Incarceration status and cancer mortality: A population-based study.

BACKGROUND: The complex relationship between incarceration and cancer survival has not been thoroughly evaluated. We assessed whether cancer diagnosis during incarceration or the immediate post-release period is associated with higher rates of mortality compared with those never incarcerated.
METHODS: We conducted a population-based study using a statewide linkage of tumor registry and correctional system movement data for Connecticut adult residents diagnosed with invasive cancer from 2005 through 2016. The independent variable was place of cancer diagnosis: during incarceration, within 12 months post-release, and never incarcerated. The dependent variables were five-year cancer-related and overall survival rates.
RESULTS: Of the 216,540 adults diagnosed with invasive cancer during the study period, 239 (0.11%) people were diagnosed during incarceration, 479 (0.22%) within 12 months following release, and the remaining were never incarcerated. After accounting for demographics and cancer characteristics, including stage of diagnosis, the risk for cancer-related death at five years was significantly higher among those diagnosed while incarcerated (AHR = 1.39, 95% CI = 1.12-1.73) and those recently released (AHR = 1.82, 95% CI = 1.57-2.10) compared to the never-incarcerated group. The risk for all-cause mortality was also higher for those diagnosed with cancer while incarcerated (AHR = 1.92, 95% CI = 1.63-2.26) and those recently released (AHR = 2.18, 95% CI = 1.94-2.45). CONCLUSIONS AND RELEVANCE: There is a higher risk of cancer mortality among individuals diagnosed with cancer during incarceration and in the first-year post-release, which is not fully explained by stage of diagnosis. Cancer prevention and treatment efforts should target people who experience incarceration and identify why incarceration is associated with worse outcomes.

Introduction

The United States has the highest per capita rate of incarceration (698 per 100,000) in the world [1], and disproportionately incarcerates people of racial and ethnic minority backgrounds and those of low socioeconomic status [2]. Cancer is the leading cause of death among people incarcerated in prison, accounting for about 30% of all deaths [3]. Prior work has suggested that being incarcerated is associated with shorter survival after a cancer diagnosis. One single-center study found that the median survival was substantially inferior (21 months for incarcerated versus 54 months in non-incarcerated) [4], although this analysis did not account for cancer type or stage at diagnosis. Moreover, the impact of incarceration on health outcomes for people with cancer may persist even after incarceration. Medicare beneficiaries recently released from correctional facilities had higher cancer-related hospitalization rates compared with the general population in the months following release [5]. Other population-based studies have also found an increased risk of cancer mortality among incarcerated individuals compared with the general population following release [6-10]. However, none of these studies ascertained place of cancer diagnosis to distinguish cancers diagnosed during or after release from incarceration.

Critical knowledge gaps regarding the relation between incarceration, clinical factors, and cancer outcomes remain [11]. On the one hand, having been incarcerated may counterintuitively improve cancer outcomes for certain populations, given a constitutionally guaranteed access to healthcare in correctional facilities [12]. On the other hand, people who are incarcerated may face barriers to cancer screening or evaluation of symptoms–either prior to or during incarceration–and therefore present at a later stage. Indeed, recent evidence from a single urban tertiary care center shows that cancer is diagnosed at more advanced stages among incarcerated people [13]. After release from correctional facilities, access to community healthcare could be impaired due to barriers in obtaining public insurance, housing, and employment compounded by the lasting impact of a criminal record, leading to low rates of screening and barriers to cancer treatment [14].

In order to identify fundamental contributors to racial and ethnic cancer disparities, it is critical to further our understanding of the relationship between the carceral system and cancer outcomes. We therefore examined whether being diagnosed with cancer during incarceration or immediately post release is associated with an increased risk of cancer-related and all-cause mortality using a statewide data linkage. We also assessed whether stage at diagnosis might account for the relationship. We hypothesized that diagnosis of cancer while incarcerated or in the immediate post release period would be independently associated with a higher likelihood of cancer mortality than those never incarcerated, and this association would be independent of stage at diagnosis.

Methods

Study design and data sources

We created a statewide linkage of administrative data from the Connecticut Tumor Registry (CTR) and Connecticut Department of Correction (CT DOC) to assess the association between place of cancer diagnosis and mortality among patients diagnosed with invasive cancer in Connecticut from 2005–2016. The CTR is a population-based registry that includes all reported cancers diagnosed in Connecticut residents since 1935, including data on the first course of treatment and follow-up data for estimating cancer survival. All medical providers, hospitals, and private pathology laboratories are required by state law to report new cancer diagnoses to the registry, including records on incarcerated persons. CT DOC is one of six unified correctional systems in the United States with all jails and prisons supervised by a single agency, making it easier to ascertain any admissions into the state’s jails or prisons.

