Review Reports - Lung Cancer Mortality and In-Hospital Outcomes in Chile: A Population-Based Cross-Sectional Study

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Manuscript ID: healthcare-4534303 The study addresses an important public-health issue and uses a large national dataset to examine lung cancer mortality and hospital outcomes in Chile. The findings are relevant, but several methodological and interpretive issues should be clarified before publication.

The word already mentioned in the title should not be repeated in keyword, select correct keyword relevant to the study.

Please clarify the rationale for including C33-C34 in the lung cancer definition, particularly C33 (tracheal cancer).
The hospitalization analysis includes repeated admissions from the same patients; please account for within-patient clustering in the regression analysis.
Important confounders such as tumor stage, smoking history, treatment, histology & comorbidities are unavailable & should be emphasized when interpreting the associations.
The category “acute respiratory complications” combines respiratory failure, pleural effusionand pulmonary embolism. These should ideally be analyzed separately.
Please reconcile the inconsistent readmission estimates reported in the main text and supplementary tables.
The reported COPD OR of 0.00 and extremely low OR for “Other medical care” suggest model instability and require clarification.
The substantial reduction in hospitalizations during 2020–2021 warrants a sensitivity analysis considering the potential impact of COVID-19.
The association between pneumonia and mortality should not be interpreted causally; the recommendation regarding vaccination should therefore be stated more cautiously.
The discussion attributes sex-specific mortality trends partly to smoking and other exposures, but these variables were not measured in the study. Such explanations should be presented as possible rather than established mechanisms.
Please strengthen the clinical interpretation by distinguishing active cancer treatment, disease progression & end-of-life/palliative admissions, which cannot currently be differentiated from the administrative data.

How the findings of the present study are relevant on a global scale.

Comments on the Quality of English Language

May be improved.

Author Response

We sincerely thank the reviewer for the careful and constructive evaluation of our manuscript. The comments were highly valuable and helped us improve the methodological clarity, interpretation, and overall presentation of the study. We have addressed each point in detail and revised the manuscript accordingly. In addition, we have carefully reviewed the English throughout the manuscript to improve grammar, clarity, and readability.

The word already mentioned in the title should not be repeated in keyword, select correct keyword relevant to the study.

Response: Thank you for this observation. We agree that keywords should complement the title rather than repeat terms already included in it. Accordingly, we have revised the keywords and replaced those overlapping with the title with more specific terms relevant to the study.

Please clarify the rationale for including C33-C34 in the lung cancer definition, particularly C33 (tracheal cancer).

Response: Thank you for pointing this out. Upon reviewing our analytical code, we confirmed that the analyses were correctly performed using ICD-10 code C34 (malignant neoplasm of bronchus and lung) only; C33 (malignant neoplasm of the trachea) was not included in the analysis. The reference to C33–C34 was therefore a reporting error that appeared in several sections of the manuscript rather than an error in the analytical definition. We have carefully reviewed the manuscript and corrected all occurrences of C33–C34 to C34, including the Abstract, Methods, Results, and figure/table descriptions where applicable. We thank the reviewer for identifying this inconsistency.

The hospitalization analysis includes repeated admissions from the same patients; please account for within-patient clustering in the regression analysis.

Response: We reanalyzed the hospitalization data accounting for repeated admissions within individuals. Hospitalization episodes were retained as the unit of analysis, and HC1 robust standard errors clustered by encrypted patient identifier were used to estimate 95% confidence intervals. The main associations remained essentially unchanged, including those for pneumonia (OR 2.05; 95% CI 1.76–2.38) and acute respiratory complications (OR 2.13; 95% CI 1.83–2.48). The Statistical Analysis section and corresponding estimates have been updated accordingly.

Important confounders such as tumor stage, smoking history, treatment, histology & comorbidities are unavailable & should be emphasized when interpreting the associations.

Response: Thank you for this important observation. We agree that the absence of key clinical variables may result in residual confounding. We have strengthened the limitations section to explicitly acknowledge the lack of information on tumor stage, smoking history, treatment, histology, performance status, and relevant comorbidities, and to clarify that residual confounding cannot be excluded when interpreting the observed associations with in-hospital mortality.

