The Impact of Social Determinants of Health Beyond Race on Survival in Metastatic Prostate Cancer: A Systematic Literature Review

J Clin Pathways. 2023;9(6):51-59. doi:10.25270/jcp.2023.11.01

Prostate cancer (PC) accounts for 7% of newly diagnosed cancers in men globally and is the fifth most common cause of male cancer mortality.^1,2 Incidence rates of PC vary greatly worldwide, ranging from 6.3 to 83.4 per 100 000 people, with higher incidences seen in nations with higher life expectancies and gross domestic products, such as the US and UK.^2,3 Although recent years have seen a decline in overall can­cer death rates worldwide,⁴ socioeconomic inequalities in cancer mortality continue to increase,^5-7 and patients with metastatic prostate cancer (mPC) face a particularly poor prognosis.⁸ In an effort to mitigate inequalities in cancer outcomes, social de­terminants of health (SDOH), or the nonmedical factors that can influence health outcomes, are increasingly at the forefront of cancer research, including PC.^5,9 Un­derstanding the impact of SDOH on survival in mPC presents an important oppor­tunity for improving patient outcomes.

Race is one SDOH that has been extensively investigated in mPC, including in a prior report by our team.¹⁰ However, recent evidence has shown that SDOH beyond race also contribute to disparities in mPC survival.¹¹ Currently, available reviews examining the impact of non-race SDOH on PC outcomes are narrative in nature or focus on PC generally, as opposed to mPC.^12-16 The diverse characteristics of PC, which can vary from indolent to aggressive, may lead to conflicting con­clusions. Hence, research focusing on a more narrowly defined group of patients, all with potentially fatal disease (ie, those with mPC), is re­quired.¹⁷ We conducted a systematic literature review (SLR) to identify and summarize evidence on the relationships be­tween SDOH (defined herein as income, socioeconomic sta­tus [SES], marital status, type of treatment facility, insurance status, location, and education) and survival outcomes in pa­tients with mPC.

Literature Search

The SLR adhered to the Cochrane guidelines¹⁸ and the Pre­ferred Reporting Items for Systematic Review and Meta- Analysis (PRISMA) statement.^19,20 Study eligibility criteria followed the population, intervention, comparator, outcome, study design (PICOS) framework (Supplementary Table 1).¹⁸ A medical information specialist designed and imple­mented the search strategy (Supplementary Table 2), which was peer reviewed using the Peer Review of Electronic Search Strategies (PRESS) checklist.²¹ Embase, MEDLINE (Epub Ahead of Print, In-Process & Other Non-Indexed Cita­tions, and Daily), and the Cochrane Database of Systematic Reviews were searched using the Ovid platform on July 7, 2022. Retrieval was limited to full-text studies published in English from 2012 onward, which was considered by expert clinical input to reflect the modern mPC treatment landscape. Conference abstracts from 2019 onward were eligible for in­clusion to capture studies not yet published as full texts. A supplemental search for relevant conference abstracts from key congresses was also conducted (Supplementary Table 3). The full study protocol was registered with PROSPERO international prospective register of systematic reviews (reg­istration number CRD42022350888).

Study Selection and Data Synthesis

The titles and abstracts of retrieved records were screened for inclusion and then assessed in full-text form against the eligibil­ity criteria (Supplementary Table 1) by two independent re­viewers using a systematic literature review software.²² A third independent reviewer was consulted when consensus could not be achieved. Although the search strategy was designed to iden­tify studies reporting on the impact of SDOH (including race) on a variety of clinical outcomes in patients with mPC, the focus of the present article was studies reporting on income, SES, marital status, type of treatment facility, insurance status, location, and education on survival outcomes. Data extraction and quality assessment was performed by a single reviewer and confirmed by a second reviewer. A standardized data extraction form in a spreadsheet-based software was used. Clinical judg­ment was employed by an experienced prostate cancer clinician and researcher (SJF) to assess the castration sensitivity status of patient populations (metastatic castration-sensitive prostate cancer [mCSPC], metastatic castration-resistant prostate can­cer [mCRPC], or mixed [both mCSPC and mCRPC]) when not explicitly stated.

Quality Assessment

The Newcastle-Ottawa Quality Assessment Scales for cohort studies were used to evaluate the study quality of full-text pub­lications.²³ Studies were given a total score between 0 and 9 and classified as low quality/high risk of bias (<5), moderate quality (5 to ≥7), or high quality/low risk of bias (≥7).²⁴

Overview of Included Studies

Database searches identified 3228 records after deduplication, with an additional 173 conference abstracts identified in the supplemental congress search. Following screening, 22 records (20 full texts^25-44 and 2 conference abstracts^45,46) representing 22 unique studies reporting on the impact of SDOH beyond race on survival outcomes were included (Figure 1). See Supple­mentary Table 4 for the list of included citations.