To identify people with a diagnosis of cancer during the study period and the place of diagnosis, we linked individuals with any incarceration history in the CT DOC master files between 2005 and 2016 to tumor registry data, from the same period, using name, date of birth, sex, race and ethnicity, and social security number. Individuals with cancer that did not have a match in the CT DOC file during the study period were considered the never-incarcerated group. For individuals in the CT DOC file that matched with the tumor registry, we used the CT DOC movement files, which has information on when individuals are admitted and released, to determine whether they were diagnosed while incarcerated or within 12 months of release. Information extracted from the CTR included age at cancer diagnosis, month and year of cancer diagnosis, cancer stage, type of cancer, vital status (deceased or alive), number of survival days from cancer diagnosis until the date of death or end of study period, and the primary cause of death. Details of our study protocol and partnership with relevant stakeholders of this study have been previously described in full detail [15]. The Yale University Institutional Review Board, the Connecticut Department of Public Health Human Investigations Committee and the CT DOC Research Advisory Council approved the study. Participant consent was waived by Yale University Institutional Review Board, and the Connecticut Department of Public Health Human Investigations Committee. This study follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

Study sample

We included only cases with invasive cancers and selected the first diagnosed invasive cancer during the study period as the index cancer. If multiple invasive cancers were diagnosed synchronously, the more advanced stage case was designated the index cancer. If a person had more than one invasive cancer diagnosed on the same day and at the same stage, the cancer associated with primary cause of death was selected as the index cancer. We excluded: (1) individuals with sex defined as “transsexual” or “other” per the Department of Public Health data use requirements that prohibit reporting data of small cell sizes to protect confidentiality of individuals and minimize incorrect conclusions; (2) individuals younger than 18 years of age; (3) in-situ cancer diagnoses; and (4) people with index cancers diagnosed more than 12 months after release from a correctional facility, as the effect of incarceration may be more difficult to disentangle from other community factors.

Measures

The independent variable in our study was place of cancer diagnosis, categorized as follows: 1) Individuals diagnosed during incarceration (incarcerated group); 2) individuals diagnosed within 12 months following release from incarceration (recently released group); and 3) individuals with no exposure to the criminal justice system during the study period (never incarcerated). The primary outcome measures were five -year cancer-related mortality, defined as death from cancer based on ICD-10 codes, and survival time.

Study covariates included participant demographic and index cancer characteristics. We created categories for age (18–30 years, >30 to 50, >50 to 70, and >70); race and ethnicity (non-Hispanic Black, Hispanic, non-Hispanic white, and other); and sex (male or female). Race and ethnicity were derived from the CTR and were ascertained from documentation provided by the healthcare provider, system, or pathology lab which reported the cancer diagnosis. To eliminate groups with small sample sizes, we described cancer stage as 1) localized, including Stage I cancers localized to organ; 2) regional, including Stage II and III cancers that are larger and involve regional nodes; 3) distant which is defined by Stage IV cancers; or 4) unknown, defined as unstaged cancers. We grouped cancers with the potential of early detection through screening or surveillance (breast, colorectal, cervical, and prostate) as screenable and all others as non-screenable. We excluded lung and liver cancer from the screenable cancer group because there were no broadly-implemented screening guidelines applicable to these two cancers during our study period. We also grouped cancers by organ systems into the following categories to avoid small cell sizes: breast, gastrointestinal, thoracic, genitourinary (non-reproductive organs), male reproductive, female reproductive, leukemia/lymphomas, central nervous system, head and neck, sarcoma, skin, and others Table in .

Data analysis

We compared demographic and cancer characteristics by place of cancer diagnosis (in a correctional facility, recently released, or in the community and never incarcerated) using Chi-square or Fisher exact tests. To maintain the confidentiality of individual records, we did not report results that include fewer than five unweighted cases in any cell [16].

We used the Kaplan-Meier method to determine the five-year survival rate by place of diagnosis. Survival time in days was calculated from the date of index cancer diagnosis through the date of death. Patients who were alive five years after the index cancer diagnosis were censored at 1,825 days, and cases with follow-up time less than five years who did not die during follow-up were censored at the end of the study period (12/31/2018). To protect patient confidentiality, we only received the month and year of the cancer diagnosis from the CTR. We performed a sensitivity analysis using the first, fifteenth, and thirtieth day of the month as the date of cancer diagnosis, which did not yield any substantive differences. Herein, we report the outcome models based on diagnosis on the fifteenth of the month.

We then used Cox proportional hazard regression models to determine differential survival risk between those diagnosed during incarceration, following release, and in the community. The first model was performed for cancer-related death and a secondary model for death from all causes. Survival time for the Cox survival analysis was calculated in a fashion similar to the 5-year survival rate, but in lieu of 5-years of follow-up duration, all patients were censored at the end of the 2018 calendar year. Cox models included the primary predictor (place of diagnosis, with those never incarcerated as a reference group), as well as all other covariates, including age group, race and ethnicity, sex, and screenable versus non-screenable cancers. We also included stage of cancer diagnosis to examine its contribution to the risk of cancer-related mortality and all-cause mortality in this population. We considered an association significant for a p-value <0.05 and calculated adjusted hazards ratios (AHRs) with 95% confidence intervals. We conducted all analyses in SPSS version 26 [17].

Results

Overall, 216,540 individuals were diagnosed with an invasive cancer in Connecticut during the study period; 239 of whom were diagnosed in prison, 479 within 12 months post-release, and 215,822 with no exposure to incarceration. The mean length of incarceration for those diagnosed while incarcerated was 4.6 years (SD = 6.1), and 1.1 years (SD = 1.1) for the recently released group. On average individuals had a cancer diagnosis 3.6 years (SD = 5.7) following admission into prison, and 5.1 months (SD = 3.5) after release Table in . The median age at diagnosis was 50 years for those incarcerated, 51 years for those within 12 months post-release and 66 years for those never incarcerated. In the incarcerated group 72.4% had been incarcerated for a year or longer, and 26.1% in recently released group were incarcerated for at least 1 year.

