Abstract
Multiple myeloma (MM) is more common among Black/African American (AA) patients than White patients, but survival rate improvements are less pronounced for AA patients. This study evaluated treatment patterns and survival among 1810 AA and 5904 White adults in the United States with ≥1 MM treatment and ≥3 months of follow-up. Median time from diagnosis to systemic treatment was longer (37 [0–3053] vs. 35 [0–3664] days) and median time to stem cell transplant (SCT) was longer for AA than White patients (255 [1–2352] vs. 225 [1–3094] days), and AA patients were less likely to receive SCT (odds ratio [OR]: 0.66; 95% confidence interval [CI]: 0.58–0.76). Despite disparities in treatment between AA and White patients, AA patients demonstrated lower risk of death (OR: 0.89; 95% CI: 0.81–0.96). These data highlight the value of equal access to care for the improvement of health outcomes in underserved populations.
Introduction
Health equity—a state in which everyone has a fair and just opportunity to attain their highest level of health—has been a key focus of the Office of Disease Prevention and Health Promotion since the 1990s [Citation1]. Decades of collecting evidence about healthcare access and practices across the United States (US) highlighted the existence of significant health disparities between races across many disease areas. For example, compared with White people, Black/African American (AA) are more likely to die from HIV/AIDS, hypertension, kidney disease, diabetes mellitus, stroke, heart disease, and cancer, including myeloma [Citation2–4].
Multiple myeloma (MM) is the most common hematologic malignancy among AA adults in the US [Citation5]. The incidence of MM is twice as high among AA patients as among White patients, with age-adjusted incidence rates of 15.5 and 7.0 per 100,000 people in 2019, respectively [Citation5]. In addition to higher incidence, AA patients are more likely to experience clinical complications associated with MM diagnosis such as anemia, renal insufficiency, and poorer prognosis [Citation6]. Although MM-specific survival rates have been historically higher among AA than White patients, improvements in survival rates over time have been less pronounced for AA patients [Citation7]. This disparity may be related to AA patients being less likely to receive novel MM therapies and/or stem cell transplant (SCT) and having a longer time from diagnosis to treatment when compared with White patients [Citation8–10].
Despite disparities in access to frontline treatment and slower progress in improved survival rates, studies have shown that overall survival (OS) rates remain comparable or better among AA patients when compared with White patients [Citation11,Citation12]. When evaluated in an equal access setting, large population-based studies have shown that AA patients with MM tend to be younger at diagnosis and have a lower incidence of high-risk cytogenetics [Citation7,Citation8,Citation13]. Although these factors may impact treatment patterns, the association between patient characteristics, treatment patterns, and survival have not been thoroughly explored. In addition, while some risk factors for MM among AA patients have been reported [Citation14], few studies have investigated predictors of clinical outcomes according to race or ethnicity in robust real-world data sets.
This study explored how real-world clinical characteristics and treatment patterns have evolved between AA and White patients with MM over a 10-year period using a large, electronic health record (EHR)-derived database. We aimed to characterize potential racial disparities in care and survival, and to identify potential factors influencing treatment patterns and survival among AA and White patients in US clinical practice.
Materials and methods
Data source
This was a retrospective observational study using Flatiron Health’s longitudinal, US-wide, EHR-derived database of de-identified, patient-level, structured and unstructured data that are curated via technology-enabled chart abstraction from physician notes and other documents [Citation15,Citation16]. The de-identified data originate from approximately 280 US cancer clinics (approximately 800 sites of care, with 80% of the patients coming specifically from community-based cancer centers). Data from December 2021 through to the end of January 2022 were extracted from the database in March 2022.
Patients
Inclusion criteria were: ≥18 years of age with a documented MM diagnosis (ICD-9-CM 203.0x; ICD-10-CM C90.0x or C90) between January 1, 2011 and October 31, 2021; reported as of AA or White racial background; had received ≥1 first-line (1 L) systemic treatment; ≥2 documented clinical visits on different days on or after January 31, 2011; and with ≥3 months of follow-up information. Patients were excluded if they had participated in a MM-targeted clinical study; they had other (including Asian, Hispanic or Latino, and Other Race), unknown, or missing race information; or, had other primary cancer diagnoses any time prior to the index date (i.e. documented initial MM diagnosis date).
