Abstract
Despite the use of autologous hematopoietic cell transplantation (AHCT) in treatment of multiple myeloma (MM) for almost 40 years and its persistence as standard of care in transplantation-eligible patients with MM even after the advent of novel agents, AHCT remains underutilized, especially in racial and ethnic minority populations. As part of a multipronged effort to quantify disparities in AHCT utilization in MM by race and ethnicity and over time in our own cancer center, we conducted an institutional review of all new patients seen at an academic transplant center for consultation for MM between 2012 and 2022, to calculate AHCT utilization and investigate the factors associated with AHCT utilization. Race and ethnicity were self-reported. Baseline characteristics were analyzed in 3 groups: non-Hispanic White (NHW), non-Hispanic Black (NHB), and Others. Reasons for not undergoing AHCT in the EHR were recorded. Multivariate analyses evaluated the effect of group on AHCT utilization, controlling for covariates related to patients not undergoing AHCT by overall cohort and consult period. Of the 1266 patients, 13.4% were NHB. The median age at consult was 66 (IQR, 23–97) years overall, 66 (IQR, 23–97) years for NHWs, 63 (IQR, 25–85) years for NHBs, and 59.5 (IQR, 31–79) years for Others (P < .01). AHCT utilization was 76% overall, 64.7% in NHBs, 76.8% in Others, and 77.8% in NHWs (P < .01). Age, cytogenetics, stage, comorbidities, and time from diagnosis to consult were associated with receipt of AHCT. From 2012–2017 to 2018–2022, NHB AHCT utilization increased from 57.5% to 69.8% (P = .10). For those who did not receive AHCT, patient preference, older age, comorbidity, early mortality, and lack of caregiver support were the most frequently documented reasons. The NHW group had greater AHCT utilization compared to the NHB group (odds ratio [OR], 3.32; 95% confidence interval [CI], 2.17–5.08; P < .0001). Absent cardiac (OR, 1.88; 95% CI, 1.35–2.62; P = .0002) or renal comorbidity (OR, 3.23; 95% CI, 2.03–5.15; P < .0001) was associated with receipt of AHCT. Older age at consult (OR, .89; 95% CI, .87-.90; P < .0001) and longer time from diagnosis to consult (OR, .97; 95% CI, .95-.98; P < .0001) were associated with lower AHCT utilization. While AHCT utilization increased from 2012–2017 to 2018–2022 in NHBs compared to NHWs, it remained significantly lower. Racial and ethnic AHCT underutilization has improved over time, but disparities persist. Younger age at consult, shorter time from diagnosis to consult, and lack of cardiac and renal comorbidities also are associated with AHCT utilization.
Keywords: Multiple myeloma, Autologous hematopoietic cell transplantation, Healthcare access, Health disparities, Health equity, Race, Ethnicity
INTRODUCTION
Although autologous hematopoietic cell transplantation (AHCT) has been used to treat multiple myeloma (MM) for almost 40 years [1] and has remained the standard of care in transplantation-eligible patients with MM even after the advent of novel agents [2], it remains underutilized [3]. National database—based studies show low overall AHCT utilization at approximately 30%, and Black and Hispanic patients consistently have delays in AHCT referral and completion and even lower AHCT utilization [3,4]. The critical importance of improving access to treatment for racial and ethnic minorities is further highlighted by that fact that with equal care, Black and Hispanic patients with MM have superior survival [5,6]. These findings illuminate a key target area for health equity research and interventions in patients with MM.
In addition to differences in treatment access, racial and ethnic minorities have significant epidemiologic differences. Both Black and Hispanic patients are diagnosed with MM at a younger age with more advanced disease compared to White patients [3,6,7]. Compared to White patients with MM, Black patients with MM have more comorbidities and worse performance status, and Hispanic patients with MM have fewer comorbidities and similar performance status [3]. We conducted an institutional analysis to investigate whether any of the above racial and ethnic epidemiologic differences are acting as drivers of racial and ethnic disparities in AHCT utilization. As part of a multipronged effort to quantify disparities in AHCT utilization in MM by race and ethnicity and over time in our own cancer center, we conducted an institutional review of all new patients who were seen at an academic transplantation center for consultation for MM to calculate AHCT utilization and investigate the factors associated with AHCT utilization.
METHODS
Study Design
This retrospective cohort study was conducted after Institutional Review Board approval.
Data Source
Electronic health record (EHR) data were obtained from an academic cancer center in the state of Wisconsin.
Cohort
Individuals with new patient consultations for MM between January 1, 2012, and December 31, 2022, were identified from an administrative institutional tumor registry. Patients who had prior stem cell transplant, had non-MM plasma cell disorders including monoclonal gammopathy of undetermined significance, smoldering multiple myeloma, concurrent amyloidosis, plasma cell leukemia, concurrent hematological malignancy, or concurrent solid tumor were excluded. A complete cohort selection flowchart is provided in Figure 1.
