Review Reports - Post-Traumatic Growth and Quality of Life Among World Trade Center Health Registry Enrollees 16 Years After 9/11

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Overall, this is a well presented, lovely analysis pushing forward the study of post trauma growth in the WTC-exposed cohort.

I have only a few thoughts and suggestions.

The phrase 'mental functioning' is used. If this phrase is used in the SF-12, please continue to use it. However, it does not really ring true to my ear. "Physical functioning' makes sense, but mental functioning does not. Perhaps using the phrase 'mental health' by itself without the word 'functioning?' "Physical functioning and mental health," perhaps? Psychiatric health could be used, or behavioral health? But following the lead of the instrument authors is the best way to go. The way things are stated in line 280 works well.

Line 39. - If you are going to talk about the literature on the impact of trauma in the WTC cohort, it is fine to start with this sentence and use the references used. However, it seems here that you are stating that some of the seminal work on the impact of trauma overall has been done by the registry. It would be good to use some other non WTC references here as the history of trauma research goes way back. If you choose to start with the WTC event, be sure to specify that.

Line 41; phrasing is somewhat confusing for those familiar with the trauma research. when speaking of variations in how individuals respond to trauma, the first thing that comes to mind is the well published risk factors for development of extended post-trauma sx. But here, you are referring to those that respond with growth, vs. not-growth. This should be clarified.

Line 46 - 'as would be the case with resilience" - this needs to have references if you are talking about resilience as a concept, vs, growth.

Line 47 - the 'phenomenon' was not developed by these two researchers, rather the label of the concept, or the instrument that captures post trauma growth.

Line 280 - It seems that PTG association with mental health and not physical health is a result that warrants more discussion.

Another significant limitation - level of exposure is not a variable in the study.

Author Response

Response to Reviewer 1 for “Posttraumatic Growth and Quality of Life Among World Trade Center Health Registry Enrollees 16 Years After 9/11”

Overall, this is a well presented, lovely analysis pushing forward the study of post trauma growth in the WTC-exposed cohort.

Comments:

The phrase 'mental functioning' is used. If this phrase is used in the SF-12, please continue to use it. However, it does not really ring true to my ear. "Physical functioning' makes sense, but mental functioning does not. Perhaps using the phrase 'mental health' by itself without the word 'functioning?' "Physical functioning and mental health," perhaps? Psychiatric health could be used, or behavioral health? But following the lead of the instrument authors is the best way to go. The way things are stated in line 280 works well.

Response: We substituted “mental function” for “mental functioning” throughout the paper.

Line 41; phrasing is somewhat confusing for those familiar with the trauma research. when speaking of variations in how individuals respond to trauma, the first thing that comes to mind is the well published risk factors for development of extended post-trauma sx. But here, you are referring to those that respond with growth, vs. not-growth. This should be clarified.

Response: We have rewritten this paragraph to emphasize how previous trauma studies differed from recent studies emphasizing Post Traumatic Growth.

Line 39. - If you are going to talk about the literature on the impact of trauma in the WTC cohort, it is fine to start with this sentence and use the references used. However, it seems here that you are stating that some of the seminal work on the impact of trauma overall has been done by the registry. It would be good to use some other non WTC references here as the history of trauma research goes way back. If you choose to start with the WTC event, be sure to specify that.

Response: We rewrote this sentence to emphasize that interest here centered on the 9/11 attacks and the role of Registry research in documenting the effects of that exposure .

Line 47 - the 'phenomenon' was not developed by these two researchers, rather the label of the concept, or the instrument that captures post trauma growth.

Response: We have corrected this mistake.

5.. Line 280 - It seems that PTG association with mental health and not physical health is a result that warrants more discussion.

Response: The present study employed the standard PTGI instrument, which focuses on the psychosocial components but not the corporeal components of trauma, so this instrument would be more likely to associate with altered mental than physical function.

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for the opportunity to evaluate this manuscript. Understanding the association between post traumatic growth and other health outcomes is an interesting subject. This study found that PTG was associated with greater mental health quality of life. I understand this this study has been redone based on the redacted version by the same name, that said, I do have some concerns/comments/questions that should be addressed before this study can be considered finalized. I’ve outlined them below in order of appearance in the manuscript – all are major unless noted.