Variables and outcomes
Demographic and clinical characteristics included age at MM diagnosis, sex, US region, Charlson Comorbidity Index (CCI) score, Eastern Cooperative Oncology Group performance status (ECOG PS), International Staging System (ISS) classification, and presence of high-risk cytogenetics according to International Myeloma Working Group (IMWG) criteria. Clinical characteristics also included CRAB measures of serum calcium, renal insufficiency (defined as serum creatinine ≥2 mg/dL), anemia (defined as hemoglobin ≤10 g/dL), and bone lesions; however, due to lack of sufficient data prior to diagnosis, bone lesions were infrequently captured and not included in subsequent analyses.
Treatment pattern information included number of treatment lines and time to treatment for patients receiving 1 L, second-line (2 L), third-line (3 L), or fourth- or greater-line (4 L+) therapy. Therapies eligible for inclusion in lines of therapy included chemotherapy, SCT, immunomodulatory imide drug (IMiD®) agents, proteasome inhibitors (PIs), and anti-CD38 monoclonal antibodies (mAbs). The start of 1 L therapy was defined as the date of the first drug administration of an eligible therapy given after the index date or up to 14 days before the index date. Subsequent lines of therapy were defined as the first eligible drug administration, plus other eligible drugs given within 28 days. The end of a line of therapy was defined as the date of the last patient-level activity (e.g., the last reported laboratory tests or results); or (in the case of patients with a recorded date of death) either the date of the last patient-level structured activity or the date of death, whichever was earlier; or the day before the start of the next line of therapy. Additional treatment data collected include receipt of SCT, time from MM diagnosis to SCT, receipt of specific antimyeloma agents (IMiD agents, PIs, anti-CD38 mAbs), and triple-class exposed status (i.e. treatment with an IMiD agent, PI, and anti-CD38 mAb).
Clinical outcomes included OS calculated from the index date to death, and documented progression-free survival (PFS) from index date to disease progression or death, whichever occurred first.
Statistical analysis
Descriptive statistics were used to analyze all continuous, categorical, and binary variables. No imputation of missing data was performed. Clinical characteristics were analyzed overall, by race, and by mutually exclusive subgroups based on year of MM diagnosis relative to former treatment patterns (<2015), newer treatment patterns including anti-CD38 mAbs (2015–2019), and current treatment patterns (≥2020). OS and PFS were estimated using the Kaplan–Meier method with censoring of patients still alive at the end of the follow-up period. Univariate and multivariate logistic regression models were used to investigate predictors of receiving SCT after MM diagnosis, and for predictors of death, adjusting for confounders of interest. All analyses were performed using the SAS version 9.4 (SAS Institute, Cary, NC, USA).
Results
Demographic and clinical characteristics
A total of 7714 patients met the eligibility criteria and were included in the analysis (Supplemental Figure S1), of whom 1810 (23%) were AA and 5904 (77%) were White. Approximately one-third of all patients were diagnosed before 2015 (n = 2593, 34%), 4008 (52%) were diagnosed between 2015 and 2019, and 1113 (14%) were diagnosed in 2020 or later () (p = 0.0036). Demographic characteristics were generally similar between AA and White patients, though AA patients tended to be younger at the time of MM diagnosis (median 66 vs. 70 years; p < 0.0001) and had a lower proportion of male patients (46% vs. 57%; p < 0.0001) (). The majority of AA and White patients had reported commercial health insurance. A greater proportion of AA patients had renal insufficiency (serum creatinine >2 mg/dL; 19% vs. 15%; p = 0.0021) and anemia (hemoglobin <10 g/dL[…]; 45% vs. 35%; p < 0.0001) when compared with White patients. In this Flatiron Health database analysis, the proportion of patients with high-risk cytogenetics was similar for both AA and White patients. The median duration of follow-up was similar between AA and White patients (35 [1–132] months vs. 34 [1–133] months; p = 0.3807) () and decreased over time: from 66 (1–132) months and 55 (1–133) months for AA and White patients diagnosed before 2015, respectively, to 36 (1–85) months for both AA and White patients diagnosed between 2015 and 2019, and 12 (1–25) months for those diagnosed from 2020 onward (data not shown). When select clinical characteristics were also analyzed by year of MM diagnosis subgroups, the proportions of patients with ISS stage III disease at diagnosis and CCI score ≥2 appeared to increase over time (Supplemental Table S1).
Table 1. Demographic and clinical characteristics.
Treatment patterns
Time from MM diagnosis to initiation of 1 L treatment was longer for AA compared with White patients overall (median [range]: 37 [0–3053] vs. 35 [0–3664] days; mean [SD]: 173 [399.7] vs. 153 [388.6] days, respectively; p = 0.0064), and particularly longer among AA patients and White patients diagnosed with MM between 2015 and 2019 (AA patients, 37 [0–2542] days; White, 32 [0–2461] days). Time from MM diagnosis to start of 1 L treatment for AA and White patients was also analyzed by 30- and 60-day periods by diagnosis period subgroups, which showed more patients in later diagnosis periods received 1 L treatment within 60 days of diagnosis ().