The category “acute respiratory complications” combines respiratory failure, pleural effusionand pulmonary embolism. These should ideally be analyzed separately.

Response: Following the reviewer’s suggestion, we performed a sensitivity analysis separating the acute respiratory complications category into respiratory failure (J96), pleural effusion (J90), and pulmonary embolism (I26). Compared with the tumor/metastasis principal-diagnosis category, respiratory failure was associated with higher odds of in-hospital death (adjusted OR 7.438; 95% CI 5.785–9.563), whereas the confidence intervals for pleural effusion (adjusted OR 0.709; 95% CI 0.501–1.004) and pulmonary embolism (adjusted OR 1.231; 95% CI 0.930–1.630) included the null value. These condition-specific results have been incorporated into the revised manuscript and Supplementary Materials.

Please reconcile the inconsistent readmission estimates reported in the main text and supplementary tables.

Response: We verified the readmission analyses, coding, and model design matrices and confirmed that first hospitalization was consistently used as the reference category. The apparently different estimates reflect different model specifications rather than inconsistent coding. We also identified and corrected an inverted comparison in the wording of Section 3.5. The Methods and supplementary table legends have been revised to explicitly specify the adjustment variables and reference categories for each model, thereby clarifying the differences across the reported readmission estimates.

The reported COPD OR of 0.00 and extremely low OR for “Other medical care” suggest model instability and require clarification.

Response: Thank you for highlighting this issue. We identified a labeling error in Table S2, where the estimates for COPD and “Other medical care” had been interchanged. The corrected COPD estimate among women is OR 1.36 (95% CI 0.87–2.07), while the near-zero estimate corresponds to Z51, for which no deaths occurred, resulting in separation. We therefore report this estimate as not estimable (NE). Among men, the Z51 estimate was based on only one death and is now explicitly identified as unstable due to sparse data. The tables and corresponding footnotes have been corrected accordingly.

The substantial reduction in hospitalizations during 2020–2021 warrants a sensitivity analysis considering the potential impact of COVID-19.

Response: We performed the requested sensitivity analysis excluding hospitalizations occurring in 2020–2021. The associations with in-hospital mortality persisted for pneumonia (OR 1.81; 95% CI 1.52–2.16) and acute respiratory complications (OR 1.97; 95% CI 1.65–2.36), although the point estimates were attenuated. We have added these findings to the Results and clarified in the Discussion that the reduction in hospitalizations during 2020–2021 may reflect pandemic-related changes in healthcare utilization.

The association between pneumonia and mortality should not be interpreted causally; the recommendation regarding vaccination should therefore be stated more cautiously.

Response: Thank you for this observation. We have revised the discussion to clarify that the association between pneumonia and in-hospital mortality should not be interpreted as causal. We have also replaced causal terminology with association-based language and revised the statement regarding vaccination to clarify that its potential benefit cannot be inferred directly from our findings.

The discussion attributes sex-specific mortality trends partly to smoking and other exposures, but these variables were not measured in the study. Such explanations should be presented as possible rather than established mechanisms.

Response: We agree that these mechanisms cannot be established from our data. Accordingly, we have revised the Discussion to frame smoking patterns, environmental exposures, and healthcare access as potential explanations supported by previous literature, while explicitly noting that these factors were not measured in our study.

Please strengthen the clinical interpretation by distinguishing active cancer treatment, disease progression & end-of-life/palliative admissions, which cannot currently be differentiated from the administrative data.

Response: Thank you for this observation. We have strengthened the clinical interpretation by clarifying in the Discussion that the administrative data cannot distinguish admissions for active cancer treatment, disease progression, or end-of-life/palliative care, which may have different clinical trajectories and mortality profiles.

How the findings of the present study are relevant on a global scale.

Response: Thank you for this important suggestion. We have expanded the Discussion to address the broader relevance of our findings. We now highlight how the sex-specific mortality patterns and hospital burden observed in Chile may be informative for other countries experiencing similar epidemiological transitions in lung cancer. We also emphasize the potential value of routinely collected mortality and hospital administrative data for monitoring lung cancer burden in settings where detailed clinical cancer registries may be limited.

