Willingness to Travel for Cellular Therapy: The Influence of Follow-Up Care Location, Oncologist Continuity, and Race

PURPOSE:

Patients weigh competing priorities when deciding whether to travel to a cellular therapy center for treatment. We conducted a choice-based conjoint analysis to determine the relative value they place on clinical factors, oncologist continuity, and travel time under different post-treatment follow-up arrangements. We also evaluated for differences in preferences by sociodemographic factors.

METHODS:

We administered a survey in which patients with diffuse large B-cell lymphoma selected treatment plans between pairs of hypothetical options that varied in travel time, follow-up arrangement, oncologist continuity, 2-year overall survival, and intensive care unit admission rate. We determined importance weights (which represent attributes' value to participants) using generalized estimating equations.

RESULTS:

Three hundred and two patients (62%) responded. When all follow-up care was at the center providing treatment, plans requiring longer travel times were less attractive (v 30 minutes, importance weights [95% CI] of –0.54 [–0.80 to –0.27], –0.57 [–0.84 to –0.29], and –0.17 [–0.49 to 0.14] for 60, 90, and 120 minutes). However, the negative impact of travel on treatment plan choice was mitigated by offering shared follow-up (importance weights [95% CI] of 0.63 [0.33 to 0.93], 0.32 [0.08 to 0.57], and 0.26 [0.04 to 0.47] at 60, 90, and 120 minutes). Black participants were less likely to choose plans requiring longer travel, regardless of follow-up arrangement, as indicated by lower value importance weights for longer travel times.

CONCLUSION:

Reducing travel burden through shared follow-up may increase patients' willingness to travel to receive cellular therapies, but additional measures are required to facilitate equitable access.

INTRODUCTION

Recent advances in cellular therapy have expanded treatment options for patients with hematologic malignancies. Between August 2017 and February 2021, four chimeric antigen receptor T-cell (CAR-T) products^1-5 received US Food and Drug Administration approval, three of which carry indications for relapsed or refractory diffuse large B-cell lymphoma (DLBCL). Both CAR-T and autologous stem-cell transplantation (ASCT) offer patients with DLBCL a chance to achieve long-term disease control,⁶ but are only available at a limited number of specialized centers.⁷

One third of the US population lives more than an hour away from a transplant center,⁸ and travel requirements may be even greater to access CAR-T.⁷ Adding to patients' travel burden, ongoing outpatient follow-up after receiving cellular therapy might also be required.^9,10 Difficulty in overcoming geographic barriers to treatment can affect receipt of cancer therapy,^11-15 increase spending,¹⁶ reduce affordability,¹⁶ and impair care continuity.¹⁶ Previous studies have identified disparities in access to hematopoietic stem-cell transplantation^17-19; transportation and other resource limitations^20,21 can similarly make the selection of a cellular therapy–based treatment plan more difficult for some patients.

When selecting between treatment plans offered at different sites, one of which requires travel to a cellular therapy center, patients must weigh competing priorities. Two aspects of patient preferences that can affect their decision are a desire to be treated close to home^22,23 and to have provider continuity.²⁴ Continuity may be especially important for Black patients, as historical and personal mistreatment²⁵ may result in lower levels of trust in health care providers,^26-28 but longstanding^29,30 and high-quality^27,31 relationships with physicians possessing strong communication skills^28,31 may help mitigate this barrier. One way to overcome geographic treatment barriers resulting from these two preferences is to enhance collaboration between cellular therapy centers and oncologists in patients' own communities. Collaborative follow-up care after cellular therapy could reduce ongoing travel requirements and help patients retain continuity with their community oncologist. Although some centers have used collaborative follow-up following hematopoietic stem-cell transplantation, utilization varies⁹ and whether such models are acceptable to patients remains unknown. Moreover, some aspects of this approach may be more important to patients than others (eg, reduced travel burden v enhanced continuity with their current oncologist) and the relative importance of these aspects may vary depending on patient characteristics. Yet, both the degree to which patients value each factor when making the complex decision of whether to travel for advanced care such as cellular therapy and whether preferences differ by demographic factors including race are unknown.

