1018 — Identifying Veterans with Delays in Diagnostic Evaluation of Cancer using Electronic Triggers

Lead/Presenter: Ashley Meyer, COIN - Houston
All Authors: Meyer AN (Houston VA Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center & Baylor College of Medicine) Murphy DR (Houston VA Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center & Baylor College of Medicine) Vaghani V (Houston VA Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center & Baylor College of Medicine) Russo E (Houston VA Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center & Baylor College of Medicine) Sittig DF (School of Biomedical Informatics & UT-Memorial Hermann Center for Healthcare Quality and Safety,The University of Texas Health Science Center, Houston) Wei L (Houston VA Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center & Baylor College of Medicine) Wu L (Houston VA Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center & Baylor College of Medicine) Singh H (Houston VA Center for Innovations in Quality, Effectiveness and Safety, Michael E. DeBakey VA Medical Center & Baylor College of Medicine)

Objectives:
To create and validate electronic "trigger" algorithms that analyze electronic health record (EHR) data in VINCI (VA's Informatics and Computing Infrastructure) to identify Veterans with possible delays in diagnostic evaluation for colorectal (CRC) or hepatocellular cancer (HCC).

Methods:
We created two trigger algorithms, each using structured clinical data to identify patient records with laboratory results suggestive of CRC or HCC. Specifically, the triggers identified iron deficiency anemia or positive fecal immunochemical tests (CRC trigger) or elevated alpha-fetoprotein (HCC trigger), then excluded patient records where follow-up would be unnecessary (e.g., those in hospice) and records where follow-up was documented within 60 days. We then validated the triggers by applying them retrospectively to EHR data from one large VA Integrated Service Network (VISN) and reviewing triggered records to verify delays in diagnostic evaluation.

Results:
Using the CRC trigger, 1,073 patient records were triggered out of 245,158 patients seen in 2013. Chart reviews on 400 randomly-selected triggered records verified 224 cases of delayed diagnostic evaluation for CRC, yielding a positive predictive value (PPV) of 56.0% (95% CI:51.0-61.0%). Within 2 years, 4 cases were diagnosed with cancer. Extrapolating to all 1,073 cases would yield 601 patients with delayed diagnostic evaluation for CRC and 19 potential cases of cancer diagnosis. Using the HCC trigger, 130 records were triggered out of 333,828 patients seen between 2011-2014, with 107 patients with delayed diagnostic evaluation verified on chart review, (PPV of 82.3% [95% CI:74.4-88.2%]) and 9 cases of HCC. Extrapolated sensitivity and specificity were 68.6% (95% CI:65.4-71.6%) and 81.1% (95% CI:79.5-82.6%), respectively for the CRC trigger, and 89.1% (95% CI: 81.8-93.8%) and 96.5% (95% CI: 94.8-97.7%) for the HCC trigger. Across both cancers, the trigger reduced the number of record reviews needed to identify a diagnostic delay by 78.7%.

Implications:
In one VISN in one year alone, use of these two triggers can identify over 600 Veterans with potentially delayed diagnostic evaluation of gastrointestinal cancers.

Impacts:
Implementation of these triggers prospectively would enable the VA to intervene in Veterans' care before they experience potentially detrimental care delays.