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for providing this opportunity to review this manuscript.The topic is important, and the use of national administrative datasets provides potentially valuable evidence. However, several methodological, statistical, and interpretative issues need to be addressed before the conclusions can be considered sufficiently supported.

Major Comments

1. Please clarify the relationship between the two datasets and clearly distinguish the mortality population from the FONASA hospitalization population.
2. The definition of lung cancer needs clarification, as ICD-10 C33–C34 also includes tracheal cancer. Please justify this classification.
3. The mortality analysis would be stronger with a formal trend analysis, such as APC/AAPC with 95% CIs, rather than comparing only the beginning and end of the study period.
4. Please specify the standard population and method used for age standardization to improve reproducibility.
5. The difference between crude and age-adjusted mortality trends requires further explanation, including the standardization approach used.
6. Because several patients contributed multiple hospitalizations, the regression analysis should account for within-patient clustering rather than treating all hospitalization episodes as independent observations.
7. The comparison between first hospitalization and readmission may be affected by survival and selection bias because readmission occurs after the initial admission. This should be addressed in the analysis and interpretation.
8. There are inconsistent readmission ORs between the main text and Tables S2/S3. Please verify the analyses, coding, reference groups, and reported estimates.
9. The COPD OR of 0.00 with a CI of 0.00–0.00 appears erroneous. Please check the underlying data and model for coding or sparse-data problems.
10. The “acute respiratory complications” category combines clinically different conditions, including respiratory failure, pleural effusion, and pulmonary embolism. Please justify this grouping or analyse these conditions separately.
11. Respiratory failure and other severe complications may be terminal events rather than independent predictors of death. The possibility of reverse causation should be discussed.
12. The rationale for combining lung cancer and metastasis as the reference category should be explained, given their potentially different prognostic implications.
13. Administrative primary diagnoses may not always represent the main clinical cause of deterioration or death. This limitation should be considered when interpreting pneumonia and respiratory complications.
14. The association between pneumonia and mortality does not demonstrate that vaccination would prevent these deaths. The vaccination-related implications should therefore be stated more cautiously.
15. Please revise causal language throughout the manuscript and consistently describe the findings as associations rather than effects or risks.
16. Different ORs between men and women do not establish effect modification. Please provide a formal interaction test before making this conclusion.
17. Important indicators of disease severity, including tumor stage, performance status, metastatic burden, treatment, ICU admission, and comorbidities, were unavailable or not included. The resulting residual confounding should be discussed.
18. The reported 17.4% mortality represents deaths among hospitalization episodes in the FONASA database, not mortality among all patients with lung cancer in Chile. Please clarify this terminology.
19. The hospitalization data are not fully representative of the Chilean healthcare system because private-sector hospitalizations are not captured. This limitation should be emphasized.
20. The substantial reduction in hospitalizations during 2020–2021 should be examined in relation to the COVID-19 pandemic and changes in healthcare utilization.
21. A sensitivity analysis excluding or separately analysing the pandemic years would help determine whether the main findings are robust.
22. The manuscript should provide more detail on missing data and the assumption that absence of an ICD-10 code represents absence of the condition.
23. Please clarify how the hospitalization diagnostic categories were defined and provide a complete ICD-10 code mapping.

Minor Comments

The title uses “Hospital Outcomes,” although the principal hospital outcome is in-hospital mortality. Consider more precise wording.
Use “sex” consistently rather than alternating between sex and gender.
Clearly distinguish population-level mortality from mortality among FONASA hospitalization episodes in the abstract.
Provide exact numerators and denominators when reporting percentages, particularly for the approximately 75% cancer/metastasis category.
Report the number of deaths within each diagnostic category in addition to percentages.
Provide confidence intervals for mortality rates and temporal estimates where appropriate.
Expand the statistical methods to specify the software version, significance threshold, variable selection, treatment of age, model diagnostics, and handling of sparse data or separation.
Emphasize effect estimates and confidence intervals rather than statistical significance alone, particularly given the large sample size.
Explanations offered for increasing mortality among women, including smoking, environmental exposures, screening, and treatment access, were not directly measured. These should be presented as possible explanations rather than study findings. Overall reference is very less. can do extensive discussion by using recent literatures
The conclusion should be more cautious and limited to the associations supported by the available administrative data.

