https://orcid.org/0000-0002-2510-3570
Figures
Abstract
Background
Evidence on the association between mental health disorders and cancer risk is inconclusive, despite well-established associations between mental health disorders and lifestyle factors such as smoking. This study examines the relationships between depression, anxiety and cancer risk, and the potential mediating effects of lifestyle factors.
Methods
A study of 34,571 participants aged 40–69 years in the CARTaGENE cohort was conducted. Depression was defined by questionnaire (PHQ-9), antidepressant use, and a composite of questionnaire, antidepressant use, or lifetime self-reported physician diagnosis. Anxiety was defined by questionnaire (GAD-7). Co-morbid depression and anxiety was also assessed. Cox regression models were used to investigate associations between mental health and risk of prostate, lung, and all cancers combined. Mediating effects of lifestyle factors were assessed using Baron and Kenny mediation criteria.
Results
There were positive associations between mental health disorders, all cancers and lung cancer risk, however with the exception of anxiety and lung cancer in women (Hazard Ratio [HR] = 1.67, 95% CI: 1.01–2.76), associations were attenuated with adjustment for sociodemographics, health status and lifestyle factors. In the mediation analysis, smoking accounted for 27%, 18%, and 26%, of the total effect between depression (PHQ-9), anxiety, and co-morbidity and lung cancer, respectively in women. In men, smoking accounted for 17% of the total effect between depression (PHQ-9, antidepressant, or lifetime self-report of physician diagnosis) and all cancers.
Conclusions
Positive associations were observed between mental health disorders, all cancer and lung cancer risk, however most relationships were attenuated with adjustment for lifestyle factors. Smoking status mediated a significant proportion of the relationships between mental health disorders and cancer risk.
Citation: Gilham K, Gadermann A, Dummer T, Murphy RA (2023) Mental health, cancer risk, and the mediating role of lifestyle factors in the CARTaGENE cohort study. PLoS ONE 18(2): e0281588. https://doi.org/10.1371/journal.pone.0281588
Editor: Chin-Kuo Chang, University of Taipei, TAIWAN
Received: May 5, 2022; Accepted: January 26, 2023; Published: February 14, 2023
Copyright: © 2023 Gilham et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data underlying the results presented in the study are available from CARTaGENE, https://www.cartagene.qc.ca/en/home.
Funding: RAM's time was funded by the Michael Smith Foundation for Health Research (grant #17644, https://www.msfhr.org/). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: RAM is a consultant to Pharmavite LLC. RAM’s consultancy does not alter adherence to PLOS ONE policies on sharing data and materials.
Introduction
It is estimated that approximately 1 in 2 Canadians will develop cancer in their lifetime, and 1 in 4 Canadians will die from cancer [1]. Cancer is also the leading cause of death in Canada, responsible for approximately 30% of all deaths [2]. The financial and emotional cost of cancer in Canada is substantial–impacting patients, their caregivers, and the healthcare system. Preventive interventions are crucial to minimize the economic and social impact of the burden of cancer in Canada.
Mental health disorders also have a high public health burden. It is estimated that 1 in 5 Canadians are affected annually by mental health disorders [3]. According to Pearson et al., 11.3% of Canadian adults have met the criteria for a depression disorder at some point in their lifetime, and 8.7% have met the criteria for generalized anxiety disorder (GAD) [4]. Individuals with mental health disorders have been shown to have an increased risk of several chronic illnesses, including cardiovascular disease and diabetes [5–7]. However, evidence on the association between cancer and mental health disorders is largely inconclusive. One meta-analysis of nine studies that assessed the relationship between depression and risk of all cancers reported high heterogeneity across studies and inadequate consideration of covariates such as smoking [8]. Another systematic review and meta-analysis of 25 studies with a pooled sample of 1,469,179 participants and 89,716 incident cases of cancer assessed the association between depression and incident cancer risk [9]. This study found that depression was significantly associated with an increased risk of all cancer, liver, and lung cancer [9]. However, no associations were found for breast, prostate, or colorectal/colon cancer [9]. The review also noted that very few studies adjusted for lifestyle risk factors, such as diet, physical activity, tobacco, or alcohol consumption.
Lifestyle factors are closely associated with anxiety, depression, and cancer. Individuals with depression and anxiety are more likely to smoke, consume higher levels of alcohol, engage in lower levels of physical activity, have lower diet quality, and experience disturbed sleep compared to those without depression or anxiety [10–18]. These lifestyle factors are key components of cancer prevention guidelines [19]. In 2015, an estimated 33% of cancer cases diagnosed were attributable to lifestyle factors; tobacco smoking, low fruit and vegetable intake, physical inactivity, and excess weight [20].
