Identification of Tobacco-Related Cancer Diagnoses among Individuals with Psychiatric Disorders: A Population-Based Matched Cohort Study Using a Competing Risks Approach from British Columbia
Round 1
Reviewer 1 Report
Identification of tobacco-related cancer diagnoses among individuals with psychiatric disorders: a population based, matched cohort study using a competing-risks approach from British Columbia
The original article by Olson and colleagues aimed to analyze whether psychiatric is related to tobacco use and the further diagnosis of several cancers. The authors recruited and normalized a large number of clinical cases in special areas. They calculated various features based on these parameters. They further claimed that individuals with psychiatric disorders have no significant value in diagnosing tobacco-related cancers. However, this manuscript requires additional clarifications, verification and comparison. This article currently lacks rigorous description. The following issues must be addressed:
Major comments:
- “Tobacco-related cancer” is rather vague. The author should carefully classify and analyze them separately. In the situation, heterogeneous cancer characteristics may offset potentially significant differences between psychiatric disorders and cancer features. For the example, lung adenocarcinoma and lung squamous carcinoma exhibit different signatures and consequence. There may be new discoveries after careful classification.
- In Table 2 and table 3, I guess they did univariate and multivariate analysis in their cohort, respectively. Strangely, these symptoms are significantly associated with poor prognosis (high risk). But after comparison, it showed the opposite results. They should discuss possible causes.
- Can another validation cohort or previous references support similar results?
- In addition to events related to psychiatric disorders, TR cancer subtypes and cancer-related parameters should also be calculated. (ex. Pulmonary function, smoking history, total vital capacity…etc.)
- What is the “total person-years of follow-up” in Table1? And how is it calculated?
Minor comments:
- The quality of figure 1 is poor, the pixels should be improved. The number of cases corresponding to each group should be marked.
Author Response
Thank you for the review.
Before their numbered items, the reviewers claim we stated "individuals with psychiatric disorders have no significant value in diagnosing tobacco-related cancers". We are confused by this statement, as we never stated that anywhere in the manuscript, and I believe the reviewer missed some of the discussion and conclusion statements.
1) This is not feasible in this very large, population based cohort of over 160,000 individuals. The focus of this manuscript is on the high-level as the first investigation of this association. We agree that sub-site analysis in a lung specific cohort is a valid goal of future research, and added this to the conclusion
2) The reviewer is misinterpreting the results. They are two totally different concepts. Table 2 is looking at the survival rates of patients, as a step to determine if we needed to do a competing risk analysis, which we did. Table 3 is the hazard ratio of developing cancer (not dying).
3) This is a great idea for future research and we have added to text
4) In this cohort of over 160,000 patients this data is not feasible to obtain through chart review; furthermore the charts are not available through this PopDataBC data pull
5) Person years are calculated as the sum of all subject's years of follow-up
A new version of Figure 1 is now included in the revision file
Reviewer 2 Report
This large, retrospective, population-based cohort study has identified a novel decreased hazard of being diagnosed with a TR cancer in individuals with depression, anxiety disorders or multiple PD diagnoses
These results were unexpected, and we hypothesize could be explained by individuals with a PD having increased barriers to a cancer diagnosis, rather a true decreased incidence
Most concerning, our limitation of the analysis to TR cancers which are usually clinically meaningful such as lung and esophageal cancer, suggests that PD stigma may prevent these patients from receiving diagnosis and care for these cancers.
Further research is needed to assess this hypothesis or look for biologically plausible reasons why PD patients may have a true decreased incidence of TR cancer
Author Response
Thank you for your positive review
Round 2
Reviewer 1 Report
The authors replied to all questions and I have no further concerns about this manuscript.