FIGURE 1.
Flow chart of the study population.
Data Collection and Definitions
Sex and self-reported race and ethnicity as documented in the EHR was recorded. Race and ethnicity were classified as non-Hispanic Black (NHB), Non-Hispanic White (NHW), and Others. The Others group comprised patients who self-reported as Hispanic, Asian, and American Indian or Alaska Native (Supplementary Table S1). Age at new patient consultation for AHCT evaluation, date of diagnosis, date of consult, and date of AHCT were recorded. The following also were recorded at the time of consultation: myeloma subtype, cytogenetic risk status, comorbidity status (cardiac, renal, hepatic, pulmonary, prior solid tumor), Karnofsky Performance Status, insurance status (commercial, Medicare, Medicaid, other), and limited English proficiency (LEP) with primary language. The reason for not undergoing AHCT was recorded as best interpreted from EHR documentation.
Early mortality was defined as death within 1 year from consultation. The following were defined as high-risk cytogenetics: del(17p), t(4;14), t(14;16), t(14;20), and gain or amplification of 1q. Cardiovascular comorbidity was defined as atrial fibrillation, atrial flutter, supraventricular tachycardia, sick sinus syndrome, heart block, ventricular arrhythmia, coronary artery disease, ejection fraction <50%, congestive heart failure, or heart valve disease (eg, moderate stenosis, moderate insufficiency, prosthetic heart valve, symptomatic mitral valve prolapse). Hepatic comorbidity was defined as elevated total bilirubin or liver function tests (at least 2 elevated values on 2 different days), hepatitis B/C, or cirrhosis. Renal comorbidity was defined as creatinine >2 mg/dL (at least 2 elevated values on 2 different days, current dialysis, or prior renal transplantation). Pulmonary comorbidity was defined as being on oxygen. Prior solid tumor did not include nonmelanoma skin cancer. Syngeneic allogeneic transplantation after consultation was categorized as AHCT for the analysis. Through a detailed chart review, reasons for not undergoing AHCT for each patient that were documented in the EHR were recorded and categorized into thematic groups, including patient preference, age, comorbidity, early mortality, lack of caregiver support, financial toxicity, and physician preference.
Statistical Analysis
We analyzed baseline characteristics using bivariate analysis between the race and ethnicity groups and clinical characteristics using the Kruskal-Wallis test for continuous variables and the Fisher exact test or Pearson χ2 test as appropriate for categorical variables. We also analyzed AHCT utilization by race by consult year period: 2012–2017 or 2018–2022. A logistic regression model with stepwise selection was developed to evaluate the effect of race and ethnicity on AHCT while controlling for other covariates. The following covariates were considered: race and ethnicity group (main effect), age at consult (as a continuous covariate), sex, cytogenetics, renal comorbidity, cardiac comorbidity, prior solid tumor, year of consult, and time from diagnosis to consult (as a continuous covariate). A significance level of .1 was used for entry into the model, and in the final model, factors with statistical significance at P < .05 were retained. We also performed a sensitivity analysis by time, constructing separate logistic regression models by consult year periods 2012–2017 and 2018–2022 were constructed. All statistical analyses were performed with SAS version 9.4 (SAS Institute).
RESULTS
Baseline Characteristics
Table 1 presents the characteristics of all patients who had new MM consultations between 2012 and 2022 by race and ethnicity. We identified 1266 patients, of whom 1040 (82.1%) were NHW, 170 (13.4%) were NHB, and 56 (4.4%) were Others. The median age at consult was 66 years overall (range, 23 to 97 years), 66 (range, 23 to 97) years for HNWs, 63 (range, 25 to 85) years for NHBs, and 59.5 (range, 31 to 79) years for Others (P < .01). NHBs included a greater proportion of female patients (55.3%) compared to NHWs (42.5%) and Others (41.1%) (P < .01). The 3 groups had similar cytogenetic risk, International Staging System (ISS) staging, myeloma subtype, comorbidities, and time from diagnosis to consult.
Table 1.
