Identifying cases of spinal cord injury or disease in a primary care electronic medical record database

This study aimed to find people with spinal cord injuries or diseases (SCI/D) by looking at electronic medical records (EMR) from primary care doctors in Ontario, Canada. Researchers wanted to create a reliable way to identify these patients using information already in their medical records. The researchers manually reviewed a portion of patient records to create a "gold standard" list of SCI/D cases. Then, they tested different methods (algorithms) that used keywords and billing codes to see which one best matched the gold standard list. The best method combined keyword searches in the free-text parts of the medical records with a specific physician billing code. This method was pretty good at finding most of the SCI/D cases without flagging too many people who didn't have the condition.

• 1
  The optimal case identification algorithm used free-text keyword searches and a physician billing code, and had 70.6% sensitivity (61.9–78.4), 98.5% specificity (97.3–99.3), 89.9% PPV (82.2–95.0), 94.7% NPV (92.8–96.3), and an F-score of 79.1.
• 2
  Relevant ICD-9 codes for SCI/D were found to be unreliable, with only 15.9% of the RSC showing an SCI/D-related physician billing code.
• 3
  Occurrences in progress notes and consultation letters of highly specific terms such as “tetraplegia,” “neurogenic bowel,” and “spinal cord injury” proved to be a valuable addition to the contents of the CPP for the purpose of case identification.