https://orcid.org/0000-0002-1721-502X
Abstract
Background
Poorer survival in cancer patients with vs. without comorbidity has been reported for various cancer sites. For patients with colorectal cancer (CRC), limited data are available so far.
Methods
Patients with CRC diagnosed between 2010 and 2018 were identified in a health claims database covering 20% of the German population. We assessed the prevalence of comorbidities at cancer diagnosis and categorized the patients into the groups: ‘none’, ‘somatic only’, ‘mental only’ or ‘both’ types of comorbidities. Hazard ratios (HR, with 95% confidence intervals) for five-year overall survival were estimated by Cox proportional hazard models, adjusted for age, sex and stage at diagnosis (advanced vs. non-advanced).
Results
We included 92,991 patients (females: 49.1%, median age: 72 years) with a median follow-up of 30 months. The proportions assigned to the groups ‘none’, ‘somatic only’, ‘mental only’ or ‘both’ were 24.7%, 65.5%, 1.4% and 8.4%. Overall, 32.8% of the patients died during follow-up. Compared to patients without comorbidities (‘none’), the adjusted HR regarding death from any cause was 1.11 (95% CI: 1.07–1.14) in the group ‘somatic only’, 1.74 (95% CI: 1.58–1.92) in the group ‘mental only’ and 1.92 (95% CI: 1.84–2.00) in the group ‘both’. For patients with ‘mental only’ comorbidities, the adjusted HR was higher in males than in females (HR = 2.19, 95% CI: 1.88–2.55 vs. HR = 1.55, 95% CI: 1.37–1.75).
Conclusions
Our results suggest that patients with CRC and with mental comorbidities, particularly males, have a markedly lower overall survival compared to those without any or only somatic comorbidities.
Background
There are several studies reporting lower survival among cancer patients who suffer from comorbidities as compared to those who do not. This has been shown for various cancer entities, including head and neck cancer [Citation1], breast cancer [Citation2–4], prostate cancer [Citation5], lung carcinoma [Citation6,Citation7], and colorectal cancer (CRC) [Citation8,Citation9]. Various reasons for this difference have been discussed, such as a higher frailty burden [Citation8], a more advanced stage at cancer diagnosis [Citation10], less intense cancer treatment or early discontinuation of cancer treatment [Citation5,Citation11].
None of these studies, however, provided results separately for patients with only somatic, only mental or both types of comorbidity. Comparing the survival of these groups indirectly between studies is hampered due to heterogeneity in the study designs and the cancers under study. The underlying study populations were either restricted regarding sex [Citation12] or age range [Citation13,Citation14], had a strong focus on mental comorbidity [Citation10,Citation15] or even on only a few, specific mental disorders [Citation16] or considered comorbidities as a whole without distinguishing between the comorbidity groups mentioned above.
For CRC, which is among the most common cancers in men and women [Citation17], the evidence on survival according to type of comorbidity is limited as well. Iversen et al. [Citation9] used hospital discharge data to investigate the association between survival and comorbidity status in patients with CRC, defined based on the Charlson Comorbidity Index. They found decreased survival in patients with higher Charlson scores, but did not provide separate results for the type of comorbidity (i.e., mental, somatic or both). Similarly, based upon a meta-analysis of 13 cohort studies on patients with CRC, Boakye et al. [Citation8] reported significantly lower overall and CRC-specific survival with increasing severity of comorbidity, but the results were not stratified by the type of comorbidity. Other studies on CRC had a strong focus on mental comorbidities but did not consider somatic comorbidities as a separate group [Citation13,Citation14,Citation18].
To shed further light on this topic, we aimed to describe and compare the overall survival of male and female patients with CRC stratified by the presence of somatic comorbidity only, mental comorbidity only, both types of comorbidity or no comorbidity.
Methods
Data source
We used the German Pharmacoepidemiological Research Database (GePaRD). GePaRD is based on claims data from four statutory health insurance providers in Germany and currently includes information on approximately 25 million persons who have been insured with one of the participating providers since 2004 or later. In addition to demographic data, GePaRD contains information on drug dispensations as well as on outpatient (i.e., from general practitioners and specialists) and inpatient services and diagnoses. Per data year, there is information on approximately 20% of the general population and all geographical regions of Germany are represented [Citation19].