Study Characteristics

Characteristics of the 22 included studies are summarized in Table 1. All were retrospective cohort studies^25-46 and reported on mCSPC (n = 18)^{25-27,29-34,36,37,40-46} or mixed mCSPC/mCRPC populations (n = 4).^28,35,38,39 Study locations included the US (n = 17),^{26-33,36-39,42-46} Europe (n = 4),^25,35,40,41 and one study con­ducted in both the US and China.³⁴ Studies investigated in­come (n = 13),^{26-29,31,32,35,36,39,41-43,45} SES (n = 2),^25,34 marital status (n = 8),^{26-28,30,31,34,43,44} type of treatment facility (n = 8),^{29,32,33,36- 39,42} insurance status (n = 12),^{29,31-33,36-39,42,43,45,46} location (n = 8),^{27,29,32,35,38,39,42,43} and education (n = 7).^{27,28,31,36,40-42} Some stud­ies reported results for multiple SDOH. A total of 14 studies reported results for overall survival (OS) only,^{25,29,30,32-39,42,44,45} 6 reported on PC-specific mortality (PCSM) only,^{26-28,31,40,46} and 2 reported both outcomes.^41,43

Assessment of Study Quality

Overall, 95% of studies (19 out of 20) published as full-text articles were determined to be of high quality/low risk of bias (score ≥7) (Supplementary Table 5).^{24,25-34,36-44}

SDOH Results

The results of each SDOH are summarized in a separate subsec­tion below. The described results reflect findings for which two or more applicable studies reported a specific comparison (eg, OS between private insurance vs uninsured, PCSM between married vs unmarried, etc). Studies must also have allowed for the impact of a single SDOH to be discerned (eg, SDOH X must have been considered as a single variable, rather than as a composite of SDOH X + SDOH Y). Results not meeting these criteria were uncommon but were included in Supplementary Table 6 for completeness.

Income

Among the studies that reported on income and OS (Figure 2A), the majority (7 of 8 studies) reported that higher income was associated with improved OS,^{29,32,36,41-43,45} whereas only one study found no significant differences between income lev­els.^35,39 Regarding PCSM, 3 of 6 studies reported that higher income was associated with improved PCSM^31,41,43 while the other three found no association.^26-28

Socioeconomic Status

Both studies (2 of 2 studies) investigating the impact of SES re­ported that higher SES was associated with improved OS (Fig­ure 2B).^25,34 No studies investigating the association between SES and PCSM were identified.

Marital Status

Being married was associated with improved OS (4 of 4 stud­ies)^30,34,43,44 and improved PCSM (3 of 4 studies)^26,27,43 compared to unmarried patients (Figure 2C).

Type of Treatment Facility

Most studies (4 of 7 studies) reported that treatment in an academic facility was associated with improved OS compared to nonacademic and/or community facilities (Figure 2D).^29,32,37,42 Additionally, no association in OS was found between patients treated in a comprehensive facility (ie, defined as those that treat 500 or more cancer patients each year)⁴⁷ compared to a community facility, in 2 of 2 studies.^36,42 No studies investigat­ing the association between treatment facility and PCSM were identified. Specific definitions used for the types of treatment facilities compared are outlined in Supplementary Table 7 and Supplementary Table 8.

Insurance Status

Within the studies that reported on insurance status and OS (Figure 3A), five categories were reported: private insurance, Medicare, Medicaid, uninsured, and insured generally (any form of insurance). Medicare is federal health insurance in the US for patients 65 years of age or older as well as those with certain disabilities or conditions, while Medicaid is joint federal and state insurance administered by each state in the US that covers medical costs for certain patients with limited income and/or resources. Although the federal government sets gen­eral rules for state Medicaid programs, variation in eligibility requirements and benefits can exist between states.⁴⁸ All studies were US-based, and all but one adjusted for various sociode­mographic factors in their analyses, including age.

For private insurance, the majority of studies reported that patients with private insurance had improved OS com­pared to those with Medicare (4 of 6 studies)^29,37,42,45 and those who were uninsured (3 of 5 studies).^29,39,45 Findings for pri­vate vs Medicaid insurance, however, were conflicting: Two studies concluded that patients with private insurance had improved OS compared to patients with Medicaid,^29,45 one study found the opposite,³⁹ and one study found no difference in survival.³⁷

For Medicare, the majority of studies reported no difference in OS between patients with Medicare vs Medicaid (2 of 3 stud­ies),^33,38 as well as between patients with Medicare vs those who were uninsured (2 of 3 studies).^33,36 None of the studies pro­vided additional information regarding the specifications/nu­ances of the different insurance types considered (eg, coverage, requirements, benefits, eligibility, etc) or whether patients were dually insured.