Characteristics of MM Patients Consulted between 2012 and 2022, by Race/Ethnicity
| Characteristic | Total | NHW | NHB | Other* | P Value |
|---|---|---|---|---|---|
| Number of patients (%) | 1266 | 1040 (82.1) | 170 (13.4) | 56 (4.4) | |
| Age at consult, yr, median (range) | 66.0 (23.0–97.0) | 66.0 (23.0–97.0) | 63.0 (25.0–85.0) | 59.5 (31.0–79.0) | <.01† |
| Age ≥75 years at consult, n (%) | 213 (16.8) | 188 (18.1) | 18 (10.6) | 7 (12.5) | .04‡ |
| Female sex, n (%) | 559 (44.2) | 442 (42.5) | 94 (55.3) | 23 (41.1) | <.01‡ |
| Myeloma subtype, n (%) | .06‡ | ||||
| IgG | 741 (58.5) | 597 (57.4) | 111 (65.3) | 33 (58.9) | |
| IgA | 240 (19.0) | 200 (19.2) | 26 (15.3) | 14 (25.0) | |
| Light chain | 237 (18.7) | 203 (19.5) | 29 (17.1) | 5 (8.9) | |
| Other | 34 (2.7) | 30 (2.9) | 3 (1.8) | 1 (1.8) | |
| Missing | 14 (1.1) | 10 (1.0) | 1 (.6) | 3 (5.4) | |
| Cytogenetics, n (%) | .32‡ | ||||
| Standard risk | 686 (54.2) | 555 (53.4) | 104 (61.2) | 27 (48.2) | |
| High risk | 498 (39.3) | 415 (39.9) | 57 (33.5) | 26 (46.4) | |
| Missing | 82 (6.5) | 70 (6.7) | 9 (5.3) | 3 (5.4) | |
| ISS stage at consult, n (%) | .86‡ | ||||
| ISS I | 637 (50.3) | 529 (50.9) | 78 (45.9) | 30 (53.6) | |
| ISS II | 333 (26.3) | 271 (26.1) | 48 (28.2) | 14 (25.0) | |
| ISS III | 201 (15.9) | 160 (15.4) | 32 (18.8) | 9 (16.1) | |
| Missing | 95 (7.5) | 80 (7.7) | 12 (7.1) | 3 (5.4) | |
| Cardiac comorbidity, n (%) | 290 (22.9) | 242 (23.3) | 40 (23.5) | 8 (14.3) | .31‡ |
| Renal comorbidity, n (%) | 119 (9.4) | 89 (8.6) | 23 (13.5) | 7 (12.5) | .08‡ |
| Pulmonary comorbidity, n (%) | 25 (2.0) | 17 (1.6) | 5 (2.9) | 3 (5.4) | .02‡ |
| Prior solid tumor, n (%) | 148 (11.7) | 126 (12.1) | 19 (11.2) | 3 (5.4) | .34‡ |
| Year of consult, n (%) | .82§ | ||||
| 2012–2017 | 578 (45.7) | 479 (46.1) | 74 (43.5) | 25 (44.6) | |
| 2018–2022 | 688 (54.3) | 561 (53.9) | 96 (56.5) | 31 (55.4) | |
| Receipt of AHCT, n (%) | 962 (76.0) | 809 (77.8) | 110 (64.7) | 43 (76.8) | <.01§ |
| Time from diagnosis to consult, mo, median (range) | 1.5 (.0–257.2) | 1.6 (.0–257.2) | 1.3 (.0–128.9) | 1.4 (.0–68.3) | .78† |
| Time from diagnosis to AHCT, mo, median (range) | 5.8 (.3–126.0) | 5.7 (3.0–126.0) | 6.4 (3.8–36.6) | 6.2 (.3–23.6) | .02† |
| Time from consult to AHCT, mo, median (range) | 4.2 (.0–125.0) | 4.1 (.7–125.0) | 4.7 (1.3–36.1) | 4.4 (.0–10.7) | .21† |
A detailed breakdown of the Others group is shown in Supplementary Table S1.
Hypothesis testing with the Kruskal-Wallis test.
Hypothesis testing with the Fisher exact test via Monte Carlo.
Hypothesis testing with the Pearson χ2 test.
AHCT Utilization
Table 1 shows overall AHCT utilization by race and ethnicity and the time from diagnosis to AHCT and time from consultation to AHCT. The AHCT utilization rate was 76% overall, with the lowest rate in NHBs at 64.7%, compared to Others at 76.8% and NHWs at 77.8% (P < .01). NHB patients with MM had a longer median time from diagnosis to AHCT compared to NHWs and Others: 6.4 (range, 3.8 to 36.6) months, 5.7 (range, 3.0 to 126.0) months, and 6.2 (range, .3 to 23.6) months, respectively (P = .02). The median time from consult to AHCT did not differ by group; 4.7 (range, 1.3 to 36.1) months for NHBs, 4.1 (range, .7 to 125.0) months for NHWs, and 4.4 (range, .0–10.7) months for Others (P = .21).
Factors Associated with AHCT Utilization
Table 2 presents characteristics of all patients by AHCT utilization. Compared to patients who did not undergo AHCT, the AHCT recipients were younger (63 years versus 74.5 years; P < .01), had higher proportions of high-risk cytogenetics (40.6% versus 35.3%; P < .01) and ISS stage I at consult (55.8% versus 32.9%; P < .01), lower proportions of NHBs (11.4% versus 19.7%; P < .01) and comorbidities, and shorter time from diagnosis to consult (1.4 months versus 2.3 months; P < .01) (Table 2).