To identify incident CRC cases in GePaRD, we used an algorithm that selects, in the first step, patients with at least one inpatient discharge diagnosis code for CRC (ICD C18-C20), which have a high validity. Patients with exclusively outpatient diagnosis codes for CRC were only classified as CRC cases if regular surveillance examinations as expected according to guidelines after endoscopic removal of pT1 cancers were recorded. To ensure that the CRC cases were incident and not prevalent, we only included patients with continuous health insurance of at least three years before the first CRC diagnosis code in GePaRD. Next, the algorithms assigned a diagnosis date to included CRC cases. This was the date of hospitalization if CRC was first coded in the inpatient setting; if the first CRC code was in the outpatient setting, the date of colonoscopy was used or the mid of the quarter if there was no colonoscopy (outpatient diagnosis codes in German claims data are only recorded on a quarterly basis). Finally, CRC diagnoses were classified into advanced and non-advanced stages considering codes for lymph node involvement and metastases (C77-C79) documented in the quarter of the first CRC code or in the following quarter. The CRC incidence and stage distribution (advanced vs. non-advanced) determined based on this algorithm has been shown to agree very well with cancer registry data [Citation20].
Study design
We included patients diagnosed with CRC between January 1, 2010 and December 31, 2018. We excluded patients if there was no or inconsistent information on sex or birth year or if they did not live in Germany. We followed up the patients from the date of diagnosis (cohort entry) until the date of (a) death, (b) the end of continuous health insurance (gap >30 d) or (c) the end of the observation period (December 31, 2018), whichever occurred first.
We assessed the presence or absence of comorbidities based on codes recorded in the five years prior to cohort entry. To define the comorbidities, we mostly used algorithms ensuring a high specificity by taking into account either inpatient codes or information on medication [Citation21]. For more details on the operationalizations of the comorbidities, please see Suppl. Table S1. Based on their comorbidity status, patients were allocated to one of the four following disjunct categories: (A) No somatic comorbidity and no mental comorbidity (‘none’), (B) any somatic comorbidity but no mental comorbidity (‘somatic only’), (C) no somatic comorbidity but any mental comorbidity (‘mental only’) and (D) both somatic and mental comorbidity (‘both’). Patients with codes for substance disorders but no other comorbidity (i.e., who did not fulfill the criteria for group B, C or D) were assigned to group A. In a sensitivity analysis, survival in these patients was assessed separately. In additional sensitivity analyses, we divided the group ‘somatic only’ into the following two subgroups and assessed survival in both subgroups separately: (A) person in group ‘somatic only’ receiving only antihypertensive treatment or lipid-modifying agents (i.e., no other somatic comorbidity), (B) remaining persons in the group ‘somatic only’.
Statistical analyses
Depending on the data levels, summary statistics consisted of counts, percentages, means and standard deviations (SD), medians, and quartiles, where appropriate. Time to cohort exit due to death was assessed based on Kaplan–Meier analyses. The survival probabilities were estimated using Cox regression modeling (adjusted for age, sex, stage at diagnosis, and presence of dementia as comorbidity), and the results expressed as Hazard Ratios (HR) with 95% confidence intervals (CI). To assess the assumption of proportional hazards, we employed the plotting of standardised Schoenfeld residuals to examine potential time-dependent effects [Citation22]. The analysis revealed no violation of the proportional hazards assumption. All statistical analyses were conducted using SAS 9.4 and R version 4.2.0 [Citation23] including the package survival 3.3-1 for the survival analyses.
Ethics
In Germany, the utilization of health insurance data for scientific research is regulated by the Code of Social Law. All involved health insurance providers as well as the German Federal Office for Social Security and the Senator for Health, Women and Consumer Protection in Bremen as their responsible authorities approved the use of GePaRD data for this study. Informed consent for studies based on claims data is required by law unless obtaining consent appears unacceptable and would bias results, which was the case in this study. According to the Ethics Committee of the University of Bremen studies based on GePaRD are exempt from institutional review board review.