Abstract – this still lists the 4760 study population. I believe it is now 2786, right? Or are these two populations both included?
In the introduction please also comment on how depression might be associated (and therefore confound) the association between PTG and HQoL. It seems their screening instruments might both be influenced by depressive symptoms.
Participants: how many people were excluded because they didn’t take all 4 surveys and were age 18+ etc (line 110)?
Participants: you state the included were the same as excluded is this excluded from group 1 and group 2 or the final 2786 vs the rest of the WTCHR population? Please add the Ns here as well so this is clear. If it’s the not the latter, please also include how these people differ from the study. For example how many people with an injury were excluded because they didn’t complete all necessary waves? You might have a survivor bias.
In the redacted version, you said “For both the physical and mental summary scores, higher values correspond to better, healthier functioning” in regards to the SF-12 why was this removed?
Minor – please include the years for each wave when describing when the covariates were measured. For those of us who do not know when the different waves took place having the exact date range there would be helpful and mean we don’t have to flip to that section.
Are both responders and survivors included in this study? If they are, please include that as a covariate or explain why that is not needed. It seems those with first responder training would have a different PTG outcome than an office worker. These are mentioned as “eligibility group” but it does not seem like this variable was included in the models based on the table footnote and the stats section.
Why are you not using the mental health screener scores instead of a report of experiencing ptsd or depression? It seems like one might endorse symptoms more easily than say they had a diagnosis due to the stigma around mental health.
Minor- It seems you should include the prevalence of PTG in the first section of the results as it is the main “exposure” of interest.
Minor – can you format table 1 in a clearer way? It is hard to tell what each of the columns are since some are % and some are scores and since it spans two pages.
Please make sure the results in table 2 are clear. I’m not sure what the top section (first 4 rows) of results are since the only results reported in the text are from the last 4 rows. The header says “sample between chambers and canal” but from line 110 I thought everyone was below chambers – from a map it looks like canal is above (north). Please clarify this table as does not stand alone in its current form.
Minor – the title says HQoL and the manuscript starts off using this term but by section 3.4 and the discussion it’s “physical and mental functioning” is there a reason for this change? It might be helpful to remind the reader that you are talking about the same thing. Additionally, in reading (really, skimming) the manual for the SF-12 it does not seem to use the terms physical functioning or mental functioning. Is this a new concept that is being applied to the SF-12?
Minor – it seems there’s a hyphen that is not needed for “primary” on like 294
Minor – in the limitations it is mentioned that gender is treated as a binary variable. This seems like the number of nonbinary participants would be small given the size of this population. Would this meaningfully bias the results? If so you can include this but it seems there are a number of other more impactful ways your results could be biased such as selection bias from the main cohort or other uncontrolled confounders.
You state that a limitation of the study is that the SF-12 was self-reported. Is there another way for the SF-12 to be administered?

Author Response

Response to Reviewer 2 for “Posttraumatic Growth and Quality of Life Among World Trade Center Health Registry Enrollees 16 Years After 9/11”

Comments and Suggestions for Authors

Abstract – this still lists the 4760 study population. I believe it is now 2786, right? Or are these two populations both included?

Response: We have corrected this error.

In the introduction please also comment on how depression might be associated (and therefore confound) the association between PTG and HQoL. It seems their screening instruments might both be influenced by depressive symptoms.

Response: Since PTG is associated with PTSD, and the latter is associated with depression, it is likely that PTG is associated with depression. Since depression is associated with HQoL, depression could be a confounder of the PTG-HQoL association. Specifically, depression could weaken the PTG-HQoL association. This was included in the introduction.

Participants: how many people were excluded because they didn’t take all 4 surveys and were age 18+ etc (line 110)?

Response: The original Registry population consisted of 71,423 enrollees. The injury study first included only all enrollees who completed all of Waves 1-4 and who were 18 years or older on 9/11 (n~36,000). So approximately 35,000 enrollees were excluded due to the above constraint.

Participants: you state the included were the same as excluded is this excluded from group 1 and group 2 or the final 2786 vs the rest of the WTCHR population? Please add the Ns here as well so this is clear. If it’s the not the latter, please also include how these people differ from the study. For example how many people with an injury were excluded because they didn’t complete all necessary waves? You might have a survivor bias.

Response: There were no differences between enrollees included in the study (n=2,786) and the WTCHR population (N=71,423).