People diagnosed with invasive cancer while incarcerated and within 12 months post-release were more likely to be younger, male, and non-Hispanic Black or Hispanic compared with those never incarcerated (p<0.001) (). Cancers originating from the gastrointestinal system (dominated by colorectal and liver) were the most common cancers in all three groups (incarcerated, 25.5%; recently released, 32.8%; and never incarcerated, 17.8%, respectively). Other cancers commonly diagnosed among those incarcerated included those in the thoracic region (97% of which were lung), the male reproductive organs (62% were prostate), and leukemia and lymphomas. Incarcerated individuals were diagnosed with cancer at a distant stage more frequently compared to those recently released or never incarcerated (42.7% vs. 28.4% vs 25.0%, p<0.001) (). The average length of incarceration for individuals diagnosed with screenable cancers at a distant stage (metastatic) in the incarcerated group was 5.9 (SD = 7.5) years, compared with 4.8 (SD = 5.9) years for those with localized screenable cancers (p = 0.58). For screenable cancers, 58.8% were diagnosed at a distant stage among incarcerated group; 40.8% for recently released group, and only 31.9% in the never incarcerated group.

The 5-year survival rate was lowest among individuals diagnosed within 12 months of release from a correctional facility (54.6%; 95% Confidence Interval (CI): 46.8% - 62.4%), compared to 63.2% (95%CI: 55.4% - 71.0%) among those diagnosed while incarcerated, and 67.2% (95%CI: 67.0% - 67.4%) among the never incarcerated group (). For the subset of patients with screenable cancers, the 5-year survival rate was 67.4% (95%CI: 53.4% - 81.5%) for individuals diagnosed while incarcerated, 77.6% (95%CI: 69.5% - 85.6%) for individuals diagnosed within 12-months post-release and 85.2% (95%CI: 84.9% - 85.5%) among those never incarcerated. Five-year survival rates for specific cancers also differed by place of diagnosis. For breast cancer, the 5-year survival rate was lowest for incarcerated patients (60%), compared to those within 12-months after release (81.8%) and those never incarcerated (89.5%); a pattern similar to that observed for all screenable cancers given the small number of women and breast cancers diagnosed in the incarcerated group.

The cumulative survival for all causes of death was similar for both groups with criminal justice exposure and significantly lower compared with the never incarcerated group (). In adjusted analyses, individuals diagnosed while incarcerated had worse all-cause mortality (Adjusted Hazard Ratio (AHR) = 1.92, 95% CI, 1.63–2.26) compared with those never incarcerated, as did those diagnosed immediately post release (AHR = 2.18, 95% CI 1.94–2.45) ()

a: Kaplan Meier curves of all-cause mortality by place of diagnosis. Green = Never incarcerated; Blue = Incarcerated; Red = Post-release b: Kaplan Meier curves of Cancer-related mortality by place of diagnosis. Green = Never incarcerated; Blue = Incarcerated; Red = Post-release.

Cancer-specific survival was also lowest among the group diagnosed after release, followed by those diagnosed while incarcerated, and highest amongst those never incarcerated (). Individuals diagnosed while incarcerated had worse cancer-related mortality (AHR) = 1.39, 95% CI, 1.12–1.73) compared with those never incarcerated. Those diagnosed immediately post release had a higher risk for cancer-related mortality (AHR = 1.82, 95% CI 1.57–2.10) compared to those never incarcerated ().

Stage of diagnosis accounted for some of the relation between incarceration status and mortality but was not the sole mediator. For instance, the risk of cancer-related death among patients who were diagnosed while incarcerated changed from 1.73 (95%CI: 1.40–2.15) to 1.39 (95%CI: 1.12–1.73) after adjusting for stage at diagnosis. Among patients diagnosed within 12 months of release, the hazard ratio changed from 1.91 (95%CI: 1.65–2.21) to 1.82 (95%CI: 1.57–2.10) after adjusting for stage at diagnosis ().

Discussion

Using data from a statewide population-based linkage, we found that being diagnosed with cancer while incarcerated or within 12 months following release was associated with higher rates of cancer-related and all-cause mortality compared with people never incarcerated, even after controlling for participant demographics, type of cancer, and stage of cancer at diagnosis. Possible reasons for the high risk of death include having limited access to high quality cancer care or even experimental cancer treatment, such as participation in clinical trials, access to palliative care, and attention to patient’s social determinants of health, including social support and food [18-22].

While prior studies have identified the association between incarceration and cancer survival time [13], our study illuminates the immediate post release period as a particularly high-risk time period. People diagnosed shortly following release from correctional settings are more likely to die earlier from cancer compared with counterparts diagnosed during incarceration, likely reflecting the significant barriers following release for obtaining timely cancer care. People are often released from correctional facilities without health insurance, medical records, or a primary care appointment, while also contending with severe structural barriers to obtaining housing, food, and employment [23, 24]. While Medicaid coverage expanded in Connecticut during this study, recent work has shown that insurance is necessary but not sufficient to engage people released from correctional settings into mental health or substance use treatment [25, 26]. This may also be the case for cancer treatment. Primary care for people recently released from correctional systems should include screening for treatable cancers, evaluation of symptoms, and addressing social determinants to mitigate these disparities in cancer related deaths.

Our findings regarding stage of cancer diagnosis and setting of cancer diagnosis are striking. Almost two-thirds of patients diagnosed with cancer while incarcerated and more than half of those diagnosed following release were found to have regional spread or metastasized cancer, including for screenable cancers such as colorectal cancer. These data support the previously reported trend of late-stage cancer diagnosis in the incarcerated population [13]. Past research has indicated that incarcerated individuals are less likely to receive age-appropriate screening and this disparity warrants immediate health system intervention [27].