Figure 1. Time from MM diagnosis to 1 L treatment by race and diagnosis time period. 1L: first line; AA: Black/African American; MM: multiple myeloma.
Median duration of 1 L treatment was 1 week longer for AA versus White patients overall (199 [1–3720] vs. 192 [1–3897] days, respectively; p = 0.9685) (). The proportions of both AA and White patients receiving single-class exposure (IMiD agent, PI, or anti-CD38 mAb) only in 1L decreased over MM diagnosis time periods whereas the proportions of those receiving IMiD agent, PI, and anti-CD38 mAb (triple-class exposure) 1L regimens was identified predominantly among patients diagnosed with MM in 2020 or more recently (). In both groups, the proportion of patients receiving 1 L IMiD agents + PIs increased over the diagnosis periods. Similar proportions of AA and White patients received 2 L and 3 L therapy; the proportion of patients receiving 2 L and 3 L therapy in both groups was lower in later diagnosis periods, likely due to patients diagnosed more recently not yet having progressed through multiple lines of therapy ().
Figure 2. Triple-class exposure status in 1 L treatment regimen by race and diagnosis time period. 1L: first line; AA: Black/African American; IMiD: immunomodulatory imide drug; mAb: monoclonal antibody; PI: proteasome inhibitor.
aPatients treated with single-class therapy received IMiD agents only, PIs only, or anti-CD38 mAb only.
bPatients treated with double-class therapy received IMiD agents and PIs, IMiD agents and anti-CD38 mAb, or PIs and anti-CD38 mAb.
cPatients treated with triple-class therapy received IMiD agents, PIs, and anti-CD38 mAb.
Table 2. Treatment patterns by race and year of MM diagnosis subgroups.
Roughly similar proportions of AA and White patients received SCT after the index date overall (29% and 32%, respectively), and both showed trends toward lower proportions receiving SCT over the diagnosis periods (<2015 to ≥2020: AA patients, 32% to 22%; White patients, 34% to 26%). Median time from MM diagnosis to first SCT was 30 days longer for AA versus White patients (255 [1–2352] vs. 225 [1–3094] days, respectively; p < 0.0001) (; Supplemental Figure S2). The logistic regression model showed lower odds of receiving SCT after MM diagnosis for AA versus White patients (p < 0.0001), women versus men (p = 0.037), those with CCI score ≥1 versus 0 (p < 0.0001), and those with higher creatinine (p = 0.001) or Lambda versus Kappa light chain (p = 0.023) (Supplemental Table S2). Higher odds of receiving SCT after diagnosis were shown for younger versus older patients (p < 0.0001), those with lower or unknown ECOG PS (p < 0.0001), patients with high-risk cytogenetics (p < 0.0001), and patients with unknown/undocumented serum creatinine versus those with ≤2 mg/dL (p = 0.766).
Survival
Median OS was longer for AA patients (74.5 [1.2–132.4] months, 95% confidence interval [CI]: 67.5–78.4) compared with White patients (63.6 [1.1–133.0] months; 95% CI: 61.5–66.5; p = 0.0020) (). The multivariate Cox proportional hazards model showed a lower risk of death for AA versus White patients (p = 0.004), younger versus older patients (p < 0.0001), women versus men (p < 0.0001), those who received SCT versus no SCT (p < 0.0001), and those with lower versus higher ECOG PS at MM diagnosis (p < 0.0001) (Supplemental Table S3). Higher risk of death was shown for patients with high-risk cytogenetics versus none (p < 0.0001), higher creatinine level (p < 0.0001), and hypercalcemia (p < 0.0001). Median PFS was longer for AA patients (40.2 [0.5–132.4] months; 95% CI: 36.5–45.2) than for White patients (37.5 [0.4–133.0] months; 95% CI: 36.1–39.5; p = 0.1261) ().
Figure 3. OS and PFS by race. (A) OS by race; (B) PFS by race. AA: Black/African American; OS: overall survival; PFS: progression-free survival.
