Systematic review of the association between socioeconomic status and bladder cancer survival with hospital type, comorbidities, and treatment delay as mediators

Abstract Objectives To review the current evidence on the relationship between three proposed mediators (comorbidities, hospital type, and treatment delays) for the relationship between socioeconomic status (SES) and bladder cancer survival. Materials and methods Six different searches using OVID (Medline and Embase) were carried out to collate information available between the proposed mediators with both SES and survival in bladder cancer. This systematic review was conducted according to a pre‐defined protocol and in line with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines. Results A total of 49 studies were included in the review across the six searches (one appeared in two searches). There was a wealth of studies investigating the relationship between each of the proposed mediators with survival in bladder cancer patients. In general, a higher SES, lower comorbidities, and a larger hospital volume were all found to be associated with a decreased risk of death in bladder cancer patients. There was, however, a paucity of studies investigating the associations between these mediators and SES in bladder cancer patients. Conclusions To gain a deeper understanding of the relationship between SES and survival identified in several observational studies, further investigations into the relationship between the proposed mediators and SES are warranted. Moreover, modifiable mediators, eg, treatment delay, highlight the importance of the standardization of clinical care across SES groups for all bladder cancer patients.

In bladder cancer, the link between SES and survival has been studied previously and patients with lower SES have been shown to have a decreased 5-year survival, 4 overall survival, 5 and higher relative risk of death. 6 Null finding has also been observed for the risk of bladder-cancer-specific mortality. 5 Around 75% of all bladder cancer patients have non-muscle-invasive bladder cancer (NMIBC) with the remaining 25% being diagnosed with muscle-invasive bladder cancer (MIBC). The main treatment choice for patients with NMIBC is usually transurethral resection of the bladder tumor (TURBT), Bacillus Calmette-Guerin (BCG) therapy or chemotherapy. 7 For non-metastatic MIBC, the main treatment choices are radical cystectomy, chemotherapy or radiotherapy. 8 The exact mechanism behind the association between SES and survival is complex and unknown; however, a review by Quaglia et al. postulated that the link could be explained by factors relating to three main groups: diagnosis, treatment modalities, and patient characteristics. 9 Despite this evidence, there remains paucity in detailed studies and comprehensive clinical investigations to elucidate the underlying mechanisms behind this association. SES is a largely unmodifiable factor; therefore, identifying potential mediators of the association between SES and survival could be used as a foundation for future interventions or recommendations to reduce the SES disparity seen in cancer survival.
Previously, using data from a cohort of Swedish bladder cancer patients, we found that Charlson Comorbidity Index (CCI), hospital type, and treatment delays mediated the association between SES and risk of death. 10 On the basis of these results, the aim of this systematic review is to collate information from existing literature about the potential mediators (hospital type, comorbidities, and treatment delay) for the association between SES and survival in bladder cancer patients.

| MATERIAL S AND ME THODS
Using similar methods as those described by Shanmugalingam et al,11 six separate searches were performed to investigate the relationship of each potential mediator (hospital type, comorbidities, and treatment delay) with both SES and survival in bladder cancer patients.
The full protocol is outlined in the Appendix. Figure 1 demonstrates the directed acyclic graph with each arrow representing a search carried out. The six searches were as follows: (1) SES and survival, (2) SES and hospital type, (3) Hospital type and survival, (4) SES and comorbidities, (5) comorbidities and survival, (6) SES and treatment delay. Number 7 in Figure 1 (the association between treatment delay and survival), however, was not carried out as the results are presented elsewhere. 12

| Search strategies
The online database Ovid Gateway was used to search both Embase Articles were included if: they were specific to bladder cancer, they were not conference abstracts or commentaries and the full text was available. Studies were excluded if they did not investigate the relevant exposure and outcome variables for each mediator and/ or were deemed of low quality after quality assessment (as explained below). Reviews were included to allow for the inclusion of as much information as possible including data from studies that might not have been captured within the current search strategy. When systematic reviews were already conducted according to the PRISMA guidelines, these were considered to overwrite individual studies included in the systematic review and the results were presented as a whole. Initially, titles were screened for relevance, then abstracts and full texts were subsequently screened. Data extraction was performed on a per mediator basis in which the year, country of study, number of patients, method for assessing comorbidities or socioeconomic status, and summary measures of results (e.g. survival F I G U R E 1 Directed acyclic graph with each number depicting a different association investigated. Socioeconomic status is the exposure variable, death is the outcome variable, while hospital type, comorbidities, and treatment delay are the potential mediators proportions, odds ratios or hazard ratios) were recorded in separate tables. All screening was performed by BR and MVH and data extraction by BR. All studies were described and compared in a narrative manner with no quantitative analyses taking place.
This review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 13 The quality of the studies was assessed using Risk of Bias in Non-Randomized Studies -Interventions (ROBINS-I) for all observational studies with death as the outcome. 14 The systematic reviews were assessed using A Measurement Tool to Assess Systematic Reviews 2 (AMSTAR 2), 15 while the narrative reviews were assessed using the Scale for the quality Assessment of Narrative Review Articles (SANRA). 16

| RE SULTS
A total of 1168 studies were initially extracted from the searches. A summary of the frequency of studies identified and subsequently included in the analysis for each search is depicted in Figure 2. No studies were identified for inclusion for the associations between SES and both comorbidities and hospital type (arrows 2 and 4, Figure 2).
Seven of the studies included both NMIBC and MIBC patients within their analyses and four of these concluded an association between a lower SES and reduced overall survival 5,18,20,23,28 (Table 1).
The two studies which did not observe this result were those of Eberle et al and Syriopoulou, although the latter only presented raw data with no level of significance reported for any results. 19,24 Four of the studies included MIBC only 22,[25][26][27] and one study included metastatic MIBC patients only. 21 All five concluded an association between a lower SES and reduced overall survival. None of the studies included any of the proposed mediators (comorbidities, hospital type or treatment delay) within their analyses.