Revisions Before Publication

Resolve the inconsistent readmission estimates across the manuscript and supplementary tables.
Reanalyse the hospitalization data while accounting for repeated admissions within individuals.
Correct and investigate the COPD estimate of OR = 0.00.
Perform a formal interaction analysis before claiming sex-based effect modification.
Strengthen the mortality trend analysis using an appropriate temporal method.
Clarify the representativeness and denominator of the FONASA hospitalization population.
Address residual confounding related to tumor stage and disease severity.
Moderate causal and preventive interpretations, particularly regarding pneumonia and vaccination.
Clarify the use of ICD-10 C33–C34 for defining lung cancer.
Address the potential influence of the COVID-19 period through discussion and, preferably, sensitivity analysis.

Author Response

We sincerely thank the reviewer for the thorough and insightful evaluation of our manuscript. The detailed comments were particularly helpful in strengthening the methodological rigor, clarifying the use and interpretation of the two administrative datasets, and improving the presentation of the statistical analyses and study limitations. We have addressed each comment carefully and revised the manuscript accordingly. We also performed a comprehensive review of the English language to improve clarity, consistency, and readability throughout the manuscript.

Major Comments

Please clarify the relationship between the two datasets and clearly distinguish the mortality population from the FONASA hospitalization population.

Response: We have clarified this distinction in the Study Design section. The revised text specifies that the two databases were analyzed independently and were not linked at the individual level. It also clearly distinguishes the national mortality population captured by DEIS from the FONASA population included in the hospitalization analysis.

The definition of lung cancer needs clarification, as ICD-10 C33–C34 also includes tracheal cancer. Please justify this classification.

Response: Thank you for pointing this out. Upon reviewing our analytical code, we confirmed that the analyses were correctly performed using ICD-10 code C34 (malignant neoplasm of bronchus and lung) only. C33 (malignant neoplasm of the trachea) was not included in the analysis. The reference to C33–C34 was therefore a reporting error in the manuscript rather than an error in the analytical definition. We have carefully reviewed the manuscript and corrected all occurrences of C33–C34 to C34 throughout the text, tables, and figure descriptions, where applicable.

The mortality analysis would be stronger with a formal trend analysis, such as APC/AAPC with 95% CIs, rather than comparing only the beginning and end of the study period.

Response: We addressed this comment by incorporating a formal log-linear trend analysis for 2002–2024, reporting APC estimates with 95% CIs for the overall population and separately by sex. We also added mortality projections through 2035 using the Nordpred age-period-cohort model.

Please specify the standard population and method used for age standardization to improve reproducibility.

Response: We have clarified the age-standardization procedure in the Methods. Mortality rates were directly age-standardized using the 2024 Chilean population as the standard population.

The difference between crude and age-adjusted mortality trends requires further explanation, including the standardization approach used.

Response: We clarified the distinction between crude and age-adjusted mortality trends and specified that direct age standardization was performed using the WHO World Standard Population (2000–2025). The Discussion was also revised to interpret the differences between crude and age-adjusted trends in the context of changes in population age structure.

Because several patients contributed multiple hospitalizations, the regression analysis should account for within-patient clustering rather than treating all hospitalization episodes as independent observations.

Response: The regression analysis was revised to account for the non-independence of repeated hospitalization episodes. We estimated HC1 robust standard errors clustered at the patient level using the encrypted patient identifier. The main estimates remained stable, with pneumonia showing an OR of 2.05 (95% CI 1.76–2.38) and acute respiratory complications an OR of 2.13 (95% CI 1.83–2.48). The Statistical Analysis section and corresponding confidence intervals have been updated accordingly.

The comparison between first hospitalization and readmission may be affected by survival and selection bias because readmission occurs after the initial admission. This should be addressed in the analysis and interpretation.

Response: We agree that the comparison between first hospitalizations and readmissions may be affected by survival and selection bias because patients must survive a previous hospitalization to subsequently contribute a readmission. We have therefore revised Section 3.4.1 to explicitly acknowledge this issue and to avoid a causal interpretation of the observed difference. We also incorporated this consideration into the Limitations section, emphasizing that hospitalization sequence should be interpreted cautiously.

