Associations Between Cancer-Related Fatigue and Healthcare Use During Cancer Follow-Up Care: A Survey-Administrative Health Data Linkage Study
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsAuthors present an interesting paper: " Associations between cancer-related fatigue and healthcare use during cancer follow-up care: a survey-administrative health data linkage study" . This study sought to examine how cancer-related fatigue impacts healthcare use after completing cancer treatment. The different sections of the manuscript are well documented and I don'have any criticism. In the Conclusions the Authors should more clearly express what they suggest based on this study to improve the risk of cancer related fatigue. Other minor suggestion is to mention that also "resilience" exists in many survivors as positive expression of being ill from cancer.
Author Response
Comment 1: The different sections of the manuscript are well documented and I don't have any criticism. Response: We thank the Reviewer for their review of our paper.
Comment 2: In the Conclusions the Authors should more clearly express what they suggest based on this study to improve the risk of cancer related fatigue. Response: Point well taken. We have now added the following sentence to the conclusions section:
This includes counselling around or referral to exercise programs, cognitive behavioral therapy, and mindfulness-based programs [22].
Comment 3: Other minor suggestion is to mention that also "resilience" exists in many survivors as positive expression of being ill from cancer. Response: We agree with the Reviewer. However, this manuscript focused on cancer-related fatigue and healthcare utilization, and we are uncertain how to incorporate psychological resilience and positive growth in the context of the current manuscript. If the Reviewer has suggestions for this, we are happy to hear them.
Reviewer 2 Report
Comments and Suggestions for AuthorsThe manuscript provides an interesting insight on a subject that is not frequent. Your manuscript should provide larger detail on some few parts:
1) why the timeframes in table 1 are different? May these differences impact on the results?
2) Larger detail on the codification of ICD (ICD-O-3 is on the 9th version)
3) Why did you assume two leverrs on the chronic conditions and for cancer related fatigues?
4) please state all the others cancer (put a simple list in the text)
5) are data normally distributed?
6) the cancer stage should be associate to TNM, can you provide the association with stage I, II and III?
7) the models should include the adjusted R squared and more detail
Author Response
Comment: Why the timeframes in table 1 are different? May these differences impact on the results?
Response: This is a great question. The timeframes were different because the periods of interest were selected at a national level, and dependent on both the 1) expected duration of cancer treatment (e.g., those with ALL have an extended period of treatment that often lasts several years) and 2) ensuring consistency in data availability and completeness in all 10 participating provinces. However, these differences, while they exist are relatively minor, should not have impacted the results given no changes in follow-up care policies or practices occurred between timeframes.
Comment: Larger detail on the codification of ICD (ICD-O-3 is on the 9th version)
Response: This is a fair question given the revisions in ICD coding over the years. The ICD-O-3 version relevant at the time of diagnosis is what would have been used by coders to classify the disease of interest in the cancer registry. For our time periods of interest, this would have been the first or second versions. If further information is required, please let us know and we are happy to provide it.
Comment: Why did you assume two levels on the chronic conditions and for cancer related fatigues?
Response: Great questions. We chose two levels for the chronic condition variable based on simplicity. For cancer-related fatigue, we considered meaningfulness of the concern from a patient perspective, wherein moderate/big cancer-related fatigue likely has some impact on patient experience and functioning whereas no/low fatigue likely has no to minimal impact on patient experience and functioning.
Comment: Please state all the others cancer (put a simple list in the text)
Response: We appreciate this question. The “other” cancers are outlined in the methods section: individuals aged 18-29 at diagnosis with a primary diagnosis of any invasive cancer (behaviour code=3), excluding: 1) stage IV at diagnosis (except for testicular cancer); 2) non-melanoma skin cancer; and 3) Kaposi's sarcoma. Nonetheless, we have now included this as a footnote to Table 2 (Cohort characteristics) to clarify what “other” means.
Comment: Are data normally distributed?
Response: Thank you for this question. The probability of discharge to primary care (Table 3 results) is a logistic regression so the distribution is logistic. The Poisson and negative binomial models have a Poisson distribution and negative binomial distribution, respectively. Each of these are skewed to the right. The negative binomial has over-inflated number of zeros.
Comment: The cancer stage should be associated to TNM, can you provide the association with stage I, II and III?
Response: We agree it is important to understand how stage is categorized. Stage at diagnosis was captured from the Nova Scotia Cancer Registry and was based on the Collaborative Stage system, a unified data collection system used by North American cancer registries to derive TNM stage. Blood cancers were based on the staging system identified in Table 1. We have now clarified this in the revised manuscript (page 2 of revised manuscript):
Stage at diagnosis for solid tumor cancers was based on the Collaborative Stage system, a unified data collection system combining clinical and pathologic data to derive a tumor (T), node (N), and metastasis (M) stage.
Comment: The models should include the adjusted R squared and more detail.
Response: We appreciate this comment. Due to the regression modelling we employed, none have an R squared calculated (unlike an approach such as the ordinary least squares method). Instead, we could report the log likelihood or AIC value. We are not certain this is typical for the journal (based on a brief review of papers published in this journal) but please let us know if the Reviewer or Editor feels differently and we can add these values.
Reviewer 3 Report
Comments and Suggestions for AuthorsThis is an interesting study which addresses important issues. However, it is unclear to me why a validated fatigue questionnaire - of which there are many - was not used.
This undermines the findings to the point at which I would need to recommend rejection.
Author Response
Comment: It is unclear to me why a validated fatigue questionnaire - of which there are many - was not used.
Response: This is an important question. There are indeed validated fatigue questionnaires. However, this national survey was developed to identify and understand the multitude of cancer survivors’ needs post-treatment. Although the survey was informed by validated surveys to assess unmet needs holistically, it did not employ validated measures to assess each individual need (e.g., fatigue, pain, insomnia, cognitive impairment, fear of recurrence, depression, anxiety, etc). While the use of validated measures would have been ideal, it would have also created an enormous burden on respondents who were asked about 20 specific needs post-treatment. Nonetheless, we feel these data are still valuable and represent respondents’ perceptions of their needs at the time of survey completion. We hope the Reviewer feels similarly. We did, however, expand on this limitation in our Discussion section, adding the following statement:
While the use of a validated measure would have been ideal, the survey asked respondents about 20 specific needs post-treatment (e.g., insomnia, cognitive function, sexual dysfunction, fear of recurrence, depression, anxiety, among others); the use of validated measures to assess each of these needs would have greatly increased the length of the survey for respondents.
Round 2
Reviewer 1 Report
Comments and Suggestions for AuthorsThe Authors followed carefully the minor suggestions moved by me. The concept of resilience was only to let you in mind but there is no role for its presence in this paper.
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors included the modifications I required in the previous review process. I am satisfy with it.
Reviewer 3 Report
Comments and Suggestions for AuthorsThank you for the response. Unfortunately, you are not able to address the fundamental issue with my review and my judgement remains the same.
