P014 VALIDATING PATIENT REPORTED UTILIZATION IN THE IBD QORUS LEARNING HEALTH SYSTEM

The IBD Qorus initiative is a learning health system built on a collaboration between patients and providers that aims to improve the healthcare value for IBD patients. A key feature of learning health systems is the use of data within a collaborative network of patients and providers to learn and disseminate high value practices. IBD Qorus relies for a large extent on patients to self-report their utilization and quality indicator data for this purpose. Therefore, the reliability of these self-reported data elements is of key-importance. The aim of this study is to validate self-reported utilization and quality measures within Qorus.

Self-reported utilization and quality indicator data in Qorus was compared to data in the EMR for validation purposes. Chart reviews were performed to extract the number of ED visits, hospitalizations, and CT scans, as well as medication use and vaccination records. The percentage agreement, sensitivity and specificity were calculated.

In total 168 self-reported records of 101 IBD patients have been reviewed at one participating academic clinic and one community clinic. Self-reported IBD-related ED visits had an 80% sensitivity and 93% specificity; IBD related hospitalizations a 95% sensitivity and 96% specificity; and IBD-related CT scans a 100% sensitivity and 86% specificity. A small number of self-reported ED visits, hospitalizations, and CT scans were not recorded in the medical record (5%, 1%, and 11% of the records, respectively). Self-reported medication use was reported with a sensitivity of 70% for both steroids and narcotics, and a specificity of 98% and 96%, respectively. Only 49% of influenza vaccinations and 60% of pneumococcal vaccinations reported by patients were recorded in the EMR.

Over 80% of patients with a known ED visit, hospitalization, or CT scan reported these events accurately. However, self-report of medication use might be less reliable as only 70% of patients who used steroids or narcotics reported this. Patients’ self-reported utilization might also help identify a small number of ED visits, hospitalizations and CT scans that providers might have been unaware of otherwise. The collection of self-reported vaccination status might augment the data in the EMR, as less than 60% of the vaccinations were recorded in the EMR, indicating that the EMR is an unreliable data source for this purpose.