Table 2.
Characteristics of MM Patients Consulted from 2012–2022, by AHCT Utilization
| Characteristic | Total | No | Yes | P Value |
|---|---|---|---|---|
| Number of patients | 1266 | 304 | 962 | |
| Age at consult, yr, median (range) | 66.0 (23.0–97.0) | 74.5 (25.0–97.0) | 63.0 (23.0–82.0) | <.01* |
| Age ≥75 yr at consult, n (%) | 213 (16.8) | 152 (50.0) | 61 (6.3) | <.01† |
| Female sex, n (%) | 559 (44.2) | 145 (47.7) | 414 (43.0) | .14† |
| Race/ethnicity, n (%) | <.01† | |||
| NHW | 1040 (82.1) | 231 (76.0) | 809 (84.1) | |
| NHB | 170 (13.4) | 60 (19.7) | 110 (11.4) | |
| Others | 56 (4.4) | 13 (4.3) | 43 (4.5) | |
| Myeloma subtype, n (%) | .29† | |||
| IgG | 741 (58.5) | 192 (63.2) | 549 (57.1) | |
| IgA | 240 (19.0) | 53 (17.4) | 187 (19.4) | |
| Light chain | 237 (18.7) | 50 (16.4) | 187 (19.4) | |
| Other | 34 (2.7) | 5 (1.6) | 29 (3.0) | |
| Missing | 14 (1.1) | 4 (1.3) | 10 (1.0) | |
| Cytogenetics, n (%) | <.01† | |||
| Standard risk | 686 (54.2) | 159 (52.3) | 527 (54.8) | |
| High risk | 498 (39.3) | 107 (35.2) | 391 (40.6) | |
| Missing | 82 (6.5) | 38 (12.5) | 44 (4.6) | |
| ISS stage at consult, n (%) | <.01† | |||
| ISS I | 637 (50.3) | 100 (32.9) | 537 (55.8) | |
| ISS II | 333 (26.3) | 96 (31.6) | 237 (24.6) | |
| ISS III | 201 (15.9) | 79 (26.0) | 122 (12.7) | |
| Missing | 95 (7.5) | 29 (9.5) | 66 (6.9) | |
| Cardiac comorbidity, n (%) | 290 (22.9) | 124 (40.8) | 166 (17.3) | <.01† |
| Renal comorbidity, n (%) | 119 (9.4) | 53 (17.4) | 66 (6.9) | <.01† |
| Pulmonary comorbidity, n (%) | 25 (2.0) | 21 (6.9) | 4 (.4) | <.01† |
| Prior solid tumor, n (%) | 148 (11.7) | 47 (15.5) | 101 (10.5) | .02† |
| Year of consult, n (%) | .93† | |||
| 2012–2017 | 578 (45.7) | 139 (45.7) | 439 (45.6) | |
| 2018–2022 | 688 (54.3) | 165 (54.3) | 523 (54.4) | |
| Time from diagnosis to consult, mo, median (range) | 1.5 (.0–257.2) | 2.3 (.0–257.2) | 1.4 (.0–120.5) | <.01* |
| Time from diagnosis to AHCT time, mo, median (range) | NE | 5.8 (.3–126.0) | ||
| Time from consult to AHCT time, mo, median (range) | NE | 4.2 (.0–125.0) |
NE indicates not evaluable.
Hypothesis testing with the Kruskal-Wallis test.
Hypothesis testing with the Pearson χ2 test.
On multivariate analysis, the NHW group was more likely to undergo AHCT compared with either the NHB group (OR, 3.32; 95% CI, 2.17 to 5.08; P < .0001) or the Other group (OR, 2.27; 95% Cl, 1.07 to 4.84; P = .0329) (Table 3). Patients also were more likely to undergo AHCT if they did not have a cardiac comorbidity (OR, 1.88; 95% CI, 1.35 to 2.62; P = .0002) or renal comorbidity (OR, 3.23; 95% CI, 2.03 to 5.15; P < .0001). On the other hand, patients were less likely to undergo AHCT if they were older at consult (OR, .89; 95% CI, .87 to .90; P < .0001) or had a longer time from diagnosis to consult (OR, .97; 95% CI, .95 to .98; P < .0001).
Table 3.
