Context
I owned data processing and CDC deliverables for Florida's Enhanced State Opioid Overdose Surveillance (ESOOS) program, working with death certificates, hospital records, labs, and EMS records.
What I built
- SAS code for state- and county-level quarterly non-fatal overdose reports, and for state-level fatal opioid cases with automatic upload into SUDORS.
- Text mining of EMS clinical notes to extract entities and structure the data, improving syndromic surveillance case definitions and predictive models.
- Analysis that identified geographically specific, statistically significant changes in non-fatal overdose rates.

Results
- Seven public-facing state surveillance reports.
- Breakout oral presentation at the 2019 CSTE Annual Conference: "Text Mining Improves Opioid Overdose Syndromic Surveillance in Florida."
- Findings also presented to the CDC EMS All-State Workgroup, at a CDC reverse site visit, and at the International Society for Disease Surveillance.