Results
Patient characteristics
Overall, we included 92,991 patients with CRC with a mean age of 70.0 years. About half of the patients (49.9%) were female. The median follow-up time after CRC diagnosis was 29.7 months (903 d). About one third (32.6%) of the patients died during follow-up. Patients who died during follow-up were significantly older at the time of the CRC diagnosis than the remaining patients (75.0 vs. 67.7 years). In 43.5% of male and 40.9% of female patients, CRC was diagnosed at an advanced stage ().
Table 1. Characteristics of included patients with colorectal cancer, stratified by sex.
Comorbidity status
shows the prevalence of the various comorbidities as well as the distribution by comorbidity status (‘somatic only’, ‘mental only’, ‘both’, ‘none’) stratified by sex. Overall, 76.3% of males and 71.7% of females suffered from at least one somatic comorbidity. Among somatic comorbidities, antihypertensive treatment showed the highest prevalence (males: 62.2%, females: 56.5%), followed by treatment with lipid-modifying agents (males: 23.6%, females: 16.6%). The prevalence of at least one mental comorbidity was 7.5% in males and 12.1% in females. The prevalence of at least one code for a substance disorder was 15.0% in males and 11.1% in females. It was 8.7% in patients without comorbidities, 13.5% in the group ‘somatic only’, 18.5% in the group ‘mental only’ and 21.4% in those with both types of comorbidities.
Table 2. Somatic and mental comorbidity status of patients in the colorectal cancer cohort (N = 92.991) up to five years prior to cohort entry, stratified by sex.
Overall, 24.7% of all patients had no somatic and no mental comorbidity prior to CRC diagnosis, 65.5% had only somatic comorbidities, 1.4% had only mental comorbidities and 8.4% were afflicted by comorbidities from both categories. Patient characteristics stratified by comorbidity status are shown in . The proportion of males was highest in the category ‘somatic only’. The mean age was below 70 years in the categories ‘none’ and ‘mental only’, while it was 72 years and 75.8 years, respectively, in the categories ‘somatic only’ and ‘both’. The proportion diagnosed at an advanced stage was 46.3% in the category ‘none’ while it was 5%–6% points lower in the other categories. Among those with advanced stage, the proportion with distant metastases was highest in the categories ‘mental only’ (64%) and ‘both’ (65%) as compared to the other two categories (∼59%).
Table 3. Characteristics of included patients with colorectal cancer, stratified by comorbidity status.
Survival analyses
displays the survival probabilities (Kaplan–Meier curves) during the follow-up period, stratified by comorbidity status and sex. Male as well as female patients from the category ‘none’ had the highest survival probability, whereas those from the category ‘both’ had the lowest. shows the results of the Cox regression analyses. Compared to the category ‘none’, the HR (aHR, Model B, i.e., adjusted for age, sex and stage at diagnosis) of dying during follow-up was 1.108 for the category ‘somatic only’, 1.743 for the category ‘mental only’ and 1.919 for the category ‘both’. After additionally adjusting for dementia (see , Model C), the aHR were further reduced while the observed patterns remained stable.
Figure 1. Survival in the colorectal cancer cohort, stratified by comorbidity status in (a) male and (b) female patients.
Table 4. Hazard ratios (with 95% confidence intervals) regarding survival in patients with colorectal cancer with and without mental and somatic comorbidities, stratified by sex.
When stratifying the analyses by sex, the aHR remained at 1.109 (men) and 1.107 (women) for the category ‘somatic only’. For the category ‘mental only’, it was 2.189 (Model C: 1.768) in men and 1.546 (Model C: 1.246) in women. For the category ‘both’, it was 2.078 (Model C: 1.613) in men and 1.806 (Model C: 1.365) in women, i.e., for these two categories, the effect of comorbidity was modified by sex. This was confirmed in separate Cox regression models comparing the survival probability of males to the survival probability of females (reference) for each comorbidity category, adjusted for age and cancer stage: ‘mental only’: aHR = 1.397 (95%CI: 1.149–1.699); ‘both’: 1.091 (95% CI: 1.004–1.185). When additionally adjusting for dementia status, these associations changed to aHR = 1.293 (95% CI: 1.063–1.573) in the ‘mental only’ group and to aHR = 1.029 (95% CI: 0.947–1.119) in the ‘both group’.