Our study also corroborates prior findings of distant-stage diagnosis of lung cancer and colorectal cancer among the incarcerated population. Despite the increased risk for lung cancer in the justice-involved population, the routine use of low-dose computed tomography scans for screening in individuals with a significant smoking history, as recommended by the US Preventive Services Task Force, has yet to be adopted in correctional facilities [28]. Colonoscopies are also not widely available [27]. Further, past studies have documented that justice involved men may be hesitant to get screened, even when colonoscopies are available, due to fatalistic thinking about cancer [29]. While there is no systematic study examining screening rates among incarcerated people or people who have been incarcerated, past single site studies have shown low rates of preventive cancer screening overall [30, 31]. Given the rising incidence of colorectal cancers among people younger than 50, especially among young Black adults, targeting this population for universal access to screening and health information about cancer prevention through innovative digital platforms, may be important to reduce cancer disparities [32]. In addition, because the majority of those diagnosed with cancer while incarcerated had been incarcerated for more than a year, targeted cancer screening for those incarcerated for more than a year should be implemented. Our findings may be pertinent to other state prison systems as the national data reveal that cancer is now the leading cause of death among incarcerated individuals [3].

Limitations

The small sample size for individuals diagnosed with cancer while incarcerated precluded examination of how cancer-specific mortality rates differed by place of diagnosis. However, the number of deaths in incarcerated individuals in our study are comparable to Bureau of Justice Statistics data on cancer deaths in correctional facilities in Connecticut [3]. Our measure of individuals’ race and ethnicity and biologic sex were derived from medical records or health care providers and systems as provided to CTR, which could be less accurate than self-report. In addition, given the lack of established correctional guidelines on lung and liver cancer screening during our study period, this cancer type was excluded from our subset analysis of screenable cancers.

Conclusion

Our study reveals that incarceration, whether at the time of receiving a cancer diagnosis or within the 12 months of release, is a significant determinant of disparities in cancer mortality. Interventions that improve preventive screening and quality cancer care for justice-involved populations are imperative components to attaining cancer health equity.

List of cancer ICD codes and grouping by organ system.

(PDF)

Click here for additional data file.

Length of incarceration and average time to cancer diagnosis.

(PDF)

Click here for additional data file.

4 May 2022

Submitted filename: PLOS-One_CTIncarcerationCancer.docx

Click here for additional data file.

3 Aug 2022

Response to Reviewers Comments

Journal Requirements

Comment 1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

Response: Thank you for the reminder. We have ensured this.

Comment 2. Please provide additional details regarding participant consent. In the ethics statement in the Methods and online submission information, please ensure that you have specified (1) whether consent was informed and (2) what type you obtained (for instance, written or verbal, and if verbal, how it was documented and witnessed). If your study included minors, state whether you obtained consent from parents or guardians. If the need for consent was waived by the ethics committee, please include this information.

Response: We have edited the manuscript study design section to include the following statement, “participant consent was waived by Yale University Institutional Review Board, and the Connecticut Department of Public Health Human Investigations Committee”

Location: Page 6, line 117-119

If you are reporting a retrospective study of medical records or archived samples, please ensure that you have discussed whether all data were fully anonymized before you accessed them and/or whether the IRB or ethics committee waived the requirement for informed consent. If patients provided informed written consent to have data from their medical records used in research, please include this information.

Response: We have edited the manuscript study design section to include the following statement, “participant consent was waived by Yale University Institutional Review Board, and the Connecticut Department of Public Health Human Investigations committee. The limited dataset accessible to us included protected health information.”

Location: Page 6, line 117-119

Comment 3. Thank you for stating the following in the Acknowledgments Section of your manuscript:

"This work is supported by the National Institutes of Health grant. Data on incident breast cancer cases used in this study were obtained from the Connecticut Tumor Registry located in the Connecticut Department of Public Health (DPH). Data on incarceration status were obtained from the Connecticut Department of Correction (DOC) master file of all individuals that interacted with the DOC during the study period and DOC movement files. The National Institutes of Health had no role in the design and conduct of the study; management, analysis, and interpretation of the data; manuscript preparation, and decision to submit the manuscript for publication. The analysis, interpretation or conclusions drawn from these data are the responsibility of the authors and do not represent the views of neither the National Institutes of Health nor the United States Department of Health and Human Services or any of its affiliates."

We note that you have provided funding information. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

"This work is supported by the National Institutes of Health R01 5R01CA230444-02, awarded to CPG and EAW. The National Institutes of Health had no role in the design and conduct of the study; management, analysis, and interpretation of the data; manuscript preparation, and decision to submit the manuscript for publication."

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

Response: We have removed the funding statement and ensured that the funding statement in the online submission is accurate and have included it in the cover letter.

Location: Page 19, lines 341- 351 and updated cover letter

Comment 4. Thank you for stating the following in the Competing Interests section:

"I have read the journal's policy and the authors of this manuscript have the following competing interests. OTO reports funding unrelated to submitted work from Radiation Oncology Institute, NRG Oncology and Bristol Meyers Squibb Foundation. CPG has received research funding NCCN Foundation (funds provided by AstraZeneca), Genentech as well as funding from Johnson and Johnson to help devise and implement new approaches sharing clinical trial data. The other authors have no competing interests to disclose."

Please confirm that this does not alter your adherence to all PLOS ONE policies on sharing data and materials, by including the following statement: "This does not alter our adherence to PLOS ONE policies on sharing data and materials.” (as detailed online in our guide for authors http://journals.plos.org/plosone/s/competing-interests). If there are restrictions on sharing of data and/or materials, please state these. Please note that we cannot proceed with consideration of your article until this information has been declared.