Given the established association between healthy lifestyle behaviours with cancer, and mental health disorders, it is possible that the inconsistent associations between depression, anxiety and cancer risk are due to lack of consideration of the impact of behavioural factors. It may also be pertinent to consider the potential modifying role of sex on relationships. There are well established differences in the prevalence of mental health disorders, symptoms and usage of mental health services [21, 22]. The incidence of many of the most common cancers such as colorectal and lung cancer differ by sex [23, 24], as does engagement in healthy lifestyle behaviours [25]. To meet these gaps, this study aimed to examine associations between mental health disorders and cancer risk. Overall cancer risk was assessed, in addition to prostate cancer and lung cancer, the two most common cancer types among men [26], and the second most common type of cancer among women (lung). Models were also tested for effect modification by sex. The secondary aim was to explore whether lifestyle factors mediated associations between mental health disorders and cancer risk.
Methods
Study sample
Participants for the present study were drawn from the CARTaGENE cohort, a study of 43,000 men and women aged 40–69 in Quebec, Canada. Participants were enrolled between 2009 and 2015 as part of the national Canadian Partnership for Tomorrow’s Health (CanPath), a prospective, longitudinal cohort study [27]. Details of the survey methods and cohort profile have previously been published [28]. Ethics approval for this project was obtained from the University of British Columbia Clinical Research Board (Certificate #H17-02706). All participants provided written, informed consent.
At study baseline participants completed self-report questionnaires that asked about sociodemographic factors, medications, and health and lifestyle factors [28]. A proportion of participants attended in-person assessments, which included measurement of height, and bio-impedance measurement of weight. Incident cancers were determined from the provincial cancer registry until 2010. After 2010, incident cancers were defined from hospitalization and medication data collected as part of the administrative health database, Régie de l’assurance maladie du Québec (RAMQ), using Tonelli criteria [29] until April 2018. Tonelli criteria are a set of validated algorithms that can be used to identify chronic conditions from administrative data [29]. The first cancer diagnostic date was used for cases meeting Tonelli criteria.
Participants were excluded if they had any incomplete information from the anxiety (GAD-7) and depression (PHQ-9) questionnaires (n = 2,509), were missing date of cancer diagnosis (n = 34), or reported a prior cancer diagnosis (n = 5,687). Participants who were diagnosed with cancer within the first six months of follow-up (n = 236) were also excluded to minimize reverse causation. After exclusion criteria was applied, the final analytical sample size was 34,571 participants. This included 2,888 total incident cancer cases, with 329 prostate and 305 lung cancer cases.
Depression
Depression at baseline was assessed using multiple definitions that capture different time frames and severity of depression. Depressive symptoms at baseline were assessed using the validated PHQ-9, a nine-item questionnaire that assesses symptoms of depression over the past two weeks [30]. A score of 10 or higher was used to identify depression [31]. This score has a has a sensitivity of 88% and a specificity of 88% for detecting major depressive disorder in an adult primary care sample [31]. In addition, PHQ-9 scores were considered in statistical models as a continuous variable (S1-S6 Tables in S1 File). Participants self-reported medication use in the CARTaGENE data was organized according to the Anatomical Therapeutic Chemical (ATC) Classification system developed by the WHO [32]. Antidepressant use was defined as current use at baseline of a drug belonging to ATC code N06A [33].
Self-reported depression data was also collected in the baseline questionnaire. Participants were asked “has a physician ever told you that you have depression?” Positive responses to this question were recorded as a lifetime prevalence of depression. Self-reported depression was considered in a composite measure of depression which was a positive response for any of PHQ-9 score ≥10, use of antidepressants, or self-report of physician diagnosis.
Anxiety
Anxiety symptoms were assessed at baseline using the validated GAD-7, a seven-item self-report scale that assesses symptoms over the past two weeks [34]. A score ≥10 was used to indicate anxiety, which has been shown to have a sensitivity of 89% and specificity of 82% in an adult primary care setting [34]. Continuous GAD-7 scores were also evaluated in statistical models (S1-S6 Tables in S1 File). The variable of self-reported lifetime diagnosis of anxiety by a physician was only collected in the second phase of the CARTaGENE study, resulting in a high proportion of missing data. Thus, anxiety in this study was only assessed using GAD-7.
Co-morbid depression and anxiety
Co-morbid depression and anxiety was defined as participants who scored a 10 or higher on both the PHQ-9 and the GAD-7 questionnaires.
Covariates
Covariates assessed for potential confounding and mediating effects were selected based on previous research and included sociodemographic factors (age, sex, ethnicity, education, marital status, and income), lifestyle factors (smoking status, alcohol consumption, sleep, physical activity, body mass index (BMI), fruit and vegetable consumption (servings/day), and health status (previous diagnosis of diabetes [35] or myocardial infarction [36]). In addition, cancer specific risk factors were evaluated. For lung cancer, chronic obstructive pulmonary disorder (COPD) and family history of lung cancer, and for prostate cancer, family history of prostate cancer. Covariates were categorized to align with established thresholds, as outlined below, or according to categories previously defined in analyses of CanPath data [37, 38].
Age was analyzed as a continuous variable. Ethnicity was dichotomized as white and non-white due to small counts in ethnicities other than white. Education was categorized as high school or lower, college, and university or higher. Marital status was dichotomized as living with partners (married or common-law), and without partners (single, divorced, or widowed). Household income was categorized as less than $50,000, $50,000–74,999, $75,000–150,000, and greater than $150,000.