The sensitivity analysis comparing patients with CRC and with substance disorders only (i.e., no other comorbidities considered in this study) and those without comorbidity showed a rather similar survival in both groups (see Suppl. Figure 1). In the sensitivity analysis splitting up the group ‘somatic only’, the aHR was 0.881 (95% CI: 0.844–0.920) for those receiving only antihypertensive or lipid-modifying treatment (i.e., no other somatic comorbidity) and 1.175 (95% CI: 138–1.212) for the remaining patients in the group ‘somatic only’ (see Suppl. Table S3).
Discussion
This is the largest study so far on the association between CRC survival and comorbidity and the first study to contrast survival of patients with CRC according to the presence of somatic comorbidity only, mental comorbidity only, both types of comorbidity and without comorbidity. Our study including 92,991 patients with CRC showed that in those with mental comorbidities, the overall risk of dying during follow-up was 70–90% higher compared to those without any of the comorbidities considered in our study. Stratification by sex showed that this difference was markedly higher in males than in females. Patients with somatic comorbidities only also had a higher risk of dying compared to those without comorbidity, but the difference was less pronounced (11% higher risk).
Boakye et al. systematically reviewed previous studies comparing the survival of patients with CRC according to the presence of comorbidity. In 85% of the 27 studies on long-term prognosis included in this review, the sample size was <15,000 (in 67% of studies N < 5000) [Citation8]. With such a sample size, stratified analyses as done in our study would lead to rather imprecise estimates, particularly because of the small group of patients with only mental comorbidities. In the larger studies included in this review, comorbidity was classified based on overall morbidity scores, so there were either no results stratified by type of comorbidity [Citation24–26] or they were restricted to univariate analyses of single comorbidities [Citation27]. Some studies with larger sample sizes specifically focused on mental comorbidities and therefore linked cancer registry data to other data sources with information on mental health, such as hospital records or health claims. For example, Bailargeon et al. [Citation13] using SEER data linked to Medicare data included 80,670 patients with CRC aged ≥67 years of whom 25.7% had a diagnosis of any mental disorder before cancer diagnosis. The adjusted overall risk of dying was 33% higher in those with vs. without mental disorder. Unlike Bailargeon et al. we used algorithms with a focus on a high specificity for the definition of mental disease to minimize bias due to misclassification, which might explain why the association was stronger in our study. Another study including 4,022 patients with CRC distinguished between two groups with mental comorbidity (A: patients with schizophrenia or bipolar affective disorder; B: others using mental health services). Compared to patients without mental comorbidity, the adjusted risk regarding death from CRC was 89% higher for group A and 25% higher for group B (similar results for death from any cause) [Citation14]. A study by Manderbacka et al. [Citation18] including 41,708 patients with CRC also compared patients without severe mental illness to different patient groups with mental illness. For those with a history of psychosis, the adjusted risk regarding death from CRC was 72% higher in men and 37% higher in women, i.e., sex was an effect modifier similar to our study.
In this regard, the strong association between mental comorbidity and survival after CRC diagnosis is consistent with the literature. Our study extends this knowledge by demonstrating—based on a direct comparison—that this association is much stronger than the association between somatic comorbidity and overall survival, particularly in men. This was even the case when we restricted the analysis to patients with more severe somatic comorbidities by excluding those from the group ‘somatic only’ who received only antihypertensive or lipid modifying drugs. The fact that survival was also substantially worse for the group ‘mental only’ and not only for the group ‘both’ further supports the relevance of mental comorbidities as a prognostic factor. This raises the question of underlying mechanisms. In a recently published review, Renzi et al. [Citation28] concluded that psychiatric disorders are potentially associated with a larger delay in cancer diagnosis, whereas patients with certain somatic conditions (e.g., hypertension) are more often diagnosed at an earlier cancer stage. However, our effect estimates as well as those mentioned above from other studies were adjusted for stage at diagnosis, so this is unlikely to be the key reason for the difference. Bailargeon et al. reported a higher probability of non-treatment among patients with CRC and with a pre-existing mental disorder but poorer survival was still observed after controlling for receipt of treatment [Citation13], i.e., treatment does not seem to be the only explanatory factor either.