We conducted a choice-based conjoint analysis to (1) determine the relative value that patients with DLBCL place on clinical factors, continuity with their current oncologist, and travel time (under each of three different post-treatment follow-up arrangements) and (2) examine for differences in preferences by sociodemographic factors. We hypothesized that sharing follow-up locally would increase patients' willingness to travel but that Black patients would place a greater value on continuity with their current oncologist, potentially lessening the appeal of this approach.

METHODS

We conducted a cross-sectional study that evaluated patients' preferences using a choice-based conjoint analysis. Conjoint analyses are frequently used in the marketing literature³² and have been applied to assess preferences in health care,^33,34 including oncology.^35-38 In conjoint analysis, participants choose between a series of treatment alternatives with varying levels of the attributes of interest (eg, efficacy or side effect profile). The relative value placed on each attribute is determined on the basis of modeling participants' choices. Conjoint analysis has several advantages over other methods to elicit preferences; most notably, it can independently and simultaneously estimate the relative value of multiple attributes that factor into a complex decision and characterize trade-offs between these attributes. This study was approved by the University of Pennsylvania's Institutional Review Board, and all patients provided informed consent to participate.

Participants

Eligible participants had DLBCL and an outpatient visit at one of the 13 oncology practices within our health system's care network in the year preceding our study. These practices represent a mix of community and academic settings located across a large two-state regional area. Patients with DLBCL were selected as it is a common hematologic malignancy for which both ASCT and CAR-T cells are currently used. Potential participants were identified through the electronic medical record using International Classification of Diseases-10 codes for DLBCL and outpatient encounter data. The team confirmed patients' eligibility via chart review and appropriateness for enrollment with their treating oncologist. Patients were excluded if they could not complete an electronically administered English-language questionnaire or if their oncologist requested they not be contacted.

Instrument Design

We first determined a list of potential attributes and levels on the basis of literature addressing access to cancer treatment,^22,24,36 including cellular therapy,²⁰ the Behavioral Model of Health Services Use,³⁹ clinical literature,^40-46 and study hypotheses. To reduce cognitive burden on respondents, five attributes were selected for inclusion: travel time (30, 60, 90, or 120 minutes of travel each way), oncologist continuity (treatment provided by a patient's current oncologist or a new oncologist), follow-up arrangement (all follow-up appointments at the Cancer Center where treatment is received, shared follow-up, or all local follow-up), efficacy (2-year overall survival rates of 30%, 35%, 40%, 45%, and 50%), and treatment toxicity (intensive care unit [ICU] admission rates of 1%, 5%, and 20%). Efficacy and toxicity levels were based on those reported for patients with relapsed or refractory DLBCL receiving CAR-T, ASCT, and noncellular therapies (targeted and chemoimmunotherapy).^40-46

We used Lighthouse Studio version 9.8.1 (Sawtooth Software, Inc) to create a full profile, balanced overlap design⁴⁷ with two versions of 12 pairs of alternative treatment options. Two versions were included so as to show participants a greater number of attribute-level combinations, thereby increasing the precision of our preference estimates.⁴⁸ The version seen by each participant was randomly assigned. After making their treatment selections, participants answered questions about sociodemographic characteristics, insurance, employment, travel distance and method of transportation to their care site, transportation barriers,⁴⁹ and disease and treatment status. The full questionnaire is shown in the Data Supplement (online only).

The questionnaire was pretested using cognitive debriefing interviews and iterative changes until saturation was reached to ensure (1) item clarity and face validity, (2) acceptable cognitive load, and (3) that no single attribute or level prevented willingness to trade off among other attributes. It was then pilot tested with a 5% sample to determine acceptability. We followed reporting guidelines for conjoint analyses in health care.⁵⁰

Recruitment and Data Collection

Participants were recruited from March 2 to October 30, 2020. Eligible patients received an initial e-mail invitation; those who did not have an e-mail address on file were called by the study team and invited to participate. Patients who did not initially respond received alternating phone calls and e-mailed reminders until a response was received, they declined to participate, or no response was received after six attempts at contact. Data were collected and managed using Research Electronic Data Capture Software.