MVA Model Factors Associated with AHCT Utilization from 2012 to 2022
| Effect | Events/N | OR | 95% CI | P Value |
|---|---|---|---|---|
| Race/ethnicity | ||||
| NHB | 110/170 | Reference | <.0001 | |
| NHW | 809/1040 | 3.32 | 2.17–5.08 | <.0001 |
| Other | 43/56 | 1.46 | .64–3.31 | .3653 |
| Contrast: NHW vs Other | 2.27 | 1.07–4.84 | .0329 | |
| Age at consult | 962/1266 | .89 | .87-.90 | <.0001 |
| Time from diagnosis to consult | 962/1266 | .97 | .95-.98 | <.0001 |
| Cardiac comorbidity | ||||
| Yes | 166/290 | Reference | .0002 | |
| No | 796/976 | 1.88 | 1.35–2.62 | .0002 |
| Renal Comorbidity | ||||
| Yes | 66/119 | Reference | <.0001 | |
| No | 896/1147 | 3.23 | 2.03–5.15 | <.0001 |
The model was built using the following covariates: race/ethnicity, age, sex, cytogenetics, renal comorbidity, cardiac comorbidity, prior solid tumor, year of consult, and time from diagnosis to consult.
Change in AHCT Utilization over Time
Figure 2 shows a visual representation of AHCT trends over time by race and ethnicity. When comparing AHCT utilization by time period, NHW AHCT utilization decreased from 79.1% in 2012–2017 to 76.7% in 2018–2022 (P = .34), while NHB AHCT utilization increased from 57.5% to 69.8% between these 2 periods (P = .10), an absolute increase of almost 13%. Table 4 shows the multivariate analysis models by time period., The likelihood of the NHB group undergoing AHCT compared to the NHW group increased from an OR of .16 (95% CI, .08 to .31; P < .0001) for 2012–2017 to .48 (95% CI .28 to .84; P = .0098) for 2018–2022 but still remained significantly lower. However, for those same time periods, the likelihood of the Other group receiving AHCT compared to the NHW group increased from an OR of .27 (95% CI, .09 to .80; P = .0181) to .74 (95% CI, .24 to 2.31; P = .6068) and was no longer significantly different. Older age at consult and longer time from diagnosis to consult independently predicted a lower likelihood of AHCT in both time periods. Lack of cardiac comorbidity was not independently associated with AHCT in 2012–2017 but was independently associated with AHCT in 2018–2022. The opposite was true for renal comorbidity.
FIGURE 2.
Trends in AHCT utilization among MM patients consulted between 2012 and 2022.
Table 4.
Multivariate Analysis Separated by Time of Consult, 2012–2017 vs 2018–2022
| Effect | 2012–2017 | 2018–2022 | ||||||
|---|---|---|---|---|---|---|---|---|
| Events/N | OR | 95% CI | P Value | Events/N | OR | 95% CI | P Value | |
| Race/ethnicity | ||||||||
| NHW | 379/479 | Reference | <.0001 | 430/561 | Reference | .0346 | ||
| NHB | 43/74 | .16 | .08-.31 | <.0001 | 67/96 | .48 | .28-.84 | .0098 |
| Other | 17/25 | .27 | .09-.80 | .0181 | 26/31 | .74 | .24–2.31 | .6068 |
| Contrast: Other vs NHB | / | .59 | .18–1.95 | .3841 | / | .65 | .19–2.18 | .4827 |
| Age at consult | 439/578 | .86 | .84-.89 | <.0001 | 523/688 | .90 | .88-.93 | <.0001 |
| Time from diagnosis to consult | 439/578 | .97 | .96-.99 | .0041 | 523/688 | .95 | .93-.98 | .0002 |
| Cardiac comorbidity | ||||||||
| Yes | 71/125 | Reference | .082 | 95/165 | Reference | .002 | ||
| No | 368/453 | 1.59 | .94–2.67 | .0823 | 428/523 | 2.27 | 1.47–3.53 | .002 |
| Renal comorbidity | ||||||||
| Yes | 31/63 | Reference | <.0001 | 35/56 | Reference | .0531 | ||
| No | 408/515 | 4.70 | 2.43–9.09 | <.0001 | 488/632 | 1.98 | .99–3.95 | .0531 |
Barriers to AHCT Utilization
To better understand barriers to AHCT, reasons for not undergoing AHCT as best determined through EHR documentation were compiled for all patients who did not undergo AHCT (Table 5). For those who did not undergo AHCT, patient preference (31%), age (28%), comorbidity (22%), early mortality (9%), and lack of caregiver support (8%) were the most frequently documented reasons in the EHR. Of note, 21.7% of patients had multiple documented reasons for not undergoing AHCT Compared to our research cohort baseline of 13.4% NHB patients, a higher proportion of NHB patients had patient preference, comorbidity, lack of caregiver support, financial toxicity, or physician preference as their reason for not undergoing AHCT.
Table 5.