The underlying mechanisms thus remain unclear, but taking into account that the impact of mental comorbidity on survival was more pronounced in male than in female patients may be useful to further reflect on potential other reasons. Dementia, which is inherently associated with an increased risk of death, is less common in men than in women [Citation29,Citation30], so this would not lead to the observed pattern. This was also demonstrated by our analyses. After adjustment for dementia, the hazard ratios were lower but the observed patterns regarding group differences did not change. Another explanation could be substance use disorders, which have been linked to poorer survival in cancer patients [Citation31,Citation32] and are more common in men than in women. In our sensitivity analysis, however, persons with substance use disorders showed hardly any difference in survival compared to patients without comorbidity (see Suppl. Figure S1). This could also be because capturing substance use disorders is suboptimal in claims data. Manderbacka et al. [Citation18] found an about 20% increased risk of death from CRC among patients with substance use disorders, but the point estimate was similar in men and women. On the other hand, sex differences regarding the impact of mental comorbidity on CRC survival might also be due to potential differences in the severity of recorded mental diseases. If differences in health seeking behavior between men and women led to a lower prevalence of diagnosed mental comorbidity in men while those diagnosed were more severe in men, such a pattern could also lead to sex differences as observed in our study. However, this is highly speculative and it is not clear whether the sex differences regarding the prevalence of mental comorbidities, reported by many studies [Citation30], could partly be due to diagnostic bias. Another hypothesis is that the sex differences regarding the impact of mental comorbidity on CRC survival could partly be due to the fact that men in general more often commit suicide than women [Citation33], but it was not possible to assess this in our database.
So overall, existing hypotheses are unlikely to fully explain the strong difference in survival between patients with CRC and with mental comorbidity and those with only somatic or no comorbidity. In addition, the reasons of the effect modification by sex remain unclear. Elucidating the reasons will be important to understand whether these differences in survival are modifiable and thus amenable to interventions (e.g., improved treatment of cancer or mental comorbidity).
In addition to the strengths and limitations already discussed, the following should be considered in the interpretation of our study. First, the options to link cancer registry and medical record data in Germany are very limited due to data privacy regulations. We therefore used claims data for the identification of patients with CRC and the ascertainment of comorbidities in these patients. Even though suboptimal (e.g., details on stage and histology, as usually available in cancer registry data, are lacking in health claims data), the validity of our case definition for incident CRC and also the rough stage classification are well supported by indirect comparison to cancer registry data [Citation34,Citation35]. The same applies to comorbidities; the information in claims data is not equivalent to medical records. Particularly in the outpatient setting, there is the problem of overcoding or unspecific coding. Furthermore, the indication of drug prescriptions is not directly recorded. We therefore used algorithms for the definition of comorbidities prioritizing a high specificity at the cost of a lower sensitivity, which is known to minimize bias in relative effect estimates. In the case of mental comorbidities, we partly classified patients only based on the type of treatment (antidepressants, antipsychotics) given that the indication of drug prescriptions is not directly recorded. To avoid reverse causality (e.g., treatment with antidepressants related to palliative cancer care) we defined comorbidities only based on information before cancer diagnosis. Furthermore, we used a long pre-observation period to minimize misclassification due to left truncation. Second, we focused on preexisting mental and somatic comorbidity but could not reasonably assess the prevalence of prior malignancies. According to a large population-based case-control study from Germany the prevalence of prior malignancies in patients with CRC is ∼11% [Citation36]. Interestingly, an