There are inconsistent readmission ORs between the main text and Tables S2/S3. Please verify the analyses, coding, reference groups, and reported estimates.

Response: We verified the coding and model design matrices and confirmed that first hospitalization was consistently used as the reference category. We identified an inverted comparison in the wording of Section 3.5, which has been corrected to “readmission vs. first hospitalization.” We also revised the Methods and the legends of the sex-stratified supplementary tables to explicitly specify the adjustment variables and reference categories for each model, thereby clarifying why the readmission ORs differ across the overall and sex-stratified analyses.

The COPD OR of 0.00 with a CI of 0.00–0.00 appears erroneous. Please check the underlying data and model for coding or sparse-data problems.

Response: We reviewed the underlying data and identified a labeling error in Table S2, where the estimates for COPD and “Other medical care” had been interchanged. The corrected COPD estimate among women is OR 1.36 (95% CI 0.87–2.07). The near-zero estimate corresponded to “Other medical care” (Z51), for which no deaths occurred among 681 hospitalizations, resulting in complete separation. This estimate is now reported as not estimable (NE), with an explanatory footnote.

The “acute respiratory complications” category combines clinically different conditions, including respiratory failure, pleural effusion, and pulmonary embolism. Please justify this grouping or analyse these conditions separately.

Response: We agree that these conditions represent clinically distinct events. We therefore performed an additional analysis separating respiratory failure (J96), pleural effusion (J90), and pulmonary embolism (I26). Respiratory failure was associated with higher odds of in-hospital death (OR 7.438; 95% CI 5.785–9.563), whereas the confidence intervals for pleural effusion (OR 0.709; 95% CI 0.501–1.004) and pulmonary embolism (OR 1.231; 95% CI 0.930–1.630) included the null value. These condition-specific results have been incorporated into the revised manuscript and Supplementary Materials.

Respiratory failure and other severe complications may be terminal events rather than independent predictors of death. The possibility of reverse causation should be discussed.

Response: We have expanded the Limitations section to acknowledge the possibility of reverse causation. We now clarify that severe acute complications, particularly respiratory failure, may reflect advanced clinical deterioration or terminal events and that their temporal relationship with mortality cannot be established from the available data.

The rationale for combining lung cancer and metastasis as the reference category should be explained, given their potentially different prognostic implications.

Response: We have clarified the rationale for selecting the tumor/metastasis category as the reference group in the Statistical Analysis section. This category was selected because it represents hospitalizations primarily attributed to the underlying malignant disease, providing a clinically relevant comparator for hospitalizations associated with respiratory and other diagnostic categories.

Administrative primary diagnoses may not always represent the main clinical cause of deterioration or death. This limitation should be considered when interpreting pneumonia and respiratory complications.

Response: We have added this limitation to the Discussion. We now clarify that the primary diagnosis recorded in administrative data may not necessarily represent the principal clinical cause of deterioration or death and that the associations observed for pneumonia and acute respiratory complications should not be interpreted as indicating that these conditions were the direct cause of death

The association between pneumonia and mortality does not demonstrate that vaccination would prevent these deaths. The vaccination-related implications should therefore be stated more cautiously.

Response: We have revised the “highlights” to avoid implying that the potential benefit of vaccination can be inferred from our findings. Specifically, we removed direct references to vaccination and reframed the finding in terms of the clinical relevance of respiratory infection prevention and management. We also clarified in the Discussion that vaccination status was not available in our data and therefore its potential impact could not be evaluated in this study.

Please revise causal language throughout the manuscript and consistently describe the findings as associations rather than effects or risks.

Response: We revised the manuscript to ensure that the findings are consistently described using non-causal, association-based language. Specifically, in the Results, Discussion, Conclusions, and Highlights, terms implying risk or causality were replaced, where appropriate, with expressions such as “associated with higher odds of in-hospital death” and “observed association.” These revisions better reflect the observational nature of the study and the effect measure estimated by the logistic regression models.

Different ORs between men and women do not establish effect modification. Please provide a formal interaction test before making this conclusion.