In the redacted version, you said “For both the physical and mental summary scores, higher values correspond to better, healthier functioning” in regards to the SF-12 why was this removed?

Response: We have reinserted this text into the paper.

Minor – please include the years for each wave when describing when the covariates were measured. For those of us who do not know when the different waves took place having the exact date range there would be helpful and mean we don’t have to flip to that section.

Response: We have reinserted years for each wave into the covariates section.

Are both responders and survivors included in this study? If they are, please include that as a covariate or explain why that is not needed. It seems those with first responder training would have a different PTG outcome than an office worker. These are mentioned as “eligibility group” but it does not seem like this variable was included in the models based on the table footnote and the stats section.

Response: Yes, both responders and survivors were included in our sample. We reran the original multiple linear regressions for the physical and mental scores, including a yes/no variable for enrollee type as a control. The regression coefficient for PTG changed very little from its original value for both physical and mental function.

Why are you not using the mental health screener scores instead of a report of experiencing ptsd or depression? It seems like one might endorse symptoms more easily than say they had a diagnosis due to the stigma around mental health.

Response: The mental health scales were not available for all waves 1-4, so we had to use reports of experiencing PTSD, depression, or anxiety to represent having mental health disease prior to the I jury study.

Minor – can you format table 1 in a clearer way? It is hard to tell what each of the columns are since some are % and some are scores and since it spans two pages.

Response: We have added text to clarify the meaning of the columns.

Please make sure the results in table 2 are clear. I’m not sure what the top section (first 4 rows) of results are since the only results reported in the text are from the last 4 rows. The header says “sample between chambers and canal” but from line 110 I thought everyone was below chambers – from a map it looks like canal is above (north). Please clarify this table as does not stand alone in its current form.

Response: We have added subsection titles to Table 2. We included results between Canal and Chambers streets because reviewers of the previous version of this paper requested it.

Minor – the title says HQoL and the manuscript starts off using this term but by section 3.4 and the discussion it’s “physical and mental functioning” is there a reason for this change? It might be helpful to remind the reader that you are talking about the same thing. Additionally, in reading (really, skimming) the manual for the SF-12 it does not seem to use the terms physical functioning or mental functioning. Is this a new concept that is being applied to the SF-12?

Response: HQoL as measured by the SF-12 has separate physical and mental components, so when we discussed results we had to mention physical and mental separately. But the start of the paper mentioned HQoL generally, so this was not necessary.

Minor – in the limitations it is mentioned that gender is treated as a binary variable. This seems like the number of nonbinary participants would be small given the size of this population. Would this meaningfully bias the results? If so you can include this but it seems there are a number of other more impactful ways your results could be biased such as selection bias from the main cohort or other uncontrolled confounders.

Response: You are correct that the number of non-binary enrollees is sufficiently small as to not bias the results. We have removed this text from the limitations section.

You state that a limitation of the study is that the SF-12 was self-reported. Is there another way for the SF-12 to be administered?

Response: You are correct that the SF-12 can only be self-reported. We have removed it’s mention from the relevant sentence.

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript is an important and useful study looking at long term psychological outcomes following the 9/11 attacks in an exceptionally large study population. By leveraging data from the World Trade Center Health Registry, the authors were able to look at the relations between post-traumatic growth and quality of life more than 15 years after the attacks. Leveraging a long-term, well-characterized cohort, is a particular strengths as it provides understanding of recovery and adaptation. The large sample and adequately powered analyses including (a) thorough data collection and (b) use of validated measures adds to the credibility and the clear public health relevance of this study.

I have organized my comments below by sections rather than by priority.

Title and abstract

The title is descriptive and the abstract provides a good overview of the study and the results. It could be improved by providing the final analytic sample in the abstract and by providing an indication of which covariates were included in the regression models. The conclusion could also be more specific with respect to future research directions.

Introduction

The introduction provides a thorough and well-structured overview of the literature on post-traumatic growth and health-related quality of life, and it successfully places the current study within the broader context of 9/11-related research. The rationale for investigating PTG in relation to both physical and mental health outcomes is clearly stated, and the justification for distinguishing between injured and non-injured participants is logical. That being said, that section could be tighter and more focused. There are several places where it reads like a short review, with long stretches of background and many citations, that could be summarized more sparsely. Tightening that section would help keep focus on the particular research gap the study is addressing. Further, the authors note that there is incongruent evidence in the literature about PTG and health-related quality of life, but do not clearly state what the particular incongruence or hypothesis that the study is testing. A clearer statement of the conceptual model or expected direction of associations in the reader’s mind would help understand the study’s contribution.