Current smokers were those who had smoked at least 100 cigarettes in their lifetime and smoked in the past 30 days. If participants had smoked at least 100 cigarettes in their lifetime, but not in the past 30 days, they were categorized as past smokers. All other participants were categorized as non-smokers. Participants were classified as abstainers (never drinking alcohol), former (drank alcohol before but not over the past 12 months), occasional (≤2–3 times/month), regular (≥ once/week but ≤2–3 times/week), and habitual drinkers (≥4–5 times/week). Average hours of sleep were categorized as <7 hours, 7–9 hours, and 9+ hours per night [39]. Physical activity was evaluated using the International Physical Activity Questionnaire (IPAQ), which asked about the number of days and usual time spent doing vigorous and moderate activity, walking and sitting in the prior 7 days. Physical activity was subsequently categorized as low, moderate, or high as per IPAQ scoring protocol [40]. BMI was categorized based on standard classification: <25 kg/m2 as underweight/normal, 25.0 to 29.9 kg/m2 as overweight, and ≥ 30 kg/m2 as obese. BMI from measured height/weight was used when available, otherwise self-reported height/weight was used. Servings of fruits and vegetables per day were dichotomized into: <five or ≥ five in line with recommendations for consumption to support overall health [41]. Self-reported diagnosis of diabetes, myocardial infarction, and COPD were recorded as “yes” or “no”. Participants were asked to indicate whether immediate blood relatives (mother, father, children, full and half brothers and sisters) had ever been diagnosed with cancer. If they answered ‘yes’, the type(s) of cancer were queried for each relative.
Statistical analysis
Missing covariate data was imputed with the ‘Multivariate Imputations by Chained Equation (MICE)’ package in R [42]. The proportion of missing was 0.02% for age, 0.37% for education, 8.36% for income, 1.64 for ethnicity, 0.30% for smoking status, 2.69% for alcohol consumption, 7.21% for physical activity, 13.8% for BMI, 0.27% for sleep, 1.30% for fruit and vegetable consumption, 0.34% for myocardial infarction, 0.45% for diabetes, and 0.46% for COPD, respectively. Sex and self-perceived health were included in the multiple imputation as auxiliary variables; variables correlated with the missing values that increase the likelihood the missing at random assumption holds. Ten iterations and 20 imputations were needed to reach convergence.
Cox proportional hazards regression analysis was used to estimate hazard ratios (HR) and 95% confidence intervals (CIs) between the exposure and outcome variables. The proportional hazards assumption was verified using Schoenfeld residuals with the ‘survival’ package in R [43]. Days of follow-up was used as the time-scale (continuous in days) included as a covariate in the model [44]. Days of follow-up was determined from the difference between first hospitalization or first claim record in the case of Tonelli criteria or date of diagnosis from the cancer registry and date of consent for the baseline questionnaire. Follow-up for non-cases was determined as the difference between date of last linkage to the RAMQ database, or date of death, and date of consent for the baseline questionnaire. The change in coefficient method was used to determine significant confounders with a stepwise procedure, with a significance level of 10%. Models (1 through 3) were progressively adjusted for sociodemographic and health factors. Effect modification was evaluated by adding an interaction term between two variables in a given model. Sex was identified as an effect modifier for all measures of depression and for anxiety, and thus, models were stratified. Analyses were also performed with estimates pooled for men and women for overall cancer and lung cancer (S2, S3, S6 and S7 Tables in S1 File).
In order to clarify the role of lifestyle factors on the causal pathway between mental health disorders and cancer risk, a mediation analysis was performed to calculate the direct, indirect, and total effect using the mediation package in R software [45]. First, potential mediating effects were screened to establish relevant exposures, mediators, and outcomes using the Baron and Kenny mediation criteria, which is the most widely used method to assess the mediation [46, 47]. For relationships that met all Baron and Kenny criteria, confidence intervals were estimated using the quasi-Bayesian Monte Carlo method with 1000 simulations [48]. The presence of exposure-mediator interactions within models were also determined to inform inferences about mediation [49]. A p-value below 0.05 (two-sided) was considered as statistically significant.
Results
A greater prevalence of high PHQ-9 scores, meeting thresholds for depression were observed in female participants, participants with an annual household income of less than $50,000, participants with a high school education or lower, participants who were single, divorced, or widowed, participants with diabetes, and participants with COPD (Table 1).
Similar to the comparison of depression status, greater proportions of GAD-7 scores that met thresholds for anxiety were observed in female participants, participants with an annual household income of less than $50,000, participants with a high school education or lower, participants who were single, divorced, or widowed, participants with type II diabetes, and participants with COPD (Table 2).
Mental health characteristics of the analytic sample are presented in Table 3. Depression defined by antidepressant use was observed in 9.2% of the analytic sample, and depression defined by PHQ-9 scores was observed in 5.6% of the sample. Self-reported lifetime depression diagnosis by a physician was observed for a higher proportion of the sample (21.5%). Anxiety defined by GAD-7 scores occurred in 4.6% of the sample, and comorbid anxiety and depression was observed in 2.7%.