Response: We performed a formal interaction analysis as suggested. The joint test for the interaction between sex and diagnostic group did not provide evidence of effect modification. We therefore removed the statement that differences in sex-specific ORs indicate effect modification and now present the stratified estimates as exploratory findings.

Important indicators of disease severity, including tumor stage, performance status, metastatic burden, treatment, ICU admission, and comorbidities, were unavailable or not included. The resulting residual confounding should be discussed.

Response: Thank you for this important comment. We agree that the lack of detailed indicators of disease severity may result in residual confounding. We have expanded the limitations section to explicitly acknowledge the unavailability of tumor stage, performance status, metastatic burden, cancer treatment, ICU admission, and relevant comorbidities. We now clarify that these unmeasured factors may influence the observed associations with in-hospital mortality and that residual confounding cannot be excluded. In addition, we have revised the Discussion to emphasize that these associations should be interpreted cautiously because adjustment for important clinical indicators of disease severity and prognosis was not possible.

The reported 17.4% mortality represents deaths among hospitalization episodes in the FONASA database, not mortality among all patients with lung cancer in Chile. Please clarify this terminology.

Response: We agree and have clarified this terminology throughout the manuscript. The 17.4% estimate represents episode-level in-hospital mortality among hospitalization episodes recorded in the GRD-FONASA database and should not be interpreted as mortality among all patients with lung cancer in Chile.

The hospitalization data are not fully representative of the Chilean healthcare system because private-sector hospitalizations are not captured. This limitation should be emphasized.

Response: We have clarified that the GRD-FONASA database represents hospitalizations financed by the Chilean public health insurance system and does not include the private healthcare sector. Accordingly, we have revised the terminology throughout the manuscript to avoid implying national representativeness of the hospitalization data and have explicitly acknowledged this limitation.

The substantial reduction in hospitalizations during 2020–2021 should be examined in relation to the COVID-19 pandemic and changes in healthcare utilization.

A sensitivity analysis excluding or separately analysing the pandemic years would help determine whether the main findings are robust.

Response (n°20 and n°21): We addressed both comments by examining the reduction in hospitalizations during the COVID-19 period and performing a sensitivity analysis excluding 2020–2021. Hospitalizations were 37.6% lower in 2020 and 40.8% lower in 2021 compared with 2019. After excluding these years, the associations with in-hospital mortality persisted for pneumonia (OR 1.81; 95% CI 1.52–2.16) and acute respiratory complications (OR 1.97; 95% CI 1.65–2.36), although the point estimates were attenuated. The Results and Discussion have been revised accordingly, with the decline in hospitalizations interpreted cautiously as potentially reflecting pandemic-related changes in healthcare utilization.

The manuscript should provide more detail on missing data and the assumption that absence of an ICD-10 code represents absence of the condition.

Response: We clarified the handling of missing data in the Methods (Missing Data section). Specifically, the absence of a corresponding diagnostic code was interpreted as the condition not being recorded for that hospitalization episode and was therefore not considered missing data. We also clarified that records with sex classified as “unknown” were excluded from the analyses.

Please clarify how the hospitalization diagnostic categories were defined and provide a complete ICD-10 code mapping.

Response: We have added a new supplementary table (Table S1) providing the complete ICD-10 mapping used to define the diagnostic categories included in the hospitalization analyses. We also added a reference to this table in the Methods to facilitate the identification and reproducibility of the diagnostic definitions used in the study.

Minor Comments

The title uses “Hospital Outcomes,” although the principal hospital outcome is in-hospital mortality. Consider more precise wording.

Response: We thank the reviewer for this helpful suggestion. We agree that emphasizing respiratory infections in the title may not fully reflect the broader scope of the study. We have therefore revised the title to “Lung Cancer Mortality and In-Hospital Outcomes in Chile: A Population-Based Cross-Sectional Study.” This title better reflects the two main components of the study: population-level lung cancer mortality trends and the analysis of in-hospital outcomes.

Use “sex” consistently rather than alternating between sex and gender.

Response: We have revised the manuscript to ensure consistent terminology throughout. “Gender” was replaced with “sex” where applicable, and terms such as “sex-specific,” “male,” and “female” are now used consistently to reflect the variable available in the administrative databases.