Methods

The methods section is clear and detailed, and benefits from being so fairly grounded in the “gold standard” procedures of the World Trade Center Health Registry. The description of inclusion and exclusion criteria is fairly transparent, and they use well-established instruments (the SF-12 and the Post-Traumatic Growth Inventory) which supports the study’s methodological foundations. The authors also include a fairly good overview of the covariates, including sociodemographic, psychosocial, and health-related factors. However, there are several areas that could be clarified or better justified:

Cross-sectional analytic design: Though the Registry does have longitudinal data available from four prior survey waves, the analysis here is cross-sectional, and only includes data from the 2017–2018 Health and Quality of Life survey. It would strengthen the paper if the authors explained why earlier measures (for example, prior PTSD or depression assessments) were not used to model trajectories or temporal associations. The current design does not allow for causal interpretation.
Sampling and response: This number of eligible patients approached to complete the HQoL survey and response (or completion) rate are not presented in the manuscript. This would allow the readers to evaluate the possible bias in the selection of the analytic sample.
Data collection mode: The authors describe the survey as self‑administered but do not indicate what mode was used (online, mailed questionnaire, telephone). The mode should be described in more detail.
Missing data: Patients with missing data were excluded, but the level of missingness and whether sensitivity analyses were performed to assess for missing data bias is not described. Even descriptive information on missing data percentages would have added to the methods section.
PTG definition: A cutoff of mean ≥ 3 to define “moderate-to-high post-traumatic growth” seems logical and similar to previous analyses, but a citation (or brief explanation) on why this threshold was chosen would be helpful.

Results

The results are well described and logically organized. The tables are well organized and provide enough information to understand how the variables relate to both physical and mental health outcomes. The authors report their main findings transparently, and the consistent association between post-traumatic growth and mental, but not physical, functioning is easy to follow. The distinction between injured and non-injured participants is a valuable addition and adds depth to the interpretation.

The results section includes a lot of numbers but offers very little narrative explanation. It would help readers if the authors highlighted key contrasts in plain language, for example, summarizing what the beta values mean in practical terms rather than relying only on statistics.

The tables could also be simplified. Table 1, in particular, contains a great deal of detail that might be more appropriate for supplementary material. A brief summary of the most relevant demographic characteristics in the text would make the section easier to read.

Finally, while the authors refer to the clinical significance threshold for the SF-12 scores, they could do a better job of explaining why the observed difference in mental functioning (around four points) is meaningful in this population. Adding a short interpretation in that context would make the findings more accessible to readers outside epidemiology.

In addition, the paper would benefit from some information on the psychometric performance of the instruments used. Although both the SF-12 and the Post-Traumatic Growth Inventory are well-validated tools, reporting internal consistency coefficients (for example, Cronbach’s alpha) for this specific sample would add credibility and also reflect the quality of the data collected. Such indicators would reassure readers that the measures functioned reliably in a 9/11-exposed population after sixteen years.

Overall, the results are strong and consistent, but the section would benefit from a bit more explanation in plain terms and a slightly lighter presentation of the data.

Discussion

The discussion section does a solid job of summarizing the key findings and placing them in relation to existing research on post-traumatic growth. The authors clearly state that PTG was linked to better mental health functioning but not to physical functioning, and they situate this result within the broader literature on trauma recovery.

The comparison with prior work, both studies with 9/11 survivors and other survivor groups (e.g., veterans, cancer survivors, is relevant and helps readers understand how the current results fit into the literature. But the discussion could be stronger in a number of ways. For example, the authors note that the analysis was cross-sectional, but they do not discuss very well what that means for interpretation. If there is not longitudinal modeling the authors cannot say whether PTG contributed to better mental functioning or whether people with better mental health just happen to report more PTG. A quick acknowledgement of the bidirectionality would help strengthen the argument.

Second, the authors allude to the fact that self-reported PTG may not accurately capture psychological change, but do not investigate that possibility further. A discussion of potential mechanisms of change (e.g., positive reappraisal, coping style, response bias) would enrich the discussion psychologically and temper the conclusions.