Please include your updated Competing Interests statement in your cover letter; we will change the online submission form on your behalf.

Response: We have added the requested statement in the form and cover letter.

Location: Online forms and cover letter.

Comment 5. In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Response: Legal restrictions to sharing data publicly has been emphasized in the edited cover letter

Location: Cover letter

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Response: We have addressed this request in our cover letter

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data:http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

Response: Thank you for the reminder.

Location: Cover letter

Comment 6. We note that you have included the phrase “data not shown” in your manuscript. Unfortunately, this does not meet our data sharing requirements. PLOS does not permit references to inaccessible data. We require that authors provide all relevant data within the paper, Supporting Information files, or in an acceptable, public repository. Please add a citation to support this phrase or upload the data that corresponds with these findings to a stable repository (such as Figshare or Dryad) and provide and URLs, DOIs, or accession numbers that may be used to access these data. Or, if the data are not a core part of the research being presented in your study, we ask that you remove the phrase that refers to these data.

Response: We have removed such phrases “data not shown” from manuscript

Location: Page 10, line 208

Comment 7. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information:http://journals.plos.org/plosone/s/supporting-information.

Response: N/A

Comment 8. Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Response: We have reviewed the reference list and it is complete and free of retracted articles.

Location: Reference list.

Reviewers' comments:

Reviewer's Responses to Questions - Comments to the Author

Comment 1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Response: Thank you for your review.

Comment 2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Response: Thank you for your comments.

Comment 3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: No

Response: We are unable to fully provide the underlying data of our findings as incarcerated individuals are considered vulnerable research population. Thus, participant privacy and legal restrictions prohibit us from sharing the underlying data for this study.

Comment 4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Response: Thank you for the feedback.

5. Review Comments to the Author – Reviewer 1

Comment: Important topic – that cancer is the leading cause of death among older prisoners – topic warrants investigation.

Intro – is it worth quantifying the cancer disparities faced by people with CJ involvement compared to the gen pop? Just to really drive home the point about significance?

Response: We thank the reviewer for this feedback regarding our in-depth introduction section highlighting cancer disparities faced by people with criminal justice involvement. We felt it was important to use this medium as an opportunity to educate readers about an overlooked population in oncology, that is not only growing in size but also the notable change in leading cause of mortality previously known to be cardiovascular disease is now cancer. Furthermore, it provides an in-depth background on the various challenges of cancer care for this population, and we hoped to highlight the critical knowledge gap regarding incarceration and cancer outcomes.

Location: No changes made.

Comment: Method – I realize it is not standard to report how the team got access to these data, but it might be useful to the field.

Otherwise, methods are clear.

Response: Thank you for this point; we agree that it is important for others in the field to know how the database was constructed. We have published our methodology for this project in full detail so that others can adopt our approach to accessing data and building a linkage registry for a population that is understandably difficult to investigate. Please see, Puglisi L, Halberstam AA, Aminawung J, Gallagher C, Gonsalves L, Schulman-Green D, et al. Incarceration and Cancer-Related Outcomes (ICRO) study protocol: using a mixed-methods approach to investigate the role of incarceration on cancer incidence, mortality and quality of care. BMJ open. 2021;11(5):e048863. In our methods section, we summarized how we created a statewide linkage registry using data from a tumor registry and department of corrections. We also described how we identified and linked individuals using pertinent identifying factors, matched movement files and selected for those with invasive cancer diagnosis.

Location: “Details of our study protocol and partnership with relevant stakeholders of this study have been previously described in full detail (15).” Page 6, Lines 114-115

Comment: Results – the gastro cancer findings are fascinating. Who knew so pervasive? Important finding. Finding that men more likely to be diagnosed also interesting.

Response: We were also surprised by this finding, and it further highlights the fact that prison health is public health. The rates of gastrointestinal cancers are increasing in the United States and now more pervasive in younger population. Thus, our findings highlight that our prison walls are not impermeable to similar trends we observe in the community.

Location: No changes made.

Comment: Discussion – how come no mention of women’s cancers? Is it not interesting that people diagnosed during incarceration more likely to be male? What’s that about?

Response: Our data sample is heavily weighted towards men due to the characteristics of the incarcerated population. Yet, it is important to highlight the cancer outcomes for women to the extent that our data allow, given that there has been too little focus on the health of women who are incarcerated. In our results (and tables), we report the incidence and five-year survival rates of women with breast and female reproductive cancers. We included cervical and breast cancer in our subgroup analysis of screenable cancers. We had to group them together given the small number of individual cases, so as to limit the likelihood of unintentionally revealing the identity of participants. In our results, we have now included this sentence to highlight the unique 5-year survival rate from breast cancer between the three cohorts: “For breast cancer, the 5-year survival rate was lowest for incarcerated patients (60%), compared to those within 12-months after release (81.8%) and those never incarcerated (89.5%); a pattern similar to that observed for all screenable cancers given the small number of women and breast cancers diagnosed in the incarcerated group.”

Location: Page 12, lines 226-229

Comment: Worth saying how and to what extent findings from CT can be extrapolated to other places?

Response: Thank you for the suggestion. We have included this sentence in the discussion in this regard: “Our findings may be pertinent to other state prison systems as the national data reveal that cancer is now the leading cause of death among incarcerated individuals.”

