BreathEasy Co-Principal Investigator and Clinical Lead
With BreathEasy’s focus on people with asthma and comorbid anxiety and/or depression, one of the challenges is how to identify potential subjects for upcoming focus groups. While we are working with two inner-city clinics using the same electronic medical record (EMR), the process has not been as simple as you might expect, as our health system is still in the process of implementing the EMR.
Ideally we could have used patients’ problem lists in the EMR to identify those with the overlapping conditions of interest. While both practices have been using the EMR for labs and prescriptions for a while now, one site began full electronic documentation just a few months ago. Since it will take a while for these practice populations to have office visits, most patients don’t yet have an electronic problem list. Identifying potential subjects starting with manual reviews of the lists in paper charts was impractical; there were 6,277 patients in our target age range!
Next was the diagnosis billing codes from recent visits. Here we encountered two pitfalls: 1) every chronic condition present or even addressed at a visit doesn’t necessarily get included in the maximum of four diagnosis codes that we submit for each visit and 2) our billing system is separate from our EMR system. Since we just had access to the latter for this project, we could only capture billing diagnoses associated with orders for labs or x-rays.
That left the medication lists. Both clinics have been using the EMR to generate prescriptions for over a year. So while we incorporated data from problem and diagnosis lists, our search strategy relied on the better populated medication lists. However, medications can be prescribed for more than one purpose. Some antidepressants are also used for sleep or chronic pain, or even conditions we needed to exclude, such as bipolar disorder. An oral medication might be used for allergies and not for asthma control.
We found 271 potential subjects based on all available electronic data in the problem lists, diagnoses, and medications that might be used for the target conditions. Because we relied so heavily on medication data and the exact indication for many prescriptions was ambiguous, review of electronic charts was necessary in 176 cases. Paper charts were reviewed for 38 patients where EMR review was not helpful. All in all, we had 151 subjects with the conditions of interest. While we didn’t have to do the whole process with paper chart review, good electronic problem lists would have made life easier. So I’ve decided to stop grumbling about building those lists when I see my own patients. Stay tuned to learn what we’ve heard from the focus groups!