Reasons for Not Undergoing AHCT Documented in the EHR*
| Reason for no AHCT | Total, n (N = 304) | NHW, n (%) (N = 231) | NHB, n (%) (N = 60) | Other, n (%) (N = 13) |
|---|---|---|---|---|
| Patient preference | 93 | 64 (68.9) | 25 (26.9) | 4 (4.3) |
| Age | 86 | 77 (89.6) | 8 (9.3) | 1 (1.2) |
| Comorbidity | 67 | 47 (70.1) | 18 (26.9) | 2 (3.0) |
| Early mortality | 27 | 23 (85.2) | 3 (11.1) | 1 (3.7) |
| Lack of caregiver support | 23 | 15 (65.2) | 8 (34.8) | 0 (0) |
| Financial toxicity | 8 | 4 (50.0) | 3 (37.5) | 1 (12.5) |
| Physician preference | 7 | 5 (71.4) | 2 (28.6) | 0 (0) |
Other: insufficient stem cells, medical mistrust, health literacy, food insecurity, home safety, transportation, comfort cares, receipt of AHCT, and unclear.
Numbers do not add up to total numbers as groups are not mutually exclusive.
DISCUSSION
In this single-center analysis of MM consults seen at an academic cancer center, we made several observations. First, we found that the NHB and Other groups with MM were younger at consult compared to NHWs, which is consistent with previously well-established epidemiology of NHB and Hispanic patients being diagnosed at a younger age [3,6,8]. Second, we found that the NHB group had a significantly longer time from diagnosis to AHCT and lower AHCT utilization compared to the NHW group, which also is consistent with prior studies [3,9]. Third, covariates associated with AHCT receipt included younger age, NHW group, high-risk cytogenetics, ISS stage I, lack of comorbidity, and a shorter time from diagnosis to consult. Fourth, patient preference, age, comorbidity, early mortality, and lack of caregiver support were the most frequently reported reasons in the EHR for not undergoing AHCT. Fifth, in multivariate analysis, the NHW group was significantly associated with AHCT utilization compared to both the NHB and Other groups. Additional significant covariates included younger age at consult, shorter time from diagnosis to consult, and lack of cardiac and renal comorbidities. Finally, while AHCT utilization in the NHB group increased significantly over time, the NHB group remains less likely to receive AHCT compared to the NHW group.
Although many studies in the literature have documented significant racial and ethnic disparities in AHCT utilization, our study represents one of few assessing racial and ethnic AHCT disparities over time [3,10,11]. Ailawadhi et al. [10] used the Surveillance Epidemiology and End Results (SEER)—Medicare (2007–2013) database and showed that true AHCT utilization increased significantly for White and African American patients with MM, with no significant differences in trends between racial and ethnic groups. Schriber et al. [3] used the Center for International Blood and Marrow Transplant Research (CIBMTR) and SEER databases to estimate a AHCT utilization rate and found that from 2008 to 2014, AHCT use increased for all racial and ethnic groups, but that the increase was substantially higher by proportion from 2008 to 2013 but remained dismally low among NHB and Hispanic patients compared to NHW patients [3]. In a recently published study, Khera et al. [11] used the CIMBTR database and showed that rate of change of AHCT use from 2009 to 2018 was 2- to 3-fold times higher in racial and ethnic minorities compared to NHW patients [11]. However, these studies were limited by a focus predominately on those who underwent HCT, with minimal data on those who did not and on reasons why AHCT was not done. In a recently published study, Ailawadhi et al. [12] used the National Cancer Database to evaluate a comprehensive list of sociodemographic factors in more than 110,000 patients with MM between 2004 and 2013 and showed that AHCT utilization increased over time for all racial and ethnic groups except the non-Hispanic Asian group [12].
Similar to that National Cancer Database study [12], the strength of our large single-center retrospective study investigating AHCT utilization over time is the ability to study both patients who did and those did not undergo AHCT. Additionally, our study builds on these prior studies by providing utilization trends in more recent years and assessing reasons for not undergoing AHCT as documented in the EHR in all patients who did not undergo AHCT. Our detailed EHR analysis captured barriers to transplantation that typically cannot be captured through retrospective analyses, such as patient preference, lack of caregiver support, financial toxicity, medical mistrust, and health literacy.
Comparing consult periods 2012–2017 and 2018–2022, NHB AHCT utilization increased by almost 13%, from 57.5% to 69.8%, but remained significantly lower than the 2018–2022 NHW AHCT utilization rate of 76.7%. Similarly, our multivariate analysis by time period showed that the likelihood of AHCT for the NHB group compared to the NHW group increased by 3-fold, from an OR of .16 in 2012–2017 to .48 in 2018–2022. Additionally, our multivariate analysis by time period showed a significant improvement in AHCT for the Other group, with an increase in OR from .27 in 2012–2017 to .74 in 2018–2022, which was no longer statistically different from the NHW group.