Lastly, both studies that compared patients with any form of insurance vs uninsured patients found that insured patients had improved OS.^32,43

With respect to PCSM, no two identified studies reported common comparisons across the various categories of insur­ance status (Supplementary Table 6).

Location

Six studies compared OS in patients with mPC across differ­ent geographical regions or locations within the US (eg, Pacific region, middle Atlantic region, mountain Pacific region, south­ern region, eastern region, etc). Although the regions that were compared between studies varied, significant differences in OS across geographic regions of the US were found in the majority of studies (5 out of 6 studies) (Figure 3B).^{29,32,38,39,42}

Additionally, a few studies were identified that assessed the impact of metropolitan, urban, or rural location on OS, with variable findings (Figure 3C). Two studies compared rural and urban locations to metropolitan locations, using data from the NCDB.^29,32,49 See Supplementary Table 9 for specific defi­nitions based on population size and location. In both studies, urban location was found to be associated with improved OS compared to metropolitan location.^29,32 For the rural and metro­politan location comparison, one study found no association,²⁹ while the other concluded that rural location was associated with improved OS compared to metropolitan.³²

An additional two studies compared only urban and rural locations,^38,39 with both using data from a national sample cre­ated by the Agency for Healthcare Research and Quality.^50,51 See Supplementary Table 10 for specific definitions based on population density and location. One study found that urban location was associated with improved OS,³⁹ while the other study found no significant differences.³⁸

Regarding PCSM, no two identified studies reported com­mon comparisons across the various categories of location (Supplementary Table 6).

Education

Of the three studies that reported on education level and OS, one study reported that higher education was associated with improved OS,⁴¹ while the two remaining studies reported no association (Figure 3D).^36,42 Similarly for PCSM, 2 of 5 stud­ies found that higher education was associated with improved PCSM^40,41; however, the remaining three studies reported no association.^27,28,31

The aim of this review was to identify and summarize recent studies investigating the relationship between SDOH beyond race (defined herein as income, SES, marital status, type of treatment facility, insurance status, location, and education) and survival outcomes in patients with mPC. A total of 22 stud­ies were identified, highlighting a dearth of research and evi­dence in this area. The majority of studies reported significant differences in survival by key SDOH, including income, SES, marital status, type of treatment facility, insurance status, and location, with the exception of education where the majority of studies reported no association with survival outcomes. Higher income and SES, private insurance, treatment in an aca­demic facility, and being married were found to be associated with a survival benefit in patients with mPC.

The overall finding of this SLR, that SDOH are associated with disparities in survival for patients with mPC, is generally consistent with the literature of PC more broadly and oncol­ogy in general. The trends and relationships on survival out­comes in mPC seen herein with income/SES, marital status, and location align with previous narrative reviews and reports in PC.^12,14,52,53 The finding that survival outcomes are improved when patients are treated at an academic facility is consistent with findings in other oncology areas.^54,55 For insurance sta­tus, improved outcomes with private insurance have also been previously discussed in the oncology literature, with evidence suggesting that some of this benefit is attributable to earlier detection and treatment.⁵⁶

With respect to education, however, the inconclusive find­ings from the current SLR differ with prior reports in the lit­erature. Notably, a recent large-scale study reported that the survival disparity between patients with higher and lower education has been increasing over time for patients with common cancers (including prostate).⁵⁷ Although inconclusive results were found in the present SLR, education is closely linked to other SDOH examined, such as SES and income, where the majority of studies reported significant survival differences. In fact, SES measures often encompass income and educational at­tainment, among other factors.⁵⁸ Prior studies in cancer gener­ally, as well as prostate cancer, have reported that both lower education and income or SES are associated with higher risk of mortality.^59,60 A large US study conducted by Singh and Je­mal found that men with low education and income were at an increased risk of all-cancer mortality and prostate cancer–spe­cific mortality.⁵⁹ Similarly, both lower neighborhood SES and lower education were found to be significantly associated with a greater risk of death in a large recent study of prostate cancer patients.⁶⁰ It is important to note, however, that none of these analyses were specific to the metastatic setting. Further research may be required to clarify this relationship between education and mortality, as well as how this relationship correlates with other related SDOH, such as income and SES.