Clearly distinguish population-level mortality from mortality among FONASA hospitalization episodes in the abstract.

Response: We revised the Abstract to clearly distinguish population-level mortality from in-hospital mortality. The mortality trend analysis is now explicitly described as based on national DEIS mortality records, whereas in-hospital mortality is identified as referring to hospitalization episodes recorded in the GRD-FONASA database.

Provide exact numerators and denominators when reporting percentages, particularly for the approximately 75% cancer/metastasis category.

Response: Thank you for this suggestion. We revised the manuscript to provide exact numerators and denominators alongside the relevant percentages and clarified the population to which these estimates refer. These changes have been incorporated throughout the Abstract, Results, and Discussion, where applicable.

Report the number of deaths within each diagnostic category in addition to percentages.

Response: Thank you for this suggestion. We added the number of in-hospital deaths and the corresponding number of hospitalizations for each diagnostic category, overall and by sex, in the Supplementary Materials. These counts complement the reported percentages and clarify their denominators.

Provide confidence intervals for mortality rates and temporal estimates where appropriate.

Response: Thank you for this suggestion. We have added 95% confidence intervals for mortality rates and temporal trend estimates where appropriate. The Results and corresponding figure descriptions have been updated accordingly.

Expand the statistical methods to specify the software version, significance threshold, variable selection, treatment of age, model diagnostics, and handling of sparse data or separation.

Response: Thank you for this suggestion. We expanded the Statistical Analysis section to specify the R version, two-sided significance threshold (P < 0.05), model covariates, absence of automated variable selection, and treatment of age as a continuous linear term. We also described the checks performed during revision, including model convergence, model-matrix rank, and outcome counts by diagnostic category. We clarified the handling of separation in women with Z51 and sparse events in men with Z51, and distinguished the confidence-interval methods used in the original models from the patient-clustered robust standard errors used in the supplementary sensitivity analyses.

Emphasize effect estimates and confidence intervals rather than statistical significance alone, particularly given the large sample size.

Response: We revised the Results to emphasize the magnitude and precision of the associations by focusing the interpretation on effect estimates and 95% confidence intervals rather than statistical significance alone. P values were retained as complementary statistical information.

Explanations offered for increasing mortality among women, including smoking, environmental exposures, screening, and treatment access, were not directly measured. These should be presented as possible explanations rather than study findings. Overall reference is very less. can do extensive discussion by using recent literatures

Response: We revised the Discussion to present smoking, environmental exposures, screening, and treatment access as possible explanations rather than study findings. We also expanded the discussion with recent literature to provide broader context for the observed sex-specific mortality trends.

The conclusion should be more cautious and limited to the associations supported by the available administrative data.

Response: We revised the Conclusion to focus on the mortality patterns and associations directly supported by the administrative data and removed broader implications extending beyond the scope of the study.

Revisions Before Publication

Resolve the inconsistent readmission estimates across the manuscript and supplementary tables.

Response: We verified the readmission analyses and confirmed that first hospitalization was consistently used as the reference category. The apparently different estimates reflect different model specifications. We also corrected an inverted comparison in the text and clarified the adjustment variables and reference categories in the Methods and supplementary tables.

Reanalyse the hospitalization data while accounting for repeated admissions within individuals.

Response:

We reanalyzed the hospitalization data accounting for repeated admissions by using patient-clustered robust standard errors. This approach accounts for the within-patient correlation arising from multiple hospitalization episodes contributed by the same individual.

Correct and investigate the COPD estimate of OR = 0.00.

Response: We appreciate the reviewer’s observation. This issue has been resolved by correcting the previously reported COPD estimate

Perform a formal interaction analysis before claiming sex-based effect modification.

Response: We appreciate the reviewer’s observation. This issue has been addressed by performing a formal interaction test between sex and hospitalization diagnosis category. The test did not provide evidence of effect modification by sex. Accordingly, the interpretation of the sex-stratified analyses was revised to avoid inferring effect modification from differences in point estimates alone.

Strengthen the mortality trend analysis using an appropriate temporal method.

Response: We addressed this recommendation by strengthening the mortality trend analysis with a formal log-linear APC analysis with 95% confidence intervals for 2002–2024. We also incorporated sex-specific mortality projections through 2035 using the Nordpred age-period-cohort model.