The reason for this improvement over time is unclear, as there were no significant institutional policy changes in terms of care of patients with MM and supportive resources offered, such as financial, housing, and transportation assistance. A possible driver in the improvement of AHCT utilization in the NHB group is that in 2018–2022, having a renal comorbidity no longer significantly decreased the likelihood of AHCT. This is important, because 13.5% of the NHB group had a renal comorbidity, compared to 8.6% of the NHW group, meaning that performing ACHT in more patients with renal disease may mean performing ACHT in more NHB patients. A 2017 CIBMTR publication in which our group was closely involved demonstrating that AHCT is safe and effective in patients with renal insufficiency might have led to more liberalization of AHCT use in those with renal insufficiency or renal failure [13].
Despite the 3-fold increase in AHCT utilization between time periods and despite referrals to a transplant center, the NHB group still was still >50% less likely than the NHW group to undergo AHCT in 2018–2022. Patients who are seen at a transplant center are already a selected subset, for which certain barriers, such as insurance, knowledge or interest in AHCT, or appropriate AHCT referral, are not applicable. These results indicate that poor access to a transplant center is not the sole barrier driving racial and ethnic AHCT utilization differences. Our EHR analysis of reasons for not undergoing AHCT may offer insight into other, more difficult to quantify AHCT barriers.
Interestingly, our findings suggest that NHB patients were more likely to have patient preference, comorbidities, lack of caregiver support, financial toxicity, and physician preference as reasons for not receiving AHCT. Other factors that were identified, including medical mistrust, health literacy, food insecurity, and transportation, merit further exploration. Social determinants of health and structural issues that could be impacting the ability of Black patients receiving AHCT even after a referral to a transplant center need further investigation. The significant percentage of patients with multiple reasons documented also highlights the complex nature of these barriers and lends further support to the need for prospective mixed-method studies.
A limitation of our study is that because our results represent the experience of a single tertiary care transplant center, they might not be generalizable to other healthcare centers. It also is important to note that patients referred to our cancer center for the most part needed to have insurance. This is an access issue across most cancer centers throughout the United States. Although our analysis examined barriers after referral for AHCT, it is important to acknowledge the significant barriers for AHCT referral, especially among those who are uninsured. Although MM is a disease of older adults for which many will have Medicare, the younger age at diagnosis in NHB patients suggests that more NHB patients may be at risk for being uninsured or underinsured, which would not allow them to be seen at transplant center and thus not have access to transplant. Additionally, social determinants of health that impact the ability to access AHCT often are not documented clearly or at all in the EHR. Thus, it is difficult to interpret any quantification of reasons for not receiving AHCT. Another limitation is related to the demographic breakdown of our cancer center patient population. Although our cohort is representative of the proportion of Black patients in our catchment area of 9% to 10% [14], the small numbers of Hispanic, Asian, and American Indian or Alaska Native did not allow for subgroup analyses, so they were combined as an “Other” category. This may be masking disparities in these racial and ethnic groups at our cancer center, especially in light of well-documented disparities in Hispanic patients with MM [15]. Owing to the significant heterogeneity of the Other group, interpreting results for this group, such as the increase in OR over time, is not straightforward. Further studies and more granular and disaggregated racial and ethnic data are needed to better quantify disparities in other minorities.
CONCLUSION
We found that NHB patients had a longer time from diagnosis to AHCT and were less likely in bivariate and multivariate analyses to undergo AHCT compared to NHW patients. However, although racial and ethnic AHCT utilization has greatly improved over time, significant disparities still persist for NHB patients. In multivariate analysis, younger age at consult, shorter time from diagnosis to consult, and lack of cardiac and renal comorbidities also predicted AHCT utilization. Future studies focused on statewide geospatial analyses and qualitative research in patients with MM and physicians treating MM are needed to better understand barriers to AHCT with a goal of continuing to improve health equity in MM.
Supplementary Material
ACKNOWLEDGMENTS
Financial disclosure:
Research presented in this work was supported by the American Society of Hematology Hematology Opportunities for the Next Generation of Research Scientists (HONORS) Award (to J.F.W.) and an Advancing a Healthier Wisconsin Endowment Collaborative for Healthcare Delivery Science grant (to A.D.). J.F.W. is supported by the National Heart, Lung, and Blood Institute (1R38HL167238-01 grant).