Our SLR identified several important data gaps. Compared to 54 studies previously identified by our team that reported on race and survival,¹⁰ the present analysis identified only 22 studies that reported on SDOH other than race. Thus, most research conducted to date on the impact of SDOH on sur­vival outcomes in mPC has focused on race, with much less attention paid to other important SDOH. Consequently, sev­eral SDOH were investigated in only 8 or fewer studies, with very few studies reporting many comparisons (eg, high vs low SES, comprehensive vs community treatment facility, Medi­care vs uninsured, etc), making it difficult to identify trends and conclusions. For example, the comparison between ur­ban and rural location was only reported in two studies, with conflicting results, making it unfeasible to discern an overall conclusion or finding for this comparison. Additionally, for 5 of the 7 SDOH of interest, there was insufficient data to assess the impact of these SDOH on PCSM, indicating a paucity of information on PCSM compared to OS. Further research in this area of other SDOH is needed, as these factors can con­tribute to disparities in PC outcomes as well as imparting in­fluences on their own.^13-15

In addition to the relatively small number of studies, there was also substantial heterogeneity in the reference groups and specific categories used to define the various SDOH across the included studies, especially for location (eg, urban/rural/metro­politan), type of treatment facility (academic/teaching facility, community facility, integrated network facility, etc), education (eg, with vs without high school diploma, basic vs higher edu­cation, etc), and income (eg, ≥$66 610 vs <$66 610, ≥$68 000 vs <$68 000, etc). For income and education, we simplified and interpreted categories as “high vs low” to enable comparison of findings across studies. Other sources of heterogeneity also in­cluded the use of different statistical tests and reporting meth­odologies, which prevented comparisons of the magnitude of the observed relationships. These factors made it challenging to directly compare findings across studies and points to the need for a more consistent classification and methodology for ex­ploring SDOH in the mPC setting. Furthermore, no identified studies were conducted in an mCRPC population alone, and very few were conducted outside the US, thus highlighting the need for further research on SDOH on survival outcomes in more advanced mPC populations and in ex-US regions.

Several strengths and limitations of this review should be noted. The present SLR provides a comprehensive and thor­ough assessment of the existing research, concentrating specifi­cally on mPC, as opposed to previous reviews that examined the impact of SDOH on PC as a whole. To incorporate and reflect the growing body of evidence, the review included full-text articles focused on real-world evidence and published within the past 10 years. In an effort to incorporate the most re­cent research in the field, conference abstracts were also includ­ed; however, they often only report limited information. As only articles published in English were included, the findings of non-English articles are not accounted for in our findings. Additionally, although the use of clinical judgment to assess the castration sensitivity status of patient populations when neces­sary is inherently subjective, this provided unique insight into the populations assessed in the available literature. Lastly, sig­nificant heterogeneity was observed between included studies and only qualitative findings are presented within this review. Future studies that quantitatively evaluate the impact of SDOH and survival to better understand and identify the relationships and disparities are needed.

In summary, this study provided a thorough systematic assess­ment of the impact of SDOH beyond race on survival in the mPC population. Results from this SLR reveal that disparities in mPC survival outcomes by key SDOH continue to exist. Higher income and SES, private insurance, treatment in an academic facility, and being married were associated with a survival advantage for patients with mPC. Despite increased attention on SDOH in recent years, surprisingly few studies investigating SDOH in this area were identified, highlighting the need for additional research to further explore the impacts of SDOH on outcomes and access to care in mPC. Continued focused efforts are needed to further understand these dispari­ties and expand the potential survival benefits to all patients.

This article has supplementary material, which can be accessed at this link.

Authors: Stephen J. Freedland, MD^1,2; Imtiaz A. Samjoo, MSc, PhD³; Emily Rosta, MSc³; Austin Lansing, MSc³; Alexandra Hall, MPH³; Manvir Rai, BSc³; Alexander Niyazov, PharmD, MPH⁴; Jonathan L. Nazari, PharmD⁴; Bhakti Arondekar, PhD, MBA⁵

Affiliations:  ¹Division of Urology, Department of Surgery, Cedars-Sinai Medical Center, Los Angeles, CA; ²Urology Section, Durham VA Medical Center, Durham, NC; ³EVERSANA, Burlington, Ontario, Canada; ⁴Pfizer, Inc, New York, NY; ⁵Pfizer, Inc, Collegeville, PA.

Address correspondence to: 

Stephen J. Freedland

8635 West 3rd Street Suite 1070W

Los Angeles, CA 90048

Email: stephen.freedland@cshs.org

Acknowledgements: The authors thank Kerise Clarke, Amrita Debnath, and Joanna Bielecki for their assistance with the literature review. All were employed by EVERSANA, Canada.

Disclosures: S.J.F. reported serving as a consultant for Pfizer, Astellas, Janssen, Bayer, Merck, AstraZeneca, Myovant, and Sanofi; and serving as a speaker for AstraZeneca and Sanofi. I.A.S., E.R., A.L., A.H., and M.R. reported being employees of EVERSANA, Canada, which was a paid consultant to Pfizer in connection with the development of this manuscript. A.N. and B.A. reported being employees of Pfizer and owning Pfizer stock. J.L.N. reported being an employee of the University of Illinois Chicago, which was a paid contractor to Pfizer in connection with the development of this manuscript through the 2021-2023 University of Illinois at Chicago- Pfizer Fellowship.