Clarify the representativeness and denominator of the FONASA hospitalization population.

Response: We appreciate the reviewer’s observation. We clarified that the hospitalization analysis is based on episodes recorded in the GRD-FONASA database and therefore represents hospitalizations among FONASA beneficiaries rather than all lung cancer hospitalizations in Chile. FONASA covers approximately 78% of the Chilean population, providing broad population coverage, although the private healthcare sector is not represented in this database. No population-based hospitalization rates were calculated from the FONASA data; therefore, a population denominator was not used for this analysis. This limitation is now explicitly acknowledged in the manuscript.

Address residual confounding related to tumor stage and disease severity.

Response: We appreciate this comment. We now explicitly acknowledge residual confounding due to unavailable measures of tumor stage and disease severity, including performance status, metastatic burden, treatment, and ICU admission.

Moderate causal and preventive interpretations, particularly regarding pneumonia and vaccination.

Response: We appreciate this suggestion. Causal interpretations regarding pneumonia were moderated, and vaccination is now discussed cautiously as a potential clinical consideration rather than a preventive effect demonstrated by our findings.

Clarify the use of ICD-10 C33–C34 for defining lung cancer.

Response: Thank you for noting this point. We clarified the ICD-10 definition used for the hospitalization cohort. Lung cancer was identified using C34, and the previous reference to C33–C34 was corrected accordingly.

Address the potential influence of the COVID-19 period through discussion and, preferably, sensitivity analysis.

Response: We appreciate this observation. We addressed the potential influence of the COVID-19 period by conducting a sensitivity analysis excluding 2020–2021. The associations for pneumonia and acute respiratory complications remained consistent, supporting the robustness of the main findings.

Reviewer 3 Report

Comments and Suggestions for Authors

Major

In the most of the countieries, lung cancer is the most prevalent cause of cancer-related death（The Current Roadmap of Lung Cancer Biology, Genomics and Racial Disparity. Int J Mol Sci. 2025 Apr 17;26(8):3818.）. There are two large histological groups of lung cancer: small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC). Recently women's lung cancer mortality rates have dramatically increased in developed countries（BMC Res Notes. 2015 Oct 3;8:530.）However, population-level evidence on long-term mortality trends and hospital outcomes in Latin America remains limited.

Therefore the current study which analyzes long-term mortality trends from LC in Chile and characterize hospitalization patterns and factors associated with in-hospital mortality, may be very important for the understanding of LC epidemiology and pathology.

Then, thy found that LC mortality trends in Chile diverged by sex, with declining age-adjusted mortality among men but increasing mortality among women. The trend of Lc mortality is similar to the trends of most of the developed counties.

They also found that respiratory infections and acute respiratory complications were strongly associated with in-hospital mortality of the patients with LC.

We agree with the authors that the prevention and clinical management of respiratory infections and acute respiratory complications is the key targets for improvement of the mortality in the patients with LC.

However, we do not think the recommendation is the core of the paper.

The title of the apper may be misunderstood by the reads of the Journal.

“Respiratory Infections, Hospital Outcomes, and Lung Cancer Mortality in Chile: A Population-Based Cross-Sectional Study” should be changed to another tile of the paper. For example, Hospital Outcomes, and Lung Cancer Mortality in Chile: A Population-Based Cross-Sectional Study-the factors for the mortaklity by the management of respiratory Infections, -

Author Response

We sincerely thank the reviewer for the thoughtful and positive assessment of our study and for recognizing the relevance of examining long-term lung cancer mortality trends and hospital outcomes in Chile. We also appreciate the reviewer’s comments regarding the observed sex-specific mortality trends and the findings related to respiratory complications. We agree that emphasizing respiratory infections in the title could give disproportionate weight to one specific finding and may not fully represent the broader scope of the study. Accordingly, we revised the title to “Lung Cancer Mortality and In-Hospital Outcomes in Chile: A Population-Based Cross-Sectional Study.” We believe that the revised title more accurately reflects the two main components of the study: population-level lung cancer mortality trends and the analysis of in-hospital outcomes.