Grant: Health System Impact Fellowships - PostDoc Western Canada 2-year

Spinal cord injuries can lead to loss of many bodily functions. They are one of the most costly conditions to treat. In Canada, a growing number of people, particularly older adults, experience spinal cord damage not from traumatic accidents, but from a variety of conditions including spinal degeneration associated with aging, tumours, or blood vessel problems. These are called non-traumatic spinal cord injuries (NTSCI). When compared to those with traumatic injuries, these patients are often underrepresented in our health data because there is no unique diagnostic code to identify them. This critical data gap means health leaders do not know the absolute number of people affected. Therefore, health service planning for NTSCI is difficult, resulting in inequitable access to care and potentially worse outcomes for this vulnerable population. Our project is intended to co-develop and validate a new strategy to identify patients with NTSCI in Alberta. First, we will screen multiple sources of health data, such as physician billing and clinical notes. Next, we will use diagnostic and procedure codes, as well as hospital encounter records, to identify possible cases. Then, we will apply artificial intelligence techniques including natural language processing to review relevant notes and enhance the accuracy. The primary objective is to develop a comprehensive case-finding strategy with improved accuracy. These studies make a critical first step towards understanding the true scope of NTSCI provincially. Using this method, health systems will be able to monitor case volumes and flag patients that belong to this underrepresented and vulnerable populations. Provincial health services will hence be able to reallocate funding, launch targeted outreach, and redesign care pathways to ensure that people with NTSCI receive the health services they need. We will transfer our knowledge, enable other jurisdictions to use, and tailor service planning to their local contexts.