Location: Page 18, lines 318-320

6. Review Comments to the Author – Reviewer 2

Methods

Comment: Page 6, line 101: Incarceration data was from 2005-2016. Was this linked to cancer registry data from the same time period?

Response: Yes, the incarceration data from the Connecticut Department of Correction was linked with the Connecticut Tumor Registry from the same period. The linkage was done using name, date of birth, sex, race, ethnicity, and social security number. This is also described in the methods section.

Location: We have clarified this in the study design section by adding “from the same period” to text can be seen on page 6, lines 105

Comment: Page 7, lines 127-131: How were patients categorized if they were previously incarcerated > 1 year from their cancer diagnosis?

Response: Patients were categorized based on the timing of their incarceration status in relation to their cancer diagnosis. The three categories are: never incarcerated, currently incarcerated, or recently released from incarceration in the last 12 months. Hence, if a patient was diagnosed with cancer while living in the community, but had been released from incarceration >12 months prior to their cancer diagnosis, we excluded them.

Location: No changes made.

Comment: Page 7, line 141-142. Screenable cancers: Recognizing the explanation provided in the limitations section, consider adding lung and liver cancer to the screenable cancers. Though these screening mechanisms are not part of incarceration screening guidelines (e.g. NCCHC, federal BOP), inclusion of these cancer types then leads to data that helps illustrate the extent to which the current cancer burden can be ameliorated with all available screening modalities.

Response: We considered the inclusion of lung and liver cancer to our analysis of screenable cancers. However, national consensus guidelines for these cancer types have only been recently refined for general population and widely adopted in community practice the last 2-3 years, which is beyond our study period ending in 2016. Because the screening of these cancers was not widely implemented during our study period, we did not classify them as such. However, in our discussion section we have added the following sentence: “We excluded lung and liver cancer from the screenable cancer group because there were no broadly-implemented screening guidelines applicable to these two cancers during our study period.”

Location: Page 8, lines 150-152

Comment: Page 9, lines 166-168. Consider using cancer category as a variable in place of screenable cancers as a variable in the overall survival model. Simplifying cancer types to “screening cancers” may mask significant imbalance of cancers like cervical, breast and colon that have very different prognoses. For instance, never-incarcerated has 16% breast cancer vs 2-5% for the other groups and breast cancer generally has a favorable prognosis compared to many other cancers. Adding at least a sensitivity analysis using cancer category in place of screenable cancer can ensure that cancer mix is not explaining survival differences. This variable could also be included in the survival model for the subset of screenable cancers for the same reason (results presented on page 11, line 206)

If available, a breakdown of cancer diagnosis / survival by incarceration in jail vs prison (if those are separate institutions within the state) would be beneficial.

Response: Thank you for the suggestion, and we agree such breakdown would be insightful. However, we are unable to perform this analysis given the Department of Correction in the state of Connecticut is a unified system (meaning jails and prisons are managed within one system), and even within one facility there are people being held in jail and in prison. Thus, we focused on contact with the Department of Correction, as we can ascertain length of time incarcerated.

Secondly, we were unable to adjust for specific cancer types due to lack of sufficient number of patients for certain types of cancers. Even when the Cox regression converged, the parameter estimates are not stable (they are subject to change if we have a different sample size) and most likely biased. In addition, gender-specific cancers by nature are confounded with sex (also included in the Cox regression as a covariate), leaving cells/blocks mutually exclusive (female reproductive type of cancer will be 0 for male participants). Thus, we decided to aggregate cancer types into larger groups, screenable vs non-screenable, for our final Cox regression models.

Location: No changes made.

Results

Comment: Page 9: Please provide median age of diagnosis.

Response: Thank you for the inquiry. We have updated our results section to include this sentence: “The median age at diagnosis was 50 years for those incarcerated, 51 years for those within 12 months post-release and 66 years for those never incarcerated.”

Location: Page 9, lines 190-191

Comment: Page 10, line 190: “For screenable cancers, 58.8% were diagnosed at a late stage…”. Please move the definition of “late stage” up from the next sentence. Consider providing this data for distant disease instead of late disease, so it can be compared to the 42.7% data point in the previous sentence. I personally think of “late stage” as representing incurable, distant disease rather than also encompassing locally advanced disease that is still potentially curable.

Response: With regards to separating late stage and distant, we used late stage and distant stage interchangeably in the paper, hence the confusion about separating locally advanced from metastatic patients. Distant disease is defined as Stage IV cancers, and this definition is stated in the “measures” section. Based on this, we do not have additional analysis to provide as we now have 4 groups of cancer stages: localized, regional, distant and unstaged/unknown. We have edited the manuscript to solely use distant stage as the descriptor for metastatic/stage 4 patients

Location: Page 10, lines 204,206; page 17, line 303.

Comment: Thinking about how this data can be used to direct implementation of interventions to improve cancer care:

Page 10 line 192-194: Among incarcerated patients with a screenable cancer that were diagnosed with a late stage, the median length of incarceration was 5.9 years, but what percentage of patients had been incarcerated for 1 year or longer? It is another way of illustrating what proportion of cancers might be caught earlier.

Response: Thank you for the feedback and insight on future intervention/policy initiative. We have included this sentence in our results section: “In the incarcerated group 72.4% had been incarcerated for a year or longer and 26.1% in recently released group were incarcerated for at least 1 year.” We have also included a sentence in our discussion about how this data can be used: “In addition, because the majority of those diagnosed with cancer while incarcerated had been incarcerated for more than a year, targeted cancer screening for those incarcerated for more than a year should be implemented.”