Women had a higher mean GAD-7 score (2.8) and PHQ-9 score (3.2) compared to men (2.2 and 2.6, respectively) (Table 3). The proportion of women with a GAD-7 score over the threshold of 10 points indicating anxiety (5.4%) was also higher than men (3.6%) (Table 3). This disparity was also reflected in the proportion of women with depression (6.3%, compared to 4.7% for men), the proportion of women diagnosed with depression by a physician in their lifetime (26.0%, compared to 16.1% for men), proportion of women using antidepressants (12.2%, compared to 5.8% for men), and proportion of women with comorbid depression and anxiety (3.2%, compared to 2.1% for men) (Table 3).
Depression, anxiety, and risk of cancer
The median (min, max) follow-up time for all cancers was 3.5 years (182 days, 8.5 years) for men and 3.4 years (181 days, 8.4 years) for women. Among all three depression definitions, the risk estimates for men and women tended to be above 1.0, although, generally associations were statistically non-significant (Table 4). Similarly, no associations were observed between PHQ-9 scores on a continuous scale and risk of cancer (S1, S2 Tables in S1 File). The composite measure of depression was the only notable exception, but risk estimates were attenuated in Model 3 with additional adjustment for lifestyle behaviours. Anxiety was not associated with cancer risk in men or women (Table 4 and S1 Table in S1 File). Risk estimates for women tended to be higher than those for men for anxiety (GAD-7), depression (PHQ-9), and comorbid anxiety and depression. Estimates however, did not reach statistical significance with the exception of Model 1 for comorbid anxiety and depression among women.
None of the conceptualizations of depression, were associated with prostate cancer risk (Table 5 and S4 Table). Associations between anxiety, and comorbid anxiety and depression with prostate cancer are not presented due to low events (N = 9 and N = 5, respectively), but were similarly null.
Depression—assessed by PHQ-9—was associated with an increased risk of lung cancer among women in Model 1: HR: 1,88 (95% CI 1.15–3.09), and Model 2: HR: 1.75 (95% CI 1.06–2,88) (Table 6 and S5 Table). Comorbid anxiety and depression were also associated with increased risk of lung cancer among women in Model 1; HR: 2.32 (95% CI: 1.28–4.20) and Model 2: HR: 2.11 (1.16–3.85). Additional adjustment for lifestyle variables in Model 3 attenuated these associations. Anxiety was associated with increased risk of lung cancer in women (Table 6). Associations remained with full adjustment for confounders (Model 3; HR: 1.67 (95% CI 1.01–2.76).
Mediation effect of lifestyle factors
The depression- or anxiety-cancer associations with a statistically significant direct effect are shown in Table 7. Smoking status was shown to partially mediate the relationship between depression (PHQ-9), anxiety, and co-morbid depression and anxiety on lung cancer in women by a proportion of 27%, 18%, and 26%, respectively. No interactions between exposures (depression, anxiety, and co-morbid depression) and mediators were observed among women, with the exception of smoking status.
Mediation effects for men were examined for only depression and all cancer as this was the only relationship that was statistically significant. Smoking status was shown to partially mediate the relationship between depression (PHQ-9, antidepressant use, or self-report of lifetime physician diagnosis) on all cancer in men by a proportion of 17%. No interactions between depression and any mediator were observed among men.
Discussion
Findings from this large study of Canadians from Quebec suggest that there may be a modest positive association between mental health disorders and risk of all cancers and lung cancer. These relationships persisted with adjustment for sociodemographic and health variables but were generally attenuated with adjustment for lifestyle factors, showing both the importance of adjustment for these factors in future studies and further underscoring the importance of healthy lifestyles for cancer prevention. Further, smoking may be an important and modifiable health behaviour that contributes in part to relationships between depression (PHQ-9), anxiety (GAD-7), and co-morbid depression and anxiety and lung cancer risk in women. Similarly in men, findings suggest that of the lifestyle variables explored in our study, smoking may be the largest contributor to relationships between depression (PHQ-9, antidepressant use, or lifetime self-report of physician diagnosis) and all cancer incidence.
Evidence on the relationship between mental health disorders and cancer risk is inconsistent; with some studies reporting significant positive relationships for some cancer subtypes but not others, and other studies presenting non-significant findings [8, 9, 50, 51]. Differences between our research and previous studies might reflect the differences in international screening practices, different lifestyle factors and assessment of lifestyle factors, varying methods of ascertaining cancer status (e.g., self-report and cancer registry) and mental health status (e.g., DSM criteria, questionnaire scores, and administrative health databases). For example, Gross et al. conducted a longitudinal study with up to 24 years of follow-up, and found that there was a significant positive association between depression and all cancer risk (HR: 1.9, 95% CI: 1.2, 3.0) [50]. However, they did not adjust for any health behaviours other than smoking, which may obscure true associations due to confounding [50]. Consistent with our results, O’Neill et al. found no significant association between depression and all cancer risk, but also found that risk estimates were higher in women than men [51].