Discussion
This retrospective observational study investigated the characteristics, treatment patterns, and survival of AA and White patients treated for MM in US clinical practice. Overall, AA patients were younger at the time of MM diagnosis, but more often experienced MM-related complications such as renal insufficiency and anemia. White patients received 1 L treatment sooner than AA patients, particularly within the group of patients diagnosed between 2015 and 2019, but AA patients had a 7-day longer duration of 1 L treatment. Overall, younger age, lower or unknown ECOG PS, and high-risk cytogenetics were predictive of receiving SCT. The regression model showed AA patients were less likely to receive SCT overall, even though they were younger at MM diagnosis; and among patients receiving SCT, White patients received SCT approximately 1 month sooner than AA patients. Despite the disparities in SCT and time to treatment, AA patients had better OS and PFS than White patients overall. Younger age, being AA, and receipt of SCT were all associated with lower odds of death, whereas the presence of high-risk cytogenetics, greater comorbidity burden, and indicators of poorer health status were associated with higher odds of death. In an equal-access setting, AA patients appear to do well in terms of progression and survival overall, but important distinctions in clinical characteristics and treatment patterns remain. Whether or not survival for AA patients with MM would be further improved by shortening the time to treatment and/or SCT remains unclear.
These findings are consistent with previous reports investigating potential health disparities based on race/ethnicity in the US, including delayed initiation of 1 L MM therapy for AA versus White patients [Citation9–11]. This study also observed a younger age at diagnosis for AA patients, as reported in prior studies [Citation7,Citation9,Citation13], and comparable or improved survival for AA versus White patients with MM [Citation7,Citation11,Citation13,Citation17]. Patients had received ≥1 L therapy, which likely represents a generally equal-access setting even though commercial insurance was slightly lower and Medicaid coverage was slightly higher among AA patients. Nonetheless, demographic and clinical characteristics could be potential factors of disparities in the timing of 1 L treatment and receipt of SCT than considerations related to access to care. Despite a longer time to initiation of 1 L treatment for AA patients, use of combination therapy and advanced treatment options increased for both AA and White patients in the more recent MM diagnosis periods, suggesting potentially comparable treatment pathways in an equal-access setting. This observation is in line with previous studies of treatment patterns in AA versus White patients with MM with equal access that reported no racial disparities in terms of use of novel agents at treatment induction or receipt of SCT [Citation7,Citation13]. Disparities in the availability of novel agents may indeed be overcome with appropriate access considerations, and although the differences in broad treatment patterns based on race were limited in this study, further analysis exploring disparities in access to 1 L therapy, in particular, are needed.
These findings should be interpreted in the context of certain methodological strengths and limitations. The EHR-derived database used in this study is the largest real-time, real-world oncology data source, making the data likely to be generalizable to the broader MM patient population in the US. However, the data are largely collected from community-based oncology clinics in the US, which may differ from US academic centers and ex-US healthcare settings. It should be noted that the date of diagnosis is the earliest date of active MM recorded by the healthcare provider in the clinic, which may not be the absolute true date of diagnosis. Information prior to patients being diagnosed with active MM is not available, including previous smoldering MM diagnoses. Although Flatiron Health’s parallel unstructured and structured data processing improves completeness compared with structured methods alone, there is the potential for missing data, such as those related to patient history and treatment details for care received outside of the participating oncology EHR. Of note, some clinical information that may be of interest is not available in the database (e.g., tumor response, sites of metastases).
This study was conducted to help raise awareness for patients and healthcare professionals on the specific considerations and need for greater attention to factors that drive care and outcomes for historically underserved populations. In an equal-access setting, when AA patients have access to innovative treatments, survival outcomes can be comparable or better than those observed for White patients. Some important distinctions remain, such as time to receipt of 1 L treatment or SCT, that merit further exploration. This study underscores the value and impact of equal access to care for the improvement of health outcomes, specifically in relation to minority populations in the US.
Supplemental Material
Download Zip (186 KB)Acknowledgments
This study was supported by Bristol Myers Squibb. Medical writing and editorial support were provided by Jeff Frimpter, MPH, of Excerpta Medica, funded by Bristol Myers Squibb.
Disclosure statement
AS, SS, KS, and TG are employees and shareholders of Bristol Myers Squibb. ASR has provided consultancy to Kangpu Biopharmaceuticals and has received direct research funding from Bristol Myers Squibb and research funding for his institution from GlaxoSmithKline, Takeda, Kangpu Biopharmaceuticals, and Biomea Fusion. K-KA and MAQ declare to have no competing interests.
Data availability
BMS policy on data sharing may be found at https://www.bms.com/researchers-and-partners/independent-research/data-sharing-request-process.html.
