ASPO Abstracts
Evaluating the OneFlorida Clinical Data Research Network as a Potential Resource to Investigate Smoking-Related Treatment Effect Modification in Cancer Patients.
Category: Survivorship & Health Outcomes/Comparative Effectiveness Research
Conference Year: 2019
Abstract Body:
Purpose. Limited surveillance data exist to examine smoking-related effect modification of treatment outcomes among cancer patients. By examining longitudinal properties of smoking status assessment, our objective was to demonstrate the potential opportunity for investigating how changes in smoking status could affect cancer treatment outcomes. Methods. The OneFlorida Clinical Research Consortium is a statewide clinical data research network that collects claims and electronic health record (EHR) data from over 15 million patients in Florida. Adult cancer patients were identified in the OneFlorida Data Trust (2012-2017) as having ≥1 encounter with an ICD-9-CM or ICD-10-CM diagnosis code for cancer and ≥1 EHR-documented assessments of smoking status within 1 year of diagnosis. Patients meeting inclusion criteria were tabulated. Current and former smoking prevalence was reported by patient demographics, cancer site, and phase of care (1 year to >30 days pre-diagnosis, within 30 days of diagnosis date, >30 days to 1 year post-diagnosis). Results. Among the 102,048 cancer patients who met inclusion criteria, 50% were female, 60% were ≥ age 65, and 64% were Caucasian. Overall, 16% were current smokers and 43% former smokers. Across cancer sites, prevalence of current smoking at diagnosis was highest for cervical cancer (30%), followed by lung cancer (27%) and liver cancer (24%). Among 16,683 cancer patients reporting current smoking at last assessment, 62% had a final smoking assessment >30 days after diagnosis and 49% had ≥2 assessments at or following a cancer diagnosis. Among patients who were current smokers at earliest available assessment in the EHR (n=18,664), 51% remained current smokers and 15% transitioned to former smokers. Conclusions. Clinical data from the OneFlorida Clinical Research Consortium representing usual care for Floridians with cancer allow for longitudinal tracking of smoking status in cancer patients that is not captured by traditional surveillance methods. Leveraging existing OneFlorida data linkages with tumor registry, healthcare utilization and health outcomes could facilitate improved understanding of how effect modification from changes in smoking behavior could affect cancer treatment.
Keywords: smoking, tobacco, cancer survivorship, electronic health record