Third, the discussion of limitations could be more informative for measurement. The well-established measures used in the study are not critically examined in terms of their reliability in this population. A discussion of lack of psychometric evaluation and the impact on interpretation of study findings would show an understanding of methodological issues. The Post-traumatic Growth Inventory is a well-validated measure, but has been criticized for both conceptual and methodological reasons: it only permits measurement of positive change and never accounts for the possibility of post-traumatic decline or stasis. In that sense, it measures perceived, but not objectively verifiable, growth. Taking this into account would have been a more sophisticated contribution.

Finally, the authors claim PTG has “implications for mental quality of life,” an overly broad statement. It would enhance the paper's contribution to propose a specific direction for future research, such as testing interventions to facilitate the development of PTG, or test whether PTG predicts better mental health over time using Registry longitudinal data.

On the whole the discussion is logically organized and appropriate but could use more critical scrutiny and a more thorough treatment of the limitations and interpretive ambiguities.

Author Response

Response to Reviewer 3 for “Posttraumatic Growth and Quality of Life Among World Trade Center Health Registry Enrollees 16 Years After 9/11”

The title is descriptive, and the abstract provides a good overview of the study and the results. It could be improved by providing the final analytic sample in the abstract and by providing an indication of which covariates were included in the regression models. The conclusion could also be more specific with respect to future research directions.

Response: We have added the study sample size and list of covariates used in the linear regressions into the abstract. We have also modified the conclusion by specifying more concrete directions for future research.

Response: We have trimmed and focused the introduction, and stated the explicit hypothesis that this study was designed to investigate.

Cross-sectional analytic design: Though the Registry does have longitudinal data available from four prior survey waves, the analysis here is cross-sectional, and only includes data from the 2017–2018 Health and Quality of Life survey. It would strengthen the paper if the authors explained why earlier measures (for example, prior PTSD or depression assessments) were not used to model trajectories or temporal associations. The current design does not allow for causal interpretatation.

Response: When the 2017–2018 Health and Quality of Life survey was conducted, it was the only data for which the Registry had measures of both PTG and SF-12, necessitating a cross-sectional design. The SF-12 will be available from the Registry’s current ongoing wave 6 survey, which will enable us to relate PTG from the 2017–2018 Health and Quality of Life survey to SF-12 data from 2026, allowing us to make causal interpretations of of the PTG-SF-12 assoication. We plan to investigate this in a future publication.

Sampling and response: This number of eligible patients approached to complete the HQoL survey and response (or completion) rate are not presented in the manuscript. This would allow the readers to evaluate the possible bias in the selection of the analytic sample.

Response: For the 2017–2018 Health and Quality of Life survey, recruitment information was presented in a previous Registry study. We have added the reference to this paper.

Data collection mode: The authors describe the survey as self‑administered but do not indicate what mode was used (online, mailed questionnaire, telephone). The mode should be described in more detail.

Response: Two survey modes were used: mailed questionnaire and online. This is described in the added publication mentioned above.

Missing data: Patients with missing data were excluded, but the level of missingness and whether sensitivity analyses were performed to assess for missing data bias is not described. Even descriptive information on missing data percentages would have added to the methods section.

Response: For the multivariable regression, missing data resulted in a sample size of 1,981 out of 2,786 enrollees, so 29% of the data was missing. Given this large fraction, we performed multiple imputation (m=5) under the MAR (missing at random) assumption as a sensitivity analysis. For the mental health component, the b for binary PTG changed from 3.92 (original) to 3.81 (imputed). We conclude that missing data, while affecting the standard error of b, had little effect on the b itself.

PTG definition: A cutoff of mean ≥ 3 to define “moderate-to-high post-traumatic growth” seems logical and similar to previous analyses, but a citation (or brief explanation) on why this threshold was chosen would be helpful.

Response: The basis for the PTG cutoff is described in reference 37 of this paper.

Response: We have added text to section 3.4 explaining the implications of the regression coefficients for metal and physical quality of life.

Response: Section 3.1 lists the main demographic characteristics for the study sample.

Response: We have added interpretation of the clinical threshold into the text.

Response: We calculated Cronbach’s a for the PTGI and SF-12 and described them in the text.

Response: We added this point to the discussion section.