| Comorbidities and survival
709 studies were extracted from the search of which 19 were deemed suitable for inclusion (Table 3). These studies included 1 review 45 and 18 cohort studies 46-63 (of which eight were from single-center data) 63 ).
Thirteen of the 19 studies investigated the association between comorbidity and survival in patients undergoing radical cystectomy, one of which also looked at MIBC patients who had undergone external beam or interstitial radiotherapy. 53 Three studies looked at this association in both NMIBC and MIBC patients; however, the study by Safarti et al 51 did not stratify the analyses by NMIBC/MIBC and so the results could not be presented separately. Two studies looked at NMIBC patients only and the remaining study looked at patients treated with TURBT; the latter study did not describe more detailed stage information beyond NMIBC. 46 None of the studies investigating comorbidities and survival included an SES indicator in their analyses.
For all MIBC patients and those undergoing curative treatment (radical cystectomy or external beam radiotherapy), 12 of the studies concluded an association between increasing comorbidity (as defined by each study) and an increased risk of overall death. [48][49][50][53][54][55][56][57][58][59][60]62 Three studies stated the same direction of association for cancer-specific death. 54,57,61 The ways in which the studies measured comorbidity for the radical cystectomy patients varied as can be seen in Table 3. Some examples of measurements used include CCI, age-adjusted CCI Two studies specifically investigated chronic kidney disease (CKD) as a major contributing factor to the link between comorbidity status and survival in bladder cancer patients. 49,63 The study by

Li et al encompassed NMIBC patients who underwent TURBT, both
with and without CKD. 63 The authors concluded that bladder cancer patients with CKD had increased odds of all-cause death when compared to patients without CKD.

| SES and treatment delay
Two studies were deemed suitable for inclusion when investigating the relationship between SES and treatment delay 1,5 (

| Quality of the included studies
Overall, the majority of the observational studies assessed using the ROBINS-I tool were deemed to have a low risk of bias (Tables S1-S3, Appendix were all positively associated with in increased risk in cancer-independent mortality after radical cystectomy. ASA (3,4), ECOG (2,3), ACE27 (3), and ACCI (>5) were all associated with an increased risk of cancer-specific mortality after radical cystectomy + Mayr et al. to be the studies that had not sufficiently adjusted for all possible confounding variables or stated crude survival proportions only. Using the AMSTAR criteria, there was deemed to be a moderate level of confidence in the results from the four systematic reviews ( Figure S1, Appendix). There was, however, a paucity in pre-defined study protocols detailing the search, definition of inclusion and exclusion criteria, and all failed to assess the risk of bias of the studies within each relative review. The two narrative reviews were both deemed of good quality according to the SANRA criteria ( Figure S2, Appendix).

| Comorbidity
Although no studies looking at SES and comorbidity were identified, outside of the bladder cancer setting evidence does exist for an association between these two variables. For example, a more generalized systematic review and meta-analysis 66

| Treatment delay
The paucity of studies investigating SES and treatment delay, together with the literature on SES, diagnostic delay, and variation in emergency presentation 67 suggests a complex mediation of SES and survival in bladder cancer. Thus, more advanced disease at diagnosis in patients with lower SES affects further treatment possibilities. 68

| Strengths and limitations
Several of the studies, in particular those investigating the association between comorbidities and survival, were single-center studies. 52,[57][58][59][60][61][62][63] Possible reasons for this include the lack of a universal comorbidity score and that different scores perform better from one disease or treatment or surgery to another. These studies may not have the same degree of external validity seen in nationwide or larger studies. Nonetheless, single-center studies are still crucial in healthcare research as they can capture data on a more granular level when compared to larger cohort studies. This review additionally benefitted from the inclusion of many large cohort studies.
These studies are inclusive of heterogeneous populations, which is especially important when studying factors such as SES.
A large proportion of the studies, particularly when looking at comorbidities, included patients undergoing radical cystectomy.
Therefore, a limitation to the current review is the low number of studies that included NMIBC patients, or those undergoing alternate treatments. Some studies did, however, include patients undergoing radiotherapy. It is important to note, however, that this may also be a limitation since patients who receive radiotherapy often do so as a result of being too frail for surgery and hence may skew the results for survival. Furthermore, none of the studies within this review used a formal mediation analysis. It is also possible that some literature may have been missed if studies included the research question of interest as a secondary or tertiary research question and were subsequently not picked up during title and abstract screening.
Meta-analyses were not deemed suitable in this review due to the heterogeneity of the studies identified.

| CON CLUS IONS
The studies identified in this review imply associations between the possible mediators of the association between SES and survival in bladder cancer patients (hospital type, comorbidities, and treatment delay