Location: Pages 9-10, line 191-193; pages 17-18, lines 316-318

Comment: Among incarcerated and formerly incarcerated patients with screenable cancers diagnosed at a late stage, what is the breakdown of cancer type? This determines which screen programs need to be beefed up.

Response: We thank the reviewer for the suggestion. However, as indicated in our response to reviewer one above, we are unable to provide a detailed breakdown by cancer type among screenable cancers. We grouped them given the small number of individual cases so that we would limit the likelihood of unintentionally revealing the identity of participants.

Location: No changes made.

Comment: The authors mention Medicaid expansion in the discussion – if the registry reports insurance status at diagnosis for recently incarcerated and the general population that would be interesting to see.

Response: Thank you for this suggestion. Unfortunately, we do not have insurance status data to answer this question. Of note, the state of Connecticut expanded Medicaid to include justice involve individuals prior to the implementation of the Affordable Care Act so many of those released likely had access to Medicaid upon release.

Location: No changes made.

Submitted filename: Incarceration - PLOS One - Response to reviewers.docx

Click here for additional data file.

2 Sep 2022

Incarceration status and cancer mortality: A population-based study

PONE-D-22-07747R1

Dear Dr. Aminawung,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Andrea Knittel

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

8 Sep 2022

PONE-D-22-07747R1

Incarceration status and cancer mortality: A population-based study

Dear Dr. Oladeru:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Andrea Knittel

Academic Editor

PLOS ONE

Table 1

Characteristics of individuals diagnosed with invasive cancer between 2005 and 2016 cancer by incarceration status at cancer diagnosis.

Characteristic	Incarcerated (N = 239)	12-months post release (N = 479)	Never incarcerated (N = 215,822)	p-value
	No. (%)	No. (%)	No. (%)
Age at diagnosis (years)				<0.001
18–30	22 (9.2)	25 (5.2)	3489 (1.6)
>30–50	102 (42.7)	197 (41.1)	28389 (13.2)
>50–70	104 (43.5)	241 (50.3)	99237 (46.0)
>70	11 4.6)	16 (3.3)	84707 (39.2)
Race and Ethnicity*				<0.001
Non-Hispanic White	112 (46.9)	236 (49.3)	181358 (84.0)
Non-Hispanic Black	83 34.7)	156 (32.6)	14913 (6.9)
Hispanic	>39 (>16.3)	>82 (>17.1)	13677 (6.3)
Other	<5 (<2.1)	<5* (<1.0)	5874 (2.7)
Sex				<0.001
Male	216 (90.4)	412 (86.0)	103317 (47.9)
Female	23 (9.6)	67 (14.0)	112505 (52.1)
Cancer Type (categories)*				<0.001
Breast	6 (2.5)	25 (5.2)	34622 (16.0)
Gastrointestinal	61 (25.5)	157 (32.8)	38376 (17.8)
Thoracic	38 (15.9)	64 (13.4)	28247 (13.1)
Male Reproductive	29 (12.1)	68 (14.2)	31120 (14.4)
Leukemia & Lymphoma	37 (15.5)	45 (9.4)	19743 (9.1)
Urinary (non-reproductive)	11 (4.6)	37 (7.7)	12683 (5.9)
Skin	5 (2.1)	6 (1.3)	10268 (4.8)
Head and Neck	28 (11.7)	40 (8.4)	14856 (6.9)
Female Reproductive	5 (2.1)	6 (1.3)	13299 (6.2)
Central Nervous System (CNS)	<5 (<2.1)	8 (1.7)	3123 (1.4)
Sarcoma	8 (3.3)	12 (2.5)	1823 (0.8)
Other	>6 (>2.5)	11 (2.3)	7662 (3.6)
Cancer stage				<0.001
Localized	72 (30.1)	185 (38.6)	103561 (48.0)
Regional	56 (23.4)	130 (27.1)	45715 (21.2)
Distant	102 (42.7)	136 (28.4)	53971 (25.0)
Unknown/unstaged	9 (3.8)	28 (5.8)	12575 (5.8)
Screenable cancer**				<0.001
Screenable	51 (21.3)	120 (25.1)	84551 (39.2)
Non-screenable	188 (78.7)	359 (74.9)	131271 (60.8)

Note: No = Number

* cell sizes containing fewer than 5 individuals are suppressed due to privacy concerns

**screenable cancer types include breast, colorectal, cervical, and prostate.

Table 2

Five-year survival rates by cancer type and by incarceration status at cancer diagnosis.