Conflict of interest statement:
B.D. reports institutional research funding from Bristol Myers Squibb, Janssen, Carsgen, Arcellx, Ichnos, Sanofi, C4 Therapeutics, Natera, and Gracell; advisory board fees from Bristol Myers Squibb, Kite, Arcellx, Janssen, Sanofi, Pfizer, Genentech, and Natera; and speakers bureau fees from Bristol Myers Squibb, Karyopharm, Janssen, Sanofi, and Pfizer. M.M. reports institutional research funding from Sanofi SA, Bristol-Myers Squibb, and Celgene; and advisory board or consulting fees from Sanofi SA, Bristol-Myers Squibb, Celgene, Pfizer, Janssen, and Legend Biotech. M.C.P. reports institutional research funding from BMS, Novartis, Kite Pharma, and Janssen and consulting fees from Novartis, Gilead, and BMS. A.D. reports institutional research funding from AbbVie, Caelum, Janssen, Novartis, Prothena, and Regeneron and advisory board or consulting fees from BMS, Janssen, Pfizer, and Prothena. The other authors have no conflicts of interest to report.
Footnotes
SUPPLEMENTARY MATERIALS
Supplementary material associated with this article can be found in the online version at doi:10.1016/j.jtct.2024.09.010.
Data availability statement:
Data that support the findings of this study are available from the corresponding author on reasonable request.
REFERENCES
- 1.Barlogie B, Hall R, Zander A, Dicke K, Alexanian R. High-dose melphalan with autologous bone marrow transplantation for multiple myeloma. Blood. 1986;67:1298–1301. [PubMed] [Google Scholar]
- 2.Cavo M, Gay F, Beksac M, et al. Autologous haematopoietic stem-cell transplantation versus bortezomib-melphalan-prednisone, with or without bortezomib-lenalidomide-dexamethasone consolidation therapy, and lenalidomide maintenance for newly diagnosed multiple myeloma (EMN02/HO95): a multicentre, randomised, open-label, phase 3 study. Lancet Haematol. 2020;7:e456–e468. [DOI] [PubMed] [Google Scholar]
- 3.Schriber JR, Hari PN, Ahn KW, et al. Hispanics have the lowest stem cell transplant utilization rate for autologous hematopoietic cell transplantation for multiple myeloma in the United States: A CIBMTR report. Cancer. 2017;123:3141–3149. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Bhatnagar V, Wu Y, Goloubeva OG, et al. Disparities in black and white patients with multiple myeloma referred for autologous hematopoietic transplantation: a single center study. Cancer. 2015;121:1064–1070. [DOI] [PubMed] [Google Scholar]
- 5.Dong J, Garacci Z, Buradagunta CS, et al. Black patients with multiple myeloma have better survival than white patients when treated equally: a matched cohort study. Blood Cancer J. 2022;12:34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kaur G, Mejia Saldarriaga M, Shah N, et al. Multiple myeloma in Hispanics: incidence, characteristics, survival, results of discovery, and validation using real-world and Connect MM Registry data. Clin Lymphoma Myeloma Leuk. 2021;21:e384–e397. [DOI] [PubMed] [Google Scholar]
- 7.Marinac CR, Ghobrial IM, Birmann BM, Soiffer J, Rebbeck TR. Dissecting racial disparities in multiple myeloma. Blood Cancer J. 2020;10:19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bhutani M, Blue BJ, Cole C, et al. Addressing the disparities: the approach to the African American patient with multiple myeloma. Blood Cancer J.. 2023;13:189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hari PN, Majhail NS, Zhang MJ, et al. Race and outcomes of autologous hematopoietic cell transplantation for multiple myeloma. Biol Blood Marrow Transplant. 2010;16:395–402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Ailawadhi S, Parikh K, Abouzaid S, et al. Racial disparities in treatment patterns and outcomes among patients with multiple myeloma: a SEER-Medicare analysis. Blood Adv.. 2019;3:2986–2994. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Khera N, Ailawadhi S, Brazauskas R, et al. Trends in volumes and survival after hematopoietic cell transplantation in racial/ethnic minorities. Blood Adv. 2024; 8:3497–3506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ailawadhi S, Adu Y, Frank RD, et al. Factors determining utilization of stem cell transplant for initial therapy of multiple myeloma by patient race: exploring intra-racial healthcare disparities. Blood Cancer J. 2024;14:86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Mahindra A, Hari P, Fraser R, et al. Autologous hematopoietic cell transplantation for multiple myeloma patients with renal insufficiency: a center for international blood and marrow transplant research analysis. Bone Marrow Transplant.. 2017;52:1616–1622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.United States Census Bureau. 2020. Census: Redistricting File (Public Law 94–171) Dataset. August 12, 2021. Accessed August 30, 2024. [Google Scholar]
- 15.Anampa-Guzmán A, Alam ST, Abuali I, Al Hadidi S. Health disparities experienced by Hispanic Americans with multiple myeloma: a systematic review. Clin Hematol Int.. 2023;5:29–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data that support the findings of this study are available from the corresponding author on reasonable request.