Analysis

We first determined whether any participants always selected either the first or the second treatment plan in a pair regardless of attribute levels (ie, always chose treatment X or treatment Y). This served as a validity check to confirm that each participant had engaged with the choice tasks, and any participants not doing so were excluded from the analysis. We then examined participants' selections for dominant preferences (when participants always select the treatment alternative containing the more desirable level of a particular attribute) in the total cohort and by race.

Preferences for specific attributes were then estimated using a generalized estimating equation (GEE) logistic regression model assuming an independence working correlation structure and robust standard errors.⁵¹ We used GEE to account for the intraparticipant correlation resulting from each participant making 12 choices. We modeled treatment choice as the binary outcome with the attributes as independent variables. Each attribute's regression coefficient (importance weight) represented its relative value to participants independent of the other attributes. Higher importance weights indicated greater value; negative importance weights indicated less desirable attributes. The model also included a travel time-by-follow-up arrangement interaction term to allow travel time's importance weight to vary depending on follow-up. We subsequently evaluated for differences in importance weights by race by adding race-by-attribute interactions.

In exploratory analyses to assess the potential impact of resource limitations on willingness to travel, we compared importance weights for those who did versus did not report problems accessing transportation, did versus did not travel to their appointments by driving or being driven, and by income. To more closely examine patients most likely to have faced this decision, we compared participants who had versus had not received CAR-T/ASCT and those with versus without relapsed or refractory disease. Data were analyzed using Stata software version 16 (Stata Corp).

RESULTS

Participants

We invited 489 eligible patients and received responses from 303 (response rate 62%). One participant was excluded during the validity check for always selecting treatment X, indicating a lack of engagement with the choice tasks. This resulted in an analytic sample of 302 participants. The patient flow diagram is presented in Appendix Figure A1 (online only). Participant characteristics are presented in Table 1. 62% of participants were male, 89% White, and 6.6% Black. 25% had relapsed or refractory disease, 13% had received an ASCT, and 10% had received CAR-T. Most (62%) traveled between 30 and 90 minutes for treatment, by driving themselves (46%) or by being driven (44%), and 19% reported at least some problems with transportation to appointments.

TABLE 1.

Participant Characteristics

Sociodemographic characteristics differed by race. Driving or being driven to appointments was more common among White than Black respondents (92% v 65%), who more frequently used public transit (30% v 4% for White respondents). Black participants reported shorter travel times to their care site (1 hour or less for 100% v 60% of White participants), but more frequently had transportation problems (45% v 17%). They were also more frequently insured by Medicaid (26% v 2%), were unemployed or unable to work (50% v 14%), and had a lower median income ($20,000 in US dollars (USD) v $100,000 (USD) overall; $30,000 (USD) v $120,000 (USD) when limited to participants currently employed).

Participants' Choices Among Treatment Options

Eighty-eight participants (29%) exhibited a dominant preference for survival, 19 (6.3%) for continuity with their current oncologist, and 8 (2.7%) for a lower ICU admission rate. The proportion of dominant preferences differed by race. Ten percent of Black participants exhibited a dominant preference for survival versus 30% of White participants; for oncologist continuity, it was 25% versus 4.8%, respectively.

Participant Preferences

Importance weights for the total cohort are presented in Figure 1. When all follow-up care was at the Cancer Center where treatment was received, treatment plans were less attractive with longer travel time (Fig 1A); however, the negative impact of travel time on choice of treatment plan was mitigated when shared follow-up care was offered (Fig 1B). When patients had all local follow-up care, there was no consistent impact of travel time to receive initial treatment on choice of treatment plan (Fig 1C). Participants also valued continuity with their current oncologist (importance weight 1.15; 95% CI, 0.99 to 1.32; Fig 1D), greater overall survival (importance weight 0.71 per 5% point increase; 95% CI, 0.63 to 0.79; Fig 1E), and lower ICU admission rate (importance weight –0.21 per 5% point increase; 95% CI, –0.25 to –0.17; Fig 1F).

