Electronic medical records represent unfulfilled promise for identifying patients who are at elevated risk for osteoporosis and fragility fractures, according to a survey of the situation in Ontario, Canada. As reported at the American Medical Informatics Association 2014 Annual Symposium, electronic medical records (EMRs) from family physicians across the province of Ontario that were contributed to a database dubbed EMRALD (EMR Administrative Data Linked Database) contain widely varying report formats for bone mineral density (BMD) results.1 EMRALD was created in 2007 by Karen Tu, MD, and her colleagues from the Department of Family and Community Medicine, University of Toronto. It is housed at the Institute for Clinical Evaluative Sciences in Toronto.
“It’s encouraging that there are a lot of BMD exam results in EMRALD, but a bit discouraging to see how variably the results and other information about osteoporosis are stored,” presenter Sonya Allin, PhD, told Value-Based Care in Rheumatology. For example, among the osteoporosis patients included in the electronic-record review “only about half had BMD exam results in the EMRs and less than two-thirds had their osteoporosis diagnosis coded in their patient profile.” Some members of the same research team showed in an earlier study that there is a woefully low rate of age-standardized BMD testing within 6 months of a fracture in Ontario patients.2
Allin, a researcher in the Department of Physical Therapy, University of Toronto, and Tu, and their coinvestigators focused on 15,365 patients in EMRALD who had diagnostic-imaging entries for BMD exams. They found that, while the rate of BMD testing recorded in EMRALD does reflect the rate shown in records from the provincial health agency, pulling diagnoses and fracture risk information from EMRALD’s BMD reports was complicated by lack of standardization. Among all the patients with BMD exam entries in EMRALD, 27.8% did not have BMD results in searchable text format, and 8% did not mention the term T-score. Even when text entries did contain a BMD exam result, the formats were “extremely varied,” noted the researchers: 27.3% were less than 200 characters long and many were short annotations from primary care physicians. Most of the BMD results were attached to the EMR either as an image or scan.
Furthermore, when members of the research team manually reviewed the EMRs of some of the patients, they found only 55.8% of osteoporosis patients’ BMD results were readily accessible. The most common place for a mention of the osteoporosis diagnoses in the EMRs, at 90.2%, was in a note from a primary care provider attached to an individual patient exam.
Fortunately, the patients in EMRALD who experienced a fragility fracture did seem to be relatively well investigated for osteoporosis.
“In the future, we’d like to relate the representation of fragility fractures in the EMR to their representation in administrative data sources; this might tell us how many fractures are not communicated to primary care providers by patients,” reflected Allin. She also noted that comparable data quality issues are likely present in American primary care physicians’ EMRs and administrative databases.
- Allin S, Munce S, Jaglal S, et al. Capture of osteoporosis and fracture information in an electronic medical record database from primary care. Presented at: American Medical Informatics Association 2014 Annual Symposium; November 15-19, 2014; Washington, DC. http://knowledge.amia.org/. Accessed January 5, 2015.
- Jaglal S, Hawker G, Croxford R, et al. Impact of a change in physician reimbursement on bone mineral density testing in Ontario, Canada: a population-based study. CMAJ Open. 2014;2(2):E45-E50. www.cmajopen.ca/content/2/2/E45.full. Accessed January 5, 2015.