Cancer type	Incarcerated	12-months post release	Never incarcerated
	Rate% (95%CI)	Rate% (95%CI)	Rate% (95%CI)
Breast	60.0 (16.9–100.0)	81.8 (64.2–99.4)	89.5 (89.5–89.5)
Gastrointestinal	43.2 (27.5–58.9)	27.8 (18.0–37.6)	47.5 (47.5–47.5)
Thoracic	42.3 (22.7–61.9)	24.1 (10.4–37.8)	28.5 (28.5–28.5)
Male Reproductive	91.8 (80.0–100.0)	100.0 (100.0–100.0)	92.2 (92.2–92.2)
Leukemia & Lymphoma	66.0 (46.4–85.6)	75.5 (57.9–93.1)	68.6 (68.6–68.6)
Genitourinary (non-reproductive)	62.5 (29.2–95.8)	70.0 (50.4–89.6)	67.0 (67.0–67.0)
Skin	100.0 (100.0–100.0)	100.0 (100.0–100.0)	88.5 (88.5–88.5)
Head and Neck	72.1 (52.5–91.7)	76.5 (60.8–92.2)	82.9 (82.9–82.9)
Female Reproductive	100.0 (100.0–100.0)	50.0 (10.8–89.2)	71.3 (71.3–71.3)
Central Nervous System	0 (0.0–0.0)	66.7 (13.8–100.0)	38.8 (36.8–40.8)
Sarcoma	66.7 (29.5–100.0)	81.8 (50.4–100.0)	66.9 (64.9–68.9)
Other	73.3 (41.9–100.0)	42.9 (5.7–80.1)	54.9 (52.9–56.9)
Screenable cancers*	67.4 (53.4–81.5)	77.6 (69.5–85.6)	85.2 (84.9–85.5)
Non Screenable cancers	56.8 (49.0–64.6)	49.8 (12.6–87.1)	55.7 (53.7–57.6)
Total	63.2 (55.4–71.0)	54.6 (46.8–62.4)	67.2 (67.0–67.4)

Note: 95%CI = 95 percent confidence interval

* screenable cancer types include breast, colorectal, cervical, and prostate.

Table 3

Hazard ratios for all-cause mortality.

	Model not adjusted for cancer stage		Model adjusted for cancer stage
	AHR (95%CI)	p-value	AHR (95%CI)	p-value
Status at diagnosis
Never incarcerated	1.00 (ref)		1.00 (ref)
Incarcerated	2.31 (1.96–2.72)	<0.001	1.92 (1.63–2.26)	<0.001
Post-release (≤12 months)	2.30 (2.05–2.59)	<0.001	2.18 (1.94–2.45)	<0.001
Age
18–30 years	1.00 (ref)		1.00 (ref)
>30–50 years	2.17 (1.97–2.40)	<0.001	2.12 (1.93–2.34)	<0.001
>50–70 year	4.68 (4.26–5.15)	<0.001	4.32 (3.93–4.75)	<0.001
>70 years	12.60 (11.47–13.85)	<0.001	10.93 (9.95–12.02)	<0.001
Race and ethnicity
Non-Hispanic White	1.00 (ref)		1.00 (ref)
Non-Hispanic Black	1.28 (1.25–1.31)	<0.001	1.19 (1.16–1.22)	<0.001
Hispanic	1.03 (1.01–1.06)	0.02	0.97 (0.94–0.99)	<0.001
Other	0.57 (0.54–0.60)	<0.001	0.53 (0.50–0.56)	<0.001
Gender
Male	1.00 (ref)		1.00 (ref)
Female	0.89 (0.88–0.90)	<0.001	0.89 (0.88–0.90)	<0.001
Cancer type
Not screenable cancer	1.00 (ref)		1.00 (ref)
Screenable* cancer	0.38 (0.37–0.38)	<0.001	0.57 (0.56–0.58)	<0.001
Cancer stage
Localized	---	---	1.00 (ref)
Regional	---	---	2.10 (2.06–2.14)	<0.001
Distant	---	---	4.25 (4.18–4.32)	<0.001
Unknown/unstaged	---	---	4.94 (4.82–5.05)	<0.001

Note: AHR = Adjusted hazard ratio; 95%CI = 95 percent confidence interval

* screenable cancer types include breast, colorectal, cervical, and prostate.

Table 4

Hazard ratios for cancer-related mortality.

	Model not adjusted for cancer stage		Model adjusted for cancer stage
	AHR (95%CI)	p-value	AHR (95%CI)	p-value
Status at diagnosis
Never incarcerated	1.00 (ref)		1.00 (ref)
Incarcerated	1.73 (1.40–2.15)	<0.001	1.39 (1.12–1.73)	0.003
Post-release (≤12 months)	1.91 (1.65–2.21)	<0.001	1.82 (1.57–2.10)	<0.001
Age
18–30 years	1.00 (ref)		1.00 (ref)
>30–50 years	2.37 (2.12–2.66)	<0.001	2.31 (2.06–2.58)	<0.001
>50–70 year	4.64 (4.15–5.18)	<0.001	4.14 (3.71–4.63)	<0.001
>70 years	9.75 (8.73–10.89)	<0.001	8.00 (7.16–8.93)	<0.001
Race and ethnicity
Non-Hispanic White	1.00 (ref)		1.00 (ref)
Non-Hispanic Black	1.26 (1.22–1.30)	<0.001	1.15 (1.12–1.18)	<0.001
Hispanic	1.00 (0.96–1.03)	0.84	0.92 (0.89–0.95)	<0.001
Other	0.56 (0.52–0.60)	<0.001	0.52 (0.49–0.56)	<0.001
Gender
Male	1.00 (ref)		1.00 (ref)
Female	0.93 (0.91–0.94)	<0.001	0.93 (0.92–0.94)	<0.001
Cancer type
Not screenable cancer	1.00 (ref)		1.00 (ref)
Screenable* cancer	0.29 (0.28–0.29)	<0.001	0.49 (0.48–0.50)	<0.001
Cancer stage
Localized	---	---	1.00 (ref)
Regional	---	---	3.06 (2.99–3.13)	<0.001
Distant	---	---	6.70 (6.55–6.84)	<0.001
Unknown/unstaged	---	---	7.53 (7.31–7.76)	<0.001

Note: AHR = Adjusted hazard ratio; 95%CI = 95 percent confidence interval

* screenable cancer types include breast, colorectal, cervical, and prostate.