The present study primarily focussed on mental health exposures as presence or absence which may not capture the range of symptoms or specific aspects of mental health that are relevant for cancer. A cohort study in Korea found inconsistent results within their research on mental health disorders and prostate cancer risk. In a dose-response analysis, mild depression was significantly associated with subsequent prostate cancer risk, while moderate and severe depression were not [52]. Interestingly, in a separate assessment of minor and major depression, only minor depression was significantly associated with increased prostate cancer risk [52]. These studies may suggest that mild forms of depression, which would not be captured in our study, may be associated with prostate cancer risk. However, our additional analyses which examined PHQ-9 scores and GAD-7 scores on a continuous scale were consistent with the main analyses of presence/absence of anxiety and depression.
The results of our mediation analysis are consistent with the findings of Trudel-Fitzgerald et al., who found that lifetime pack-years of smoking mediated 38% of the overall association between depressive symptoms and lung cancer risk in women [53]. Since smoking is the predominant risk factor for lung cancer [54, 55] this finding is biologically plausible. To our knowledge, there is no comparable study that assessed this relationship in a cohort with men. A study by Burns et al. assessed the relationship between mental health difficulties, which was defined as any emotional, nervous, or psychiatric problems, and smoking-related diseases, which included respiratory disease, cardiovascular disease, or any smoking-related cancer [56]. They found that smoking had no significant mediating role in the relationship between mental health difficulties and smoking-related diseases [56]. However, this study grouped different mental health difficulties and smoking-related diseases together as a single exposure and outcome, limiting its comparability [56]. Given the limited number of studies that have assessed this relationship, additional research is warranted. Future studies that consider a wider range of potential mediators may help to provide insight on mental health-cancer links since smoking explained less than a third of the associations in this study.
A strength of this study was the large cohort size that included men and women reporting a wide range of income and education, which increases the generalizability of the findings [27]. A further strength was the collection of information on mental health disorders at baseline using standard validated questionnaires, along with detailed information about health behaviours using self-reported and measured data (BMI). Previous studies have seldom considered health behaviours as potential confounders in analyses. The availability of multiple variables related to depression at baseline allowed for conceptualizations of depression that may capture different dimensions or chronicity of depression. For example, anti-depressive medication use may reflect more severe or longer depressive symptoms versus the PHQ-9 which reflects depressive symptoms in the prior two weeks. The general consistency of findings across different depression definitions increases the confidence in our results.
Limitations of this study include the short overall follow-up time, which given the long latency of most cancers may not have accurately captured the association between mental health disorders and cancer incidence. The short follow-up was also reflected in a modest number of incident cancers. This may have obscured more modest associations between mental health and cancer risk in the analyses of cancer sites and further stratification by sex. Although individuals with an incident cancer in the first six months of follow-up were excluded, it is possible that participants had asymptomatic or early stages of cancer that were not captured at baseline. Another limitation was the limited availability on incidence of cancer sites beyond prostate and lung cancer due to restrictions on data availability at the time of our analysis. Mental health variables and lifestyle factors were measured simultaneously at baseline within the CARTaGENE cohort. Assessment of exposure and mediator variables at the same time point weakens the argument for causation in the mediation analysis, which limits the robustness of findings from this part of the study [57]. Caution is also warranted in the interpretation of direct and indirect effects of smoking in the mediation analysis in women, given the presence of an exposure-mediator interaction [49].
Additional studies which consider cancers that are strongly associated with lifestyle factors such as breast and colorectal cancer [54, 58] may help further elucidate potential mediators of mental health disorders and cancer risk. Other cancer outcomes, such as mortality and survival, are also of interest for future research. Similar to cancer incidence, there is a reasonable body of evidence suggesting a positive association between mental health disorders and cancer outcomes [59, 60]. Based on observations from previous literature and results from the present study, future research investigating this relationship should aim to stratify by sex, assess lifestyle behaviours, and explore alternative conceptualizations of mental health.
Conclusion
This study demonstrated a modest positive association between mental health disorders and lung cancer risk in women, however only anxiety and lung cancer in women remained significant with full adjustment of sociodemographic, health and lifestyle factors. The mediation analysis found that smoking may contribute to associations between anxiety, co-morbid depression and anxiety, and depression with incident lung cancer in women, and associations between depression with all cancers in men. This provides further emphasizes the importance of ongoing public health efforts to reduce smoking.
Supporting information
S1 File. Supporting information contains all the S1-S7 Tables.
https://doi.org/10.1371/journal.pone.0281588.s001
(DOCX)
References
- 1.
Canadian Cancer Society’s Advisory Committee on Cancer [Internet]. Canadian Cancer Society. [cited 2021 May 10]. Available from: https://action.cancer.ca/en/research/cancer-statistics/cancer-statistics-at-a-glance
- 2.
Government of Canada SC. Leading causes of death, total population, by age group [Internet]. 2018 [cited 2020 Jun 19]. Available from: https://www150.statcan.gc.ca/t1/tbl1/en/tv.action?pid=1310039401
- 3.