analysis based on SEER data including 550,325 CRC patient did not find a worse survival among CRC patients with a history of a prior non-leukemic malignancy (corresponding to the vast majority of patients with a prior malignancy) [Citation37]. We therefore do not expect that consideration of prior malignancies would have had a relevant impact on our results. We do also not expect that additional consideration of comorbidities with a low prevalence (e.g., epilepsy or multiple sclerosis) would have altered our conclusion. Finally, it should be noted that the age and sex distribution of the CRC cases in our study is not fully representative of all CRC cases in Germany (higher proportion of females, slightly older). This was expected due to the age and sex structure of the populations covered by the health insurance providers providing data to GePaRD, but it is not a limitation of our study as we conduced all analyses stratified by sex and adjusted for age. In terms of representativeness of cancer care, we do not expect variation in the quality of cancer care between patients with colorectal cancer insured by statutory health insurance providers participating in GePaRD and other patients with colorectal cancer. In Germany, about 90% of the general population in Germany are covered by a statutory health insurance provider. Core characteristics of the German health insurance system are uniform access to all levels of care and free choice of providers. The intention of this system is to provide equal care to patients irrespective of their socioeconomic status. In line with this, we have previously shown that the drug dispensations in GePaRD are representative of drug dispensations among all patients with statutory health insurance in Germany [Citation38].
In conclusion, our study showed that patients with CRC and with mental comorbidities, particularly males, have a markedly lower overall survival compared to those without or with only somatic comorbidities. Neither our nor other studies suggest the existence of simple reasons explaining this finding (e.g., stage distribution). Identifying potential mechanisms will be important to understand whether these differences in survival are modifiable and thus amenable to interventions.
Authors contributions
OR wrote the first draft of the manuscript and conceptualized the statistical analyses. JV conducted the statistical analyses including sensitivity analyses and reviewed the draft of the manuscript. UH reviewed the draft of the manuscript and provided a critique.
Supplemental Material
Download MS Word (12.8 KB)Supplemental Material
Download JPEG Image (592.2 KB)
Supplemental Material
Download MS Word (25.8 KB)Supplemental Material
Download MS Word (26.9 KB)Supplemental Material
Download MS Word (28.6 KB)Acknowledgement
The authors would like to thank all statutory health insurance providers which provided data for this study, namely AOK Bremen/Bremerhaven, DAK-Gesundheit, Techniker Krankenkasse (TK), and hkk Krankenkasse.
Disclosure statement
All authors of this study are working at an independent, non-profit research institute, the Leibniz Institute for Prevention Research and Epidemiology – BIPS. Unrelated to this study, BIPS occasionally conducts studies financed by the pharmaceutical industry. Almost exclusively, these are post-authorization safety studies (PASS) requested by health authorities. The design and conduct of these studies as well as the interpretation and publication are not influenced by the pharmaceutical industry. The study presented was not funded by the pharmaceutical industry and was performed in line with the ENCePP Code of Conduct. The authors declare no conflict of interest
Additional information
Funding
References
- Schimansky S, Lang S, Beynon R, et al. Association between comorbidity and survival in head and neck cancer: results from head and neck 5000. Head Neck. 2019;41(4):1053–1062. doi: 10.1002/hed.25543.
- Land LH, Dalton SO, Jorgensen TL, et al. Comorbidity and survival after early breast cancer. A review. Crit Rev Oncol Hematol. 2012;81(2):196–205. doi: 10.1016/j.critrevonc.2011.03.001.
- Ording AG, Garne JP, Nystrom PMW, et al. Comorbid diseases interact with breast cancer to affect mortality in the first year after diagnosis-A Danish nationwide matched cohort study. PLOS One. 2013;8(10):e76013. doi: 10.1371/journal.pone.0076013.
- Woelfel IA, Fernandez LJ, Idowu MO, et al. A high burden of comorbid conditions leads to decreased survival in breast cancer. Gland Surg. 2018;7(2):216–227. doi: 10.21037/gs.2018.02.03.
