Reliability of Electronic Medical Record to Assess Patient's Eligibility for Lung Cancer Screening: Analysis of Two Pilot Trials.

INTRODUCTION: Accurate smoking history documentation in electronic medical records (EMRs) is essential for determining lung cancer screening (LCS) eligibility. We investigated the completeness of EMR smoking data, its concordance with patients' self-reports, and its reliability for LCS eligibility after 2021 LCS United States Preventive Services Task Force guideline change. METHODS: Between October 2023 and November 2024, patients aged 50 to 80 years with a primary care visit in academic and community practices of a health system, who were current or former smokers or had an unknown smoking history based on EMR, and who were being screened for two clinical trials on interventions to improve LCS were enrolled. Participants completed an LCS eligibility verification survey, and EMR smoking data from the most recent visit before PCP encounter were extracted for comparison. RESULTS: In total, 322 patients were included (mean age 62.4 ± 7.9; 58.1% male; 15.6% Asian, 1.2% Black, 61.4% White, and 21.8% other races; 45.7% Hispanic). Overall, 39.4% (127 of 322) had incomplete EMR smoking history to determine LCS eligibility. Among 167 smokers with both EMR and self-reported pack-year data, 85.6% (143 of 167) showed discordance in pack-year. Further, 26.2% (51 of 195) of patients with determined EMR eligibility (15.8% of 322 participants) had discordant eligibility based on self-report. Hispanic (versus non-Hispanic) patients were more likely to have incomplete EMR data to assess LCS eligibility (odds ratio, 1.68; 95% confidence interval 1.01-2.80) and discordant EMR eligibility with self-report (odds ratio, 2.34; 95% confidence interval 1.10-5.08). CONCLUSIONS: This study highlights significant gaps in smoking history documentation within EMRs, with high rates of incomplete and discordant data, particularly among Hispanic populations.