Errors in Table 2
JAMA Network Open · 22.04.2026 · 9 (4) · e2613456
DOI: 10.1001/jamanetworkopen.2026.13456 ↗Kurz gefasst
Korrektur in der Zeitschrift JAMA Network Open, online erschienen am 22.04.2026. Die Arbeit nennt 1 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar.
Abstract (Originalsprache)
In Reply We sincerely appreciate the engagement of these thoughtful clinicians to push the interpretation of the findings of our cohort study of chemotherapy in patients with non–small cell lung cancer1 toward greater meaning. One of the greatest challenges of retrospective study is the inability to know why patients were treated the way they were, as the selection process embedded in the treatment planning may introduce bias. In regard to the “apples to oranges” analogy brought up by Dai and colleagues, if the “observed” patients were medically unfit for chemotherapy, the same health issues would pose independent risk to survival and confound the mortality risk from cancer (ie, make the observed patients appear worse). We attempted to mitigate this bias using standard techniques (propensity matching and Cox proportional hazards) using the wide array of sociodemographic variables and comorbidity index data from the National Cancer Database (NCDB).1,2 We also leveraged a unique advantage of the NCDB, which allows patients unfit for treatment to be excluded. In the current study, patients for whom “chemotherapy was not recommended or administered because it was contraindicated owing to patient risk factors”1 were excluded to minimize the influence of poor baseline health on effect estimates. Moreover, we restricted the study population to only those patients who had undergone and survived surgery to create a cohort of patients who had a reasonable performance status (in the study by Kawaguchi et al,3 75% of patients who underwent surgery had a performance status of 0). Although it was not possible to control for all the variables that might have factored into the decision to offer adjuvant chemotherapy, we believe that the influence of health-related bias on survival was minimal in the present study.1 We agree with both Dai et al and Inamura that variables such as histologic subtype of adenocarcinoma and tumor invasion size could potentially influence the benefit of adjuvant chemotherapy in early-stage lung cancer. Because the NCDB does not capture all of the nuanced histopathologic features, future studies using more granular data sets could, perhaps, shed more light into this important question. Ultimately, we agree with our colleagues that the findings of the present study1 should not be interpreted to represent the last word in chemotherapy eligibility. Rather, they highlight the need to consider multiple parameters when engaging in shared decision-making conversations over the risks and benefits of adjuvant chemotherapy in patients with earlystage lung cancer.
Zitieren
APA
(2026). Errors in Table 2. JAMA Network Open, 9(4), e2613456. https://doi.org/10.1001/jamanetworkopen.2026.13456
Vancouver
Errors in Table 2. JAMA Network Open. 2026;9(4):e2613456. doi:10.1001/jamanetworkopen.2026.13456
BibTeX
@article{euvimed2026,
title = {Errors in Table 2},
journal = {JAMA Network Open},
year = {2026},
volume = {9},
number = {4},
pages = {e2613456},
doi = {10.1001/jamanetworkopen.2026.13456}
}
EUVIMED