- Bradley CJ, Dahman B, Anscher M. Prostate cancer treatment and survival evidence for men with prevalent comorbid conditions. Med Care. 2014;52(6):482–489. doi: 10.1097/MLR.0000000000000113.
- Luchtenborg M, Jakobsen E, Krasnik M, et al. The effect of comorbidity on stage- specific survival in resected non-small cell lung cancer patients. Eur J Cancer. 2012;48(18):3386–3395. doi: 10.1016/j.ejca.2012.06.012.
- Shieh SH, Probst JC, Sung FC, et al. Decreased survival among lung cancer patients with co-morbid tuberculosis and diabetes. BMC Cancer. 2012;12(1):174. doi: 10.1186/1471-2407-12-174.
- Boakye D, Rillmann B, Walter V, et al. Impact of comorbidity and frailty on prognosis in colorectal cancer patients: a systematic review and meta-analysis. Cancer Treat Rev. 2018;64:30–39. doi: 10.1016/j.ctrv.2018.02.003.
- Iversen LH, Norgaard M, Jacobsen J, et al. The impact of comorbidity on survival of danish colorectal cancer patients from 1995 to 2006-A population-based cohort study. Dis Colon Rectum. 2009;52(1):71–78. doi: 10.1007/DCR.0b013e3181974384.
- Kisely S, Crowe E, Lawrence D. Cancer-related mortality in people with mental illness. JAMA Psychiatry. 2013;70(2):209–217. doi: 10.1001/jamapsychiatry.2013.278.
- Boakye D, Nagrini R, Ahrens W, et al. The association of comorbidities with administration of adjuvant chemotherapy in stage III Colon cancer patients: a systematic review and meta-analysis. Ther Adv Med Oncol. 2021;13:1758835920986520. doi: 10.1177/1758835920986520.
- Batty GD, Whitley E, Gale CR, et al. Impact of mental health problems on case fatality in male cancer patients. Br J Cancer. 2012;106(11):1842–1845. doi: 10.1038/bjc.2012.150.
- Baillargeon J, Kuo YF, Lin YL, et al. Effect of mental disorders on diagnosis, treatment, and survival of older adults with Colon cancer. J Am Geriatr Soc. 2011;59(7):1268–1273. doi: 10.1111/j.1532-5415.2011.03481.x.
- Cunningham R, Sarfati D, Stanley J, et al. Cancer survival in the context of mental illness: a national cohort study. Gen Hosp Psychiatry. 2015;37(6):501–506. doi: 10.1016/j.genhosppsych.2015.06.003.
- Kisely S, Forsyth S, Lawrence D. Why do psychiatric patients have higher cancer mortality rates when cancer incidence is the same or lower? Aust N Z J Psychiatry. 2016;50(3):254–263. doi: 10.1177/0004867415577979.
- Chang CK, Hayes RD, Broadbent MTM, et al. A cohort study on mental disorders, stage of cancer at diagnosis and subsequent survival. BMJ Open. 2014;4(1):e004295. doi: 10.1136/bmjopen-2013-004295.
- Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209–249. doi: 10.3322/caac.21660.
- Manderbacka K, Arffman M, Lumme S, et al. The effect of history of severe mental illness on mortality in colorectal cancer cases: a register-based cohort study. Acta Oncol. 2018;57(6):759–764. doi: 10.1080/0284186X.2018.1429649.
- Haug U, Schink T. German pharmacoepidemiological research database (GePaRD). In: sturkenboom M, Schink T, editors. Databases for pharmacoepidemiological research. Cham, Switzerland: Springer; 2021.
- Schwarz S, Hornschuch M, Pox C, et al. Colorectal cancer after screening colonoscopy: 10-year incidence by site and detection rate at first repeat colonoscopy. Clin Transl Gastroenterol. 2023;14(1):e00535. doi: 10.14309/ctg.0000000000000535.
- Schröder H, Brückner G, Schüssel K, et al. Monitor: vorerkrankungen mit erhöhtem Risiko für schwere COVID-19-Verläufe. Verbreitung in der Bevölkerung Deutschlands und seinen Regionen. Berlin, 2020.