Response: We added material related to this point in the relevant discussion section. We also noted that “One way to attempt to answer this question is to determine if self-reported PTG predicts outcomes of interest (e.g., QoL, self-reported chronic disease, hospitalizations).” and our intention to run this analysis in the future.

Third, the discussion of limitations could be more informative for measurement. The well-established measures used in the study are not critically examined in terms of their reliability in this population. A discussion of lack of psychometric evaluation and the impact on interpretation of study findings would show an understanding of methodological issues. The Post-traumatic Growth Inventory is a well-validated measure but has been criticized for both conceptual and methodological reasons: it only permits measurement of positive change and never accounts for the possibility of post-traumatic decline or stasis. In that sense, it measures perceived, but not objectively verifiable, growth. Taking this into account would have been a more sophisticated contribution.

Response: As requested above, we evaluated Cronbach’s a for PTG and the SF-12 physical and mental sub-scales and found that PTG and SF-12 have satisfactory internal reliability for the present study population. We also added material to address the reviewer’s conceptual concerns regarding the PTG construct.

Response: We have proposed future research to address issues such as the direction of causation between PTG and SF-12 and whether PTG predicts mental health conditions such as PTSD, depression, and psychological distress. The latter research project is underway.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thank you for your response however I believe not all of the prior points were fully address and to the degree that I know this research team as a whole is capable of doing. To ease in this review, I have copied the prior review and the authors responses below. My new comments are below after your responses.

Abstract – this still lists the 4760 study population. I believe it is now 2786, right? Or are these two populations both included?

Response: We have corrected this error.

Reviewer response: Thank you

In the introduction please also comment on how depression might be associated (and therefore confound) the association between PTG and HQoL. It seems their screening instruments might both be influenced by depressive symptoms.

Reviewer response: Thank you

Participants: how many people were excluded because they didn’t take all 4 surveys and were age 18+ etc (line 110)?

Reviewer response: was this added to the manuscript?

Participants: you state the included were the same as excluded is this excluded from group 1 and group 2 or the final 2786 vs the rest of the WTCHR population? Please add the Ns here as well so this is clear. If it’s the not the latter, please also include how these people differ from the study. For example how many people with an injury were excluded because they didn’t complete all necessary waves? You might have a survivor bias.

Response: There were no differences between enrollees included in the study (n=2,786) and the WTCHR population (N=71,423).

Reviewer response: was this included in the results? And on what variables did you test this?

In the redacted version, you said “For both the physical and mental summary scores, higher values correspond to better, healthier functioning” in regards to the SF-12 why was this removed?

Response: We have reinserted this text into the paper.

Reviewer response: thank you

Minor – please include the years for each wave when describing when the covariates were measured. For those of us who do not know when the different waves took place having the exact date range there would be helpful and mean we don’t have to flip to that section.

Response: We have reinserted years for each wave into the covariates section.

Reviewer response: thank you

Are both responders and survivors included in this study? If they are, please include that as a covariate or explain why that is not needed. It seems those with first responder training would have a different PTG outcome than an office worker. These are mentioned as “eligibility group” but it does not seem like this variable was included in the models based on the table footnote and the stats section.

Reviewer response: thank you, did you include it in the final model? It seems like it would be a confounder a priori.

Why are you not using the mental health screener scores instead of a report of experiencing ptsd or depression? It seems like one might endorse symptoms more easily than say they had a diagnosis due to the stigma around mental health.

Reviewer response: ok thank you.

Minor – can you format table 1 in a clearer way? It is hard to tell what each of the columns are since some are % and some are scores and since it spans two pages.

Response: We have added text to clarify the meaning of the columns.

Reviewer response: this is still confusing since it spans two pages. If it’s ok with the journal standards, please make the two columns (n and %) one column with n (%) format with the percent rounded.

Please make sure the results in table 2 are clear. I’m not sure what the top section (first 4 rows) of results are since the only results reported in the text are from the last 4 rows. The header says “sample between chambers and canal” but from line 110 I thought everyone was below chambers – from a map it looks like canal is above (north). Please clarify this table as does not stand alone in its current form.

Response: We have added subsection titles to Table 2. We included results between Canal and Chambers streets because reviewers of the previous version of this paper requested it.