Smetanin P, Stiff D, Briante C, Adair CE, Ahmad S, Khan M. The Life and Economic Impact of Major Mental Illnesses in Canada: 2011 to 2041 [Internet]. Risk Analytica, on behalf of the Mental Health Commission of Canada; 2011. Available from: https://www.mentalhealthcommission.ca/sites/default/files/MHCC_Report_Base_Case_FINAL_ENG_0_0.pdf
- 4.
Government of Canada SC. Mental and substance use disorders in Canada [Internet]. 2013 [cited 2021 May 20]. Available from: https://www150.statcan.gc.ca/n1/pub/82-624-x/2013001/article/11855-eng.htm
- 5. Correll CU, Solmi M, Veronese N, Bortolato B, Rosson S, Santonastaso P, et al. Prevalence, incidence and mortality from cardiovascular disease in patients with pooled and specific severe mental illness: a large-scale meta-analysis of 3,211,768 patients and 113,383,368 controls. World Psychiatry. 2017;16(2):163–80. pmid:28498599
- 6. De Hert M, Detraux J, Vancampfort D. The intriguing relationship between coronary heart disease and mental disorders. Dialogues Clin Neurosci. 2018 Mar;20(1):31–40. pmid:29946209
- 7. Das‐Munshi J, Ashworth M, Dewey ME, Gaughran F, Hull S, Morgan C, et al. Type 2 diabetes mellitus in people with severe mental illness: inequalities by ethnicity and age. Cross-sectional analysis of 588 408 records from the UK. Diabet Med. 2017;34(7):916–24. pmid:27973692
- 8. Ahn HK, Bae JH, Ahn HY, Hwang IC. Risk of cancer among patients with depressive disorder: a meta-analysis and implications. Psychooncology. 2016;25(12):1393–9. pmid:26810736
- 9. Jia Y, Li F, Liu YF, Zhao JP, Leng MM, Chen L. Depression and cancer risk: a systematic review and meta-analysis. Public Health. 2017 Aug;149:138–48. pmid:28641155
- 10. Gibson-Smith D, Bot M, Brouwer IA, Visser M, Penninx BWJH. Diet quality in persons with and without depressive and anxiety disorders. J Psychiatr Res. 2018 Nov 1;106:1–7.
- 11. Rahe C, Baune BT, Unrath M, Arolt V, Wellmann J, Wersching H, et al. Associations between depression subtypes, depression severity and diet quality: cross-sectional findings from the BiDirect Study. BMC Psychiatry. 2015 Mar 4;15(1):38. pmid:25886444
- 12. Teasdale SB, Ward PB, Samaras K, Firth J, Stubbs B, Tripodi E, et al. Dietary intake of people with severe mental illness: systematic review and meta-analysis. Br J Psychiatry. 2019 May;214(5):251–9. pmid:30784395
- 13. Firth J, Rosenbaum S, Stubbs B, Gorczynski P, Yung AR, Vancampfort D. Motivating factors and barriers towards exercise in severe mental illness: a systematic review and meta-analysis. Psychol Med. 2016 Oct;46(14):2869–81. pmid:27502153
- 14. Stubbs B, Williams J, Gaughran F, Craig T. How sedentary are people with psychosis? A systematic review and meta-analysis. Schizophr Res. 2016 Mar 1;171(1):103–9. pmid:26805414
- 15. Stubbs B, Vancampfort D, Rosenbaum S, Ward PB, Richards J, Soundy A, et al. Dropout from exercise randomized controlled trials among people with depression: A meta-analysis and meta regression. J Affect Disord. 2016 Jan 15;190:457–66. pmid:26551405
- 16. Haynes JC, Farrell M, Singleton N, Meltzer H, Araya R, Lewis G, et al. Alcohol consumption as a risk factor for anxiety and depression: results from the longitudinal follow-up of the National Psychiatric Morbidity Survey. Br J Psychiatry J Ment Sci. 2005 Dec;187:544–51. pmid:16319407
- 17. Jansson-Fröjmark M, Lindblom K. A bidirectional relationship between anxiety and depression, and insomnia? A prospective study in the general population. J Psychosom Res. 2008 Apr;64(4):443–9. pmid:18374745
- 18. Azevedo Da Silva M, Singh-Manoux A, Brunner EJ, Kaffashian S, Shipley MJ, Kivimäki M, et al. Bidirectional association between physical activity and symptoms of anxiety and depression: the Whitehall II study. Eur J Epidemiol. 2012 Jul;27(7):537–46. pmid:22623145
- 19.