- Collett D. Modelling survival data in medical research. 3rd ed. Boca Raton: CRC Press, Taylor & Francis Group, Chapman & Hall; 2015.
- R Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2022.
- Nayak P, Luo RL, Elting L, et al. Impact of rheumatoid arthritis on the mortality of elderly patients who develop cancer: a population-based study. Arthritis Care Res. 2017;69(1):75–83. doi: 10.1002/acr.22997.
- Rabeneck L, Souchek J, El-Serag HB. Survival of colorectal cancer patients hospitalized in the veterans affairs health care system. Am J Gastroenterol. 2003;98(5):1186–1192. doi: 10.1111/j.1572-0241.2003.07448.x.
- Shack LG, Rachet B, Williams EMI, et al. Does the timing of comorbidity affect colorectal cancer survival? A population based study. Postgrad Med J. 2010;86(1012):73–78. doi: 10.1136/pgmj.2009.084566.
- Marventano S, Grosso G, Mistretta A, et al. Evaluation of four comorbidity indices and charlson comorbidity index adjustment for colorectal cancer patients. Int J Colorectal Dis. 2014;29(9):1159–1169. doi: 10.1007/s00384-014-1972-1.
- Renzi C, Kaushal A, Emery J, et al. Comorbid chronic diseases and cancer diagnosis: disease-specific effects and underlying mechanisms. Nat Rev Clin Oncol. 2019;16(12):746–761. doi: 10.1038/s41571-019-0249-6.
- Ott A, Breteler MMB, van Harskamp F, et al. Incidence and risk of dementia – The rotterdam study. Am J Epidemiol. 1998;147(6):574–580. doi: 10.1093/oxfordjournals.aje.a009489.
- Wittchen HU, Jacobi F, Rehm J, et al. The size and burden of mental disorders and other disorders of the brain in Europe 2010. Eur Neuropsychopharmacol. 2011;21(9):655–679. doi: 10.1016/j.euroneuro.2011.07.018.
- Manderbacka K, Arffman M, Suvisaari J, et al. Effect of stage, comorbidities and treatment on survival among cancer patients with or without mental illness. Br J Psychiatry. 2017;211(5):304–309. doi: 10.1192/bjp.bp.117.198952.
- Ribe AR, Laurberg T, Laursen TM, et al. Ten-year mortality after a breast cancer diagnosis in women with severe mental illness: a danish population-based cohort study. PLOS One. 2016;11(7):e0158013. doi: 10.1371/journal.pone.0158013.
- Rich CL, Ricketts JE, Fowler RC, et al. Some differences between men and women who commit suicide. Am J Psychiatry. 1988;145(6):718–722. doi: 10.1176/ajp.145.6.718.
- Oppelt KA, Luttmann S, Kraywinkel K, et al. Incidence of advanced colorectal cancer in Germany: comparing claims data and cancer registry data. BMC Med Res Methodol. 2019;19(1):142. doi: 10.1186/s12874-019-0784-y.
- Schwarz S, Hornschuch M, Pox C, et al. Polyp detection rate and cumulative incidence of post-colonoscopy colorectal cancer in Germany. Int J Cancer. 2023;152(8):1547–1555. doi: 10.1002/ijc.34375.
- Al-Husseini MJ, Saad AM, Mohamed HH, et al. Impact of prior malignancies on outcome of colorectal cancer; revisiting clinical trial eligibility criteria. BMC Cancer. 2019;19(1):863. doi: 10.1186/s12885-019-6074-6.
- Walter V, Jansen L, Ulrich A, et al. Alcohol consumption and survival of colorectal cancer patients: a population-based study from Germany. Am J Clin Nutr. 2016;103(6):1497–1506. doi: 10.3945/ajcn.115.127092.
- Fassmer A, Schink T. Repräsentativität von ambulanten Arzneiverordnungen in der German Pharamcoepidemiological Research Database (GePaRD). Paper presented at the 9. Jahrestagung der Deutschen Gesellschaft für Epidemiologie (DGEpi). Ulm, 17-20. 2014.