FIG 1.

Importance weights of travel time by follow-up arrangement: (A) when all follow-up appointments are at the Cancer Center where a patient received treatment, (B) when follow-up care is shared between the Cancer Center and a local hospital, and (C) when all follow-up appointments are at a local hospital. Importance weight of (D) oncologist continuity, (E) OS, and (F) ICU admission rate. Higher importance weights indicate greater value; negative importance weights indicate less desirable attributes. Error bars represent 95% CI of estimated importance weights. ICU, intensive care unit; OS, overall survival.

Heterogeneity in Preferences by Race

Importance weights accounting for heterogeneity by race through the incorporation of race-by-attribute interactions are presented in Figure 2. Participants who were Black were less likely to choose treatment plans requiring greater travel compared with White participants, regardless of follow-up arrangement, as indicated by lower value importance weights for longer travel time versus White patients across follow-up arrangements (Figs 2A-2C). Black patients were also less likely than White patients to choose treatment plans that offered lower continuity with their current oncologist (importance weights 2.50 [95% CI, 1.74 to 3.27] v 1.09 [95% CI, 0.92 to 1.25], respectively [Fig 2D]), and, when making choices that required trade-offs, treatment efficacy was a weaker driver of treatment plan preferences for Black than White patients (importance weights 0.34 [95% CI, 0.14 to 0.55] v 0.75 [95% CI, 0.66 to 0.84] per 5% point increase in overall survival for Black v White patients, respectively [Fig 2E]).

FIG 2.

Importance weights of travel time by follow-up arrangement for White and Black respondents: (A) when all follow-up appointments are at the Cancer Center where a patient received treatment, (B) when follow-up care is shared between the Cancer Center and a local hospital, and (C) when all follow-up appointments are at a local hospital. Importance weight of (D) oncologist continuity, (E) OS, and (F) ICU admission rate for White and Black respondents. Higher importance weights indicate greater value; negative importance weights indicate less desirable attributes. Error bars represent 95% CI of estimated importance weights. ICU, intensive care unit; OS, overall survival.

Exploratory Analyses

Importance weights were similar across travel times and follow-up arrangements for those who did versus did not report problems accessing transportation to their appointments, those who did versus did not travel to their appointments by driving or being driven, and by income. They were also similar for those who had versus had not received CAR-T/ASCT and for those with versus without relapsed or refractory disease.

DISCUSSION

In this conjoint analysis of stated preferences for cancer treatment, patients with DLBCL were more likely to choose therapies with the clinical characteristics of cellular therapy that were only offered at distant cancer centers if follow-up care could be shared locally. However, for Black patients, regardless of follow-up arrangement, travel remained an important independent negative factor in choosing a treatment only offered at a distant cancer center. Black patients also valued receipt of treatment with their current oncologist more highly than White patients did independently of the other attributes we examined. When making choices that required trade-offs, they prioritized maintaining this continuity over other treatment plan attributes. Taken together, these findings suggest that collaborative follow-up between specialized centers and a patient's local oncologist could improve access to cellular therapy. Our findings also indicate that strategies that reduce the travel burden required to receive both the initial treatment and follow-up care, while also enhancing provider continuity, may best facilitate equitable access to these treatments. Continuous investigations into multifaceted and personalized strategies that promote expanded and equitable access to cellular therapy are needed.

Our finding that patients value receiving cancer therapy closer to home is consistent with previous studies.^22,23 But our analysis also demonstrates that a patients' follow-up arrangement is an important determinant of how far they are willing to travel for treatment. This suggests that, for some patients, collaborative follow-up may facilitate access to more distant treatments. However, because the choices that Black participants made indicate less willingness to travel for cancer therapy regardless of follow-up arrangement, attention to unintended consequences is needed to avoid exacerbating existing disparities in access to cellular therapies.^17-19

One potential explanation for the differences we observed is that resource limitations such as having a lower income or poorer access to transportation⁵² represent a barrier to increased travel. However, we did not note a difference in willingness to travel by income, between those with and without transportation problems to medical appointments, or among those driving to appointments as opposed to relying on other means of transportation. Although these analyses were exploratory, this suggests that barriers other than resource limitations could be the primary mechanism for the observed preference heterogeneity.