World Cancer Research Fund. Cancer Prevention Recommendations [Internet]. WCRF International. [cited 2022 Jan 27]. Available from: https://www.wcrf.org/diet-and-cancer/cancer-prevention-recommendations/
- 20. Poirier AE, Ruan Y, Volesky KD, King WD, O’Sullivan DE, Gogna P, et al. The current and future burden of cancer attributable to modifiable risk factors in Canada: Summary of results. Prev Med. 2019 May 1;122:140–7. pmid:31078167
- 21. Pattyn E, Verhaeghe M, Bracke P. The gender gap in mental health service use. Soc Psychiatry Psychiatr Epidemiol. 2015 Jul;50(7):1089–95. pmid:25788391
- 22. Eaton NR, Keyes KM, Krueger RF, Balsis S, Skodol AE, Markon KE, et al. An invariant dimensional liability model of gender differences in mental disorder prevalence: evidence from a national sample. J Abnorm Psychol. 2012 Feb;121(1):282–8. pmid:21842958
- 23. Jemal A, Miller KD, Ma J, Siegel RL, Fedewa SA, Islami F, et al. Higher Lung Cancer Incidence in Young Women Than Young Men in the United States. N Engl J Med. 2018 May 24;378(21):1999–2009. pmid:29791813
- 24. White A, Ironmonger L, Steele RJC, Ormiston-Smith N, Crawford C, Seims A. A review of sex-related differences in colorectal cancer incidence, screening uptake, routes to diagnosis, cancer stage and survival in the UK. BMC Cancer. 2018 Dec;18(1):906. pmid:30236083
- 25. Varì R, Scazzocchio B, D’Amore A, Giovannini C, Gessani S, Masella R. Gender-related differences in lifestyle may affect health status. Ann Ist Super Sanita. 2016;52(2):158–66. pmid:27364389
- 26. Xie L, Semenciw R, Mery L. Cancer incidence in Canada: trends and projections (1983–2032). Health Promot Chronic Dis Prev Can Res Policy Pract. 2015 Mar;35(Suppl 1):1–187. pmid:26011811
- 27. Dummer TJB, Awadalla P, Boileau C, Craig C, Fortier I, Goel V, et al. The Canadian Partnership for Tomorrow Project: a pan-Canadian platform for research on chronic disease prevention. CMAJ. 2018 Jun 11;190(23):E710–7. pmid:29891475
- 28. Awadalla P, Boileau C, Payette Y, Idaghdour Y, Goulet JP, Knoppers B, et al. Cohort profile of the CARTaGENE study: Quebec’s population-based biobank for public health and personalized genomics. Int J Epidemiol. 2013 Oct 1;42(5):1285–99. pmid:23071140
- 29. Tonelli M, Wiebe N, Fortin M, Guthrie B, Hemmelgarn B, James M, et al. Methods for identifying 30 chronic conditions: Application to administrative data Healthcare Information Systems. BMC Med Inform Decis Mak. 2015 Apr 17;15:31.
- 30. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001 Sep;16(9):606–13. pmid:11556941
- 31. Manea L, Gilbody S, McMillan D. Optimal cut-off score for diagnosing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. CMAJ Can Med Assoc J J Assoc Medicale Can. 2012 Feb 21;184(3):E191–196. pmid:22184363
- 32.
WHO Collaborating Centre for Drugs Statistics Methodology. ATC/DDD Index. Available from: https://www.whocc.no/atc_ddd_index/. Cited 2022 Jan 19.
- 33. Wong J, Motulsky A, Eguale T, Buckeridge DL, Abrahamowicz M, Tamblyn R. Treatment Indications for Antidepressants Prescribed in Primary Care in Quebec, Canada, 2006–2015. JAMA. 2016 May 24;315(20):2230–2. pmid:27218634
- 34. Spitzer RL, Kroenke K, Williams JBW, Löwe B. A Brief Measure for Assessing Generalized Anxiety Disorder: The GAD-7. Arch Intern Med. 2006 May 22;166(10):1092. pmid:16717171
- 35. Giovannucci E, Harlan DM, Archer MC, Bergenstal RM, Gapstur SM, Habel LA, et al. Diabetes and cancer: a consensus report. Diabetes Care. 2010 Jul;33(7):1674–85. pmid:20587728
- 36. Li N, Huang Z, Zhang Y, Sun H, Wang J, Zhao J. Increased cancer risk after myocardial infarction: fact or fiction? A systemic review and meta-analysis. Cancer Manag Res. 2019;11:1959–68. pmid:30881121
- 37. Gilham K, Gu Q, Dummer TJB, Spinelli JJ, Murphy RA. Diet Quality and Neighborhood Environment in the Atlantic Partnership for Tomorrow’s Health Project. Nutrients. 2020 Oct 21;12(10):3217. pmid:33096731
- 38. Gu Q, Dummer TBJ, Spinelli JJ, Murphy RA. Diet Quality among Cancer Survivors and Participants without Cancer: A Population-Based, Cross-Sectional Study in the Atlantic Partnership for Tomorrow’s Health Project. Nutrients. 2019 Dec 11;11(12):3027. pmid:31835839
- 39. Hirshkowitz M, Whiton K, Albert SM, Alessi C, Bruni O, DonCarlos L, et al. National Sleep Foundation’s updated sleep duration recommendations: final report. Sleep Health. 2015 Dec;1(4):233–43. pmid:29073398
- 40. Craig CL, Marshall AL, Sjöström M, Bauman AE, Booth ML, Ainsworth BE, et al. International physical activity questionnaire: 12-country reliability and validity. Med Sci Sports Exerc. 2003 Aug;35(8):1381–95. pmid:12900694
- 41. Wang DD, Li Y, Bhupathiraju SN, Rosner BA, Sun Q, Giovannucci EL, et al. Fruit and Vegetable Intake and Mortality: Results From 2 Prospective Cohort Studies of US Men and Women and a Meta-Analysis of 26 Cohort Studies. Circulation. 2021 Apr 27;143(17):1642–54. pmid:33641343
- 42. Buuren S van, Groothuis-Oudshoorn K. mice: Multivariate Imputation by Chained Equations in R. J Stat Softw. 2011;45(3). Available from: http://www.jstatsoft.org/v45/i03/. Cited 2018 Mar 14
- 43.