Additional information
Funding
References
- Gómez CA, Kleinman DV, Pronk N, et al. Addressing health equity and social determinants of health through healthy people 2030. J Public Health Manag Pract. 2021;27(Suppl 6):S249–S257. doi:10.1097/PHH.0000000000001297
- Snowden LR, Graaf G, Keyes L, et al. Did Medicaid expansion close African American-white health care disparities nationwide? A scoping review. BMC Public Health. 2022;22(1):1638. doi:10.1186/s12889-022-14033-8
- Pathak EB. Mortality among black men in the USA. J Racial Ethn Health Disparities. 2018;5(1):50–61. doi:10.1007/s40615-017-0341-5
- DeSantis CE, Miller KD, Goding Sauer A, et al. Cancer statistics for African Americans, 2019. CA Cancer J Clin. 2019;69(3):211–233. doi:10.3322/caac.21555
- National Cancer Institute. Compare cancer sites: SEER 5-year age-adjusted incidence rates, 2015-2019. 2022.(Accessed November 15, 2022). Available from: https://seer.cancer.gov/statistics-network/explorer/application.html?site=9999&data_type=1&graph_type=10&compareBy=site&chk_site_83=83&chk_site_90=90&chk_site_89=89&chk_site_86=86&series=race&chk_race_9=9&chk_race_8=8&chk_race_3=3&chk_race_2=2&sex=1&age_range=1&hdn_stage=101&advopt_precision=1&advopt_show_ci=on&hdn_view=1.
- Fillmore NR, Yellapragada SV, Ifeorah C, et al. With equal access, African American patients have superior survival compared to white patients with multiple myeloma: a VA study. Blood. 2019;133(24):2615–2618. doi:10.1182/blood.2019000406
- Waxman AJ, Mink PJ, Devesa SS, et al. Racial disparities in incidence and outcome in multiple myeloma: a population-based study. Blood. 2010;116(25):5501–5506. doi:10.1182/blood-2010-07-298760
- Abouzaid S, Parikh K, Zhou Z-Y, et al. Disparities in treatment patterns and outcomes between Caucasian and African American patients with multiple myeloma (MM). J Clin Oncol. 2016;34(Suppl 15?):8022–8022. doi:10.1200/JCO.2016.34.15_suppl.8022
- Ailawadhi S, Frank RD, Advani P, et al. Racial disparity in utilization of therapeutic modalities among multiple myeloma patients: a SEER-medicare analysis. Cancer Med. 2017;6(12):2876–2885. doi:10.1002/cam4.1246
- Fiala MA, Wildes TM, Vij R. Racial disparities in the utilization of novel agents for frontline treatment of multiple myeloma. Clin Lymphoma Myeloma Leuk. 2020;20(10):647–651. doi:10.1016/j.clml.2020.04.018
- Ailawadhi S, Aldoss IT, Yang D, et al. Outcome disparities in multiple myeloma: a SEER-based comparative analysis of ethnic subgroups. Br J Haematol. 2012;158(1):91–98. doi:10.1111/j.1365-2141.2012.09124.x
- Ailawadhi S, Frank RD, Sharma M, et al. Trends in multiple myeloma presentation, management, cost of care, and outcomes in the medicare population: a comprehensive look at racial disparities. Cancer. 2018;124(8):1710–1721. doi:10.1002/cncr.31237
- Ailawadhi S, Jagannath S, Lee HC, et al. Association between race and treatment patterns and survival outcomes in multiple myeloma: a connect MM registry analysis. Cancer. 2020;126(19):4332–4340. doi:10.1002/cncr.33089
- Sonderman JS, Bethea TN, Kitahara CM, et al. Multiple myeloma mortality in relation to obesity among African Americans. J Natl Cancer Inst. 2016;108(10):djw120. doi:10.1093/jnci/djw120
- Birnbaum B, Nussbaum N, Seidl-Rathkopf K, et al. Model-assisted cohort selection with bias analysis for generating large-scale cohorts from the EHR for oncology research. 2020. (Accessed November 7, 2022). Available from: https://arxiv.org/abs/2001.09765.
- Ma X, Long L, Moon S, et al. Comparison of population characteristics in real-world clinical oncology databases in the US: flatiron Health, SEER, and NPCR. 2020.(Accessed November 7, 2022). Available from: https://doi.org/10.1101/2020.03.16.20037143.
- Ailawadhi S, Jacobus S, Sexton R, et al. Disease and outcome disparities in multiple myeloma: exploring the role of race/ethnicity in the cooperative group clinical trials. Blood Cancer J. 2018;8(7):67. doi:10.1038/s41408-018-0102-7