Another potential mechanism may relate to the differential importance of provider continuity and the degree to which participants prioritized this over other factors. One potential explanation for why Black participants prioritized continuity is that longstanding^29,30 and high-quality^27,31 physician-patient relationships are needed to overcome mistrust of health care providers.^26-28 Difficulty in establishing trust with a new oncologist might have led Black patients to value their established relationship more highly than White respondents.

Our study has several limitations. First, the findings are only representative of patients who responded to our survey and nonresponse bias is a potential concern. However, our recruitment methodology included repeated, multimodality attempts to contact nonresponders and resulted in a high response rate. Second, included patients were treated within a single health system and the degree to which our findings apply to patients from other catchment areas will need to be investigated. However, because participants were recruited from a large number of sites comprising both academic and community practices across a two-state regional area, our study included patients treated in a variety of clinical settings. There is also the potential for hypothetical bias, in which participants' stated treatment selections in our study differ from their revealed preferences when faced with a similar choice during the course of treatment. To reduce hypothetical bias, we chose to use a choice-based conjoint because of its face validity for modeling clinical decisions,³⁵ incorporated a previously validated technique to reduce the use of decision heuristics⁵³ into our questionnaire, and recruited patients with DLBCL, a population that either has faced or could face this decision. Additionally, although exploratory, our analysis did not identify preference differences by relapsed or refractory status or prior receipt of CAR-T/ASCT. However, a prospective validation study to demonstrate the association of stated preferences with real-world decisions would further support our findings. Finally, although the differences observed by race may reflect structural racism-driven access inequities, the relatively small subsample of Black patients and model complexity constraints limited our ability to analyze multiple factors simultaneously. Therefore, further work is needed to determine whether these differences were due to the effects of structural racism on measured factors (eg, income) versus unmeasured factors (eg, trust) and whether the findings generalize beyond our sample.

Our findings may inform future research into programs to expand access to cellular therapies and reduce disparities in their use. The preference heterogeneity that we observed by race suggests the need for a multifaceted and individualized approach. For some patients, a model of shared follow-up care may increase willingness to travel to a more distant center to receive cellular therapy. For others, efforts to leverage longstanding oncologist-patient relationships, provide cellular therapies at a greater number of sites over a wider geographic area, or reduce the costs of travel to and from a cellular therapy center, such as through temporary housing assistance as is sometimes offered to patients undergoing transplantation,⁵⁴ may be more effective.

In summary, our findings suggest that shared follow-up care may increase patients' willingness to travel to receive cellular therapies only available at distant cancer centers but that reducing the travel burden required for both initial treatment and follow-up care while also enhancing provider continuity may best facilitate equitable access. This indicates a need to further investigate drivers of preference heterogeneity so as to inform the design and evaluation of multifaceted and personalized strategies to increase patient access to and promote equity in the receipt of cellular therapy.

Appendix

FIG A1.

Patient flow diagram. DLBCL, diffuse large B-cell lymphoma; ICD, International Classification of Diseases.

AUTHOR CONTRIBUTIONS

Conception and design: Zachary A. K. Frosch, Raghuram Iyengar, Carmen E. Guerra, Marilyn M. Schapira

Financial support: Zachary A. K. Frosch

Administrative support: Sunita D. Nasta

Provision of study materials or patients: Sunita D. Nasta, Daniel J. Landsburg

Collection and assembly of data: Zachary A. K. Frosch, Esin C. Namoglu

Data analysis and interpretation: Zachary A. K. Frosch, Nandita Mitra, Daniel J. Landsburg, Sunita D. Nasta, Justin E. Bekelman, Raghuram Iyengar, Carmen E. Guerra, Marilyn M. Schapira

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Willingness to Travel for Cellular Therapy: The Influence of Follow-up Care Location, Oncologist Continuity, and Race

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).