Therneau T. A package for survival analysis in R. 2021. Available from: https://rdrr.io/cran/survival/f/inst/doc/survival.pdf. Cited 2021 Apr 18.
- 44.
Canchola A, Stewart S, Center NCC, Bernstein L. Cox Regression Using Different Time Scales. Available from: https://www.lexjansen.com/wuss/2003/DataAnalysis/i-cox_time_scales.pdf. Cited 2021 April 13.
- 45. Tingley D, Yamamoto T, Hirose K, Keele L, Imai K. mediation: R Package for Causal Mediation Analysis. J Stat Softw. 2014;59(5). Available from: http://www.jstatsoft.org/v59/i05/. Cited 2021 Apr 18.
- 46. MacKinnon DP, Fairchild AJ, Fritz MS. Mediation Analysis. Annu Rev Psychol. 2007;58:593. pmid:16968208
- 47. Baron RM, Kenny DA. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J Pers Soc Psychol. 1986 Dec;51(6):1173–82. pmid:3806354
- 48. Yuan Y, MacKinnon DP. Bayesian mediation analysis. Psychol Methods. 2009;14(4):301–22. pmid:19968395
- 49. Valeri L, Vanderweele TJ. Mediation analysis allowing for exposure-mediator interactions and causal interpretation: theoretical assumptions and implementation with SAS and SPSS macros. Psychol Methods. 2013 Jun;18(2):137–50. pmid:23379553
- 50. Gross AL, Gallo JJ, Eaton WW. Depression and cancer risk: 24 years of follow-up of the Baltimore Epidemiologic Catchment Area sample. Cancer Causes Control. 2010 Feb;21(2):191–9. pmid:19885645
- 51. O’Neill S, Posada-Villa J, Medina-Mora ME, Al-Hamzawi AO, Piazza M, Tachimori H, et al. Associations between DSM-IV mental disorders and subsequent self-reported diagnosis of cancer. J Psychosom Res. 2014 Mar;76(3):207–12. pmid:24529039
- 52. Chang HY, Keyes KM, Mok Y, Jung KJ, Shin YJ, Jee SH. Depression as a risk factor for overall and hormone-related cancer: The Korean cancer prevention study. J Affect Disord. 2015 Mar 1;173:1–8. pmid:25462388
- 53. Trudel-Fitzgerald C, Zevon ES, Kawachi I, Tucker-Seeley RD, Kubzansky LD. Depression, smoking, and lung cancer risk over 24 years among women. Psychol Med. 2022 Oct;52(13):2510–2519. pmid:33267930
- 54. Canadian Population Attributable Risk of Cancer (ComPARe). Percentage of cancers that are preventable in Canada. 2019. Available from: https://prevent.cancer.ca. Cited 2021 Apr 16.
- 55. Carbone D. Smoking and cancer. Am J Med. 1992 Jul 15;93(1, Supplement 1):S13–7.
- 56. Burns A, Strawbridge JD, Clancy L, Doyle F. Exploring smoking, mental health and smoking-related disease in a nationally representative sample of older adults in Ireland–A retrospective secondary analysis. J Psychosom Res. 2017 Jul 1;98:78–86. pmid:28554376
- 57. Lapointe-Shaw L, Bouck Z, Howell NA, Lange T, Orchanian-Cheff A, Austin PC, et al. Mediation analysis with a time-to-event outcome: a review of use and reporting in healthcare research. BMC Med Res Methodol. 2018 Oct 29;18(1):118. pmid:30373524
- 58. Chen Z, Wang PP, Woodrow J, Zhu Y, Roebothan B, Mclaughlin JR, et al. Dietary patterns and colorectal cancer: results from a Canadian population-based study. Nutr J. 2015 Jan 15;14(1):8. pmid:25592002
- 59. Pinquart M, Duberstein PR. Depression and cancer mortality: a meta-analysis. Psychol Med. 2010 Nov;40(11):1797–810. pmid:20085667
- 60. Yang L, Korhonen K, Moustgaard H, Silventoinen K, Martikainen P. Pre-existing depression predicts survival in cardiovascular disease and cancer. J Epidemiol. 2018;72(7):617–22. pmid:29483141