Reviewer response: Was this included in the methods section? I’m not seeing this in your stats section. I understand that a reviewer asked for it but I’m still not seeing how the canal to chambers participants are in your cohort of those below chambers. Can you please explain this? Additionally, please include the breakdown of these two groups in table 1 as this seems to be an important variable as you are stratifying by it.

Minor – the title says HQoL and the manuscript starts off using this term but by section 3.4 and the discussion it’s “physical and mental functioning” is there a reason for this change? It might be helpful to remind the reader that you are talking about the same thing. Additionally, in reading (really, skimming) the manual for the SF-12 it does not seem to use the terms physical functioning or mental functioning. Is this a new concept that is being applied to the SF-12?

Reviewer response: It still seems a little confusing but that might just be me.

Minor – in the limitations it is mentioned that gender is treated as a binary variable. This seems like the number of nonbinary participants would be small given the size of this population. Would this meaningfully bias the results? If so you can include this but it seems there are a number of other more impactful ways your results could be biased such as selection bias from the main cohort or other uncontrolled confounders.

Response: You are correct that the number of non-binary enrollees is sufficiently small as to not bias the results. We have removed this text from the limitations section.

Reviewer response: can you please include the more important limitations likes uncontrolled confounders as mentioned above

You state that a limitation of the study is that the SF-12 was self-reported. Is there another way for the SF-12 to be administered?

Response: You are correct that the SF-12 can only be self-reported. We have removed it’s mention from the relevant sentence.

Reviewer response: was this removed? Please focus on what your results actually mean and how they might be biased. It seems like this limitation section is a little bit of a throwaway list.

Author Response

Response to Reviewers

Participants: how many people were excluded because they didn’t take all 4 surveys and were age 18+ etc (line 110)?	The original Registry population consisted of 71,423 enrollees. The injury study first included only all enrollees who completed all of Waves 1-4 and who were 18 years or older on 9/11 (n~36,000). So approximately 35,000 enrollees were excluded due to the above constraint. This was added to the manuscript.
Participants: you state the included were the same as excluded is this excluded from group 1 and group 2 or the final 2786 vs the rest of the WTCHR population? Please add the Ns here as well so this is clear. If it’s the not the latter, please also include how these people differ from the study. For example how many people with an injury were excluded because they didn’t complete all necessary waves? You might have a survivor bias.	There were no differences between enrollees included in the study (n=2,786) and the WTCHR population (N=71,423). This was added to the manuscript, with an explanation for the conclusion of no differences between the two groups.
Are both responders and survivors included in this study? If they are, please include that as a covariate or explain why that is not needed. It seems those with first responder training would have a different PTG outcome than an office worker. These are mentioned as “eligibility group” but it does not seem like this variable was included in the models based on the table footnote and the stats section.	Yes, both responders and survivors were included in our sample. We ran the original multiple linear regressions for the physical and mental scores, including a yes/no variable for enrollee type as a control. The regression coefficient for PTG changed very little from its original value for both physical and mental function. However, we did include enrollee type in our final model.
Minor – can you format table 1 in a clearer way? It is hard to tell what each of the columns are since some are % and some are scores and since it spans two pages.	We have added text to clarify the meaning of the columns. We hope Table 1 is now sufficiently clear.
Please make sure the results in table 2 are clear. I’m not sure what the top section (first 4 rows) of results are since the only results reported in the text are from the last 4 rows. The header says “sample between chambers and canal” but from line 110 I thought everyone was below chambers – from a map it looks like canal is above (north). Please clarify this table as does not stand alone in its current form.	We have added subsection titles to Table 2. We included results between Canal and Chambers Streets because reviewers of the previous version of this paper requested it.
Was this included in the methods section? I’m not seeing this in your stats section. I understand that a reviewer asked for it but I’m still not seeing how the canal to chambers participants are in your cohort of those below chambers. Can you please explain this? Additionally, please include the breakdown of these two groups in table 1 as this seems to be an important variable as you are stratifying by it.	The region from canal to chambers streets is distinct from the region below chambers Street. The main analysis concerned the latter region. A previous reviewer asked us to compare regression results between the two regions, so we added that information to Table 2. We added material to the statistics section to describe this additional calculation.
can you please include the more important limitations likes uncontrolled confounders as mentioned above	The reviewer mentioned that depression could confound the PTG/SF-12 association. We included depression as a control variable in the multivariable regressions described in Table 2.