We read with great interest the publication titled “Inaccuracies in electronic health records smoking data and a potential approach to address resulting underestimation in determining lung cancer screening eligibility” by Kukhareva et al.1
We and others agree that the current manner in which electronic health records (EHRs) document smoking history is inadequate and may at least partially explain low rates of lung cancer screening.2,3 As a result, we previously sought to identify the value of different means of documenting smoking exposure on lung cancer screening eligibility and risk.4 We similarly assessed the value of longitudinal approaches to smoking exposure and their impact on screening eligibility and found that more patients were eligible for screening when longitudinal approaches were used instead of last-entered static values. We then assessed the impact of longitudinal smoking approaches on the real-world performance of standard screening criteria as well as risk prediction models. In our study population, the sensitivity to detect lung cancer using National Lung Screen Trial (NLST) criteria increased from 64.5% to 68.6% (P < .001), while specificity decreased from 56.4% to 53.8% (P < .001).5
With regards to model-based screening approaches, we found model-based predictions for risk of lung cancer were less dependent upon fluctuations in smoking history. This is likely because they have less of an effect on lung cancer risk than other variables (eg, demographic data and past medical history) available in the EHR. These model-based predictions also performed better than NLST-based criteria. We continue to advocate for model-based approaches to lung cancer screening in general.
We applaud Khukhareva and colleagues for bringing more attention to this important issue and agree with their suggestion that EHR vendors re-evaluate how smoking exposure data is presented. We have been discussing this issue with our vendor (Epic) and are hopeful that a better solution will soon be available.
CONFLICT OF INTEREST STATEMENT
None declared.
Contributor Information
Yasir Tarabichi, Center for Clinical Informatics Research and Education, MetroHealth, Cleveland, Ohio, USA.
J Daryl Thornton, Center for Clinical Informatics Research and Education, MetroHealth, Cleveland, Ohio, USA.
References
- 1. Kukhareva PV, Caverly TJ, Li H, et al. Inaccuracies in electronic health records smoking data and a potential approach to address resulting underestimation in determining lung cancer screening eligibility. J Am Med Inform Assoc 2022; 29 (5): 779–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Tarabichi Y, Kats DJ, Kaelber DC, Thornton JD.. The impact of fluctuations in pack-year smoking history in the electronic health record on lung cancer screening practices. Chest 2018; 153 (2): 575–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Modin HE, Fathi JT, Gilbert CR, et al. Pack-year cigarette smoking history for determination of lung cancer screening eligibility: comparison of the electronic medical record versus a shared decision making conversation. Ann Am Thorac Soc 2017; 14 (8): 1320–5. [DOI] [PubMed] [Google Scholar]
- 4. Kats DJ, Adie Y, Tlimat A, Greco PJ, Kaelber DC, Tarabichi Y.. Assessing different approaches to leveraging historical smoking exposure data to better select lung cancer screening candidates: a retrospective validation study. Nicotine Tob Res 2021; 23 (8): 1334–40. [DOI] [PubMed] [Google Scholar]
- 5. Aberle DR, Adams AM, Berg CD, et al. ; National Lung Screening Trial Research T. Reduced lung-cancer mortality with low-dose computed tomographic screening. N Engl J Med 2011; 365 (5): 395–409. [DOI] [PMC free article] [PubMed] [Google Scholar]