Driving distances and loss to follow‐up after hematopoietic cell transplantation

Abstract In a recent multicenter analysis, long geographic distances predicted loss to follow‐up (LTF) among allogeneic hematopoietic cell transplantation (HCT) survivors. We hypothesized that lower frequencies of patient interactions (including in‐person appointments and telemedicine encounters) would predict LTF rather than long driving distances. However, in our retrospective single‐center analysis of 263 HCT survivors, the only predictors of LTF were residence in the furthest driving‐distance quartile and Medicaid insurance (but not annualized frequencies of patient interactions). Our findings suggest that telemedicine may not necessarily "rescue" long‐distance HCT survivors from LTF. Other solutions, for example patient‐specific partnerships with local providers, may be helpful.


INTRODUCTION
Allogeneic hematopoietic cell transplantation (HCT) for hematologic malignancies requires long-term follow-up for chronic graft-versushost disease (cGVHD) management and optimal survivorship care.
However, in a recent Center for International Blood and Marrow Transplant Research (CIBMTR) analysis, loss to follow-up (LTF) rates among allogeneic HCT survivors rose to 5% within 5 years of HCT and 13% within 10 years of HCT [1]. Long geographic distances between patient residences and HCT centers predicted LTF in this analysis; however, data on appointment frequencies were not available for incorporation into CIBMTR regression modeling. Previous research suggests that long-distance HCT survivors may attend fewer in-person appointments compared to nearer-living patients [2]. Additionally, because telemedicine-based follow-up appointments are a recommended component of post-HCT care for long-distance survivors [3], it is unknown whether this observed reduction in attended in-person appointments is offset by a compensatory rise in telemedicine encounters among long-distance patients. We hypothesized that lower frequencies of patient interactions, a term including both in-person appointments and telemedicine encounters, would independently predict LTF among HCT survivors rather than long driving distances.

METHODS
We conducted a retrospective landmark analysis of adult allogeneic

RESULTS
We analyzed 263 HCT recipients as described in Overall, 17 patients (6% of our cohort) were found to be LTF. As depicted in Figure 1

DISCUSSION
Contrary to our initial hypothesis, we found long driving distance to be a superior predictor of LTF among HCT survivors compared to annualized interaction frequencies (encompassing both in-person appointments and telemedicine encounters). Compared to a recent multicenter CIBMTR analysis demonstrating a 5% rate of LTF 5 years after allogeneic HCT [1], our corresponding 6% LTF rate with median follow-up of 49 months was similar. Additionally, our findings were consistent with the CIBMTR's models, suggesting that driving distances of 100+ miles and public insurance status were associated with LTF. While we defined long driving distances based on the 75th percentile rather than a predefined cutoff of 100+ miles as used by the CIBMTR, these definitions were analogous given that 100 miles constituted the 71st percentile of driving distance in the CIBMTR analysis [1]. As such, our research builds upon this previous analysis by demonstrating that long driving distances and Medicaid insurance retain their association with higher rates of LTF even after accounting for annualized interaction frequencies and cGVHD (data about which were not available to the CIBMTR).
Limitations of our retrospective cross-sectional analysis include its relatively small sample size from a single institution, because of which we were only able to identify 17 LTF patients. Given that 7% of patients in our cohort had originally undergone HCT at a different center (similar to the proportion at another academic US center) [4], pandemic.
Our findings nevertheless highlight a potential limitation of previous studies that have investigated driving distances and outcomes in malignant hematology. These studies have generally found no impact of long distances on either clinical or patient-reported outcomes [2,4-7. If certain long-distance patients are disproportionately both at risk of LTF and worsened outcomes, retrospective analyses with censoring may fail to capture the entirety of their post-HCT courses. From a clinical perspective, our data also suggest that telemedicine itself may not necessarily "rescue" long-distance HCT survivors from higher LTF rates. Other solutions to coordinate care, for example proactive patient-specific partnerships with local providers using direct email/phone communication channels [3,8, [11], important follow-up information after telemedicine encounters can potentially be mailed to patients at potential risk of LTF as well.

CONCLUSION
We found that long driving distances among HCT survivors are associated with higher rates of LTF even after adjusting for how frequently these patients interact in person or virtually with their HCT center.
Mitigating the risk of LTF among long-distance survivors may require a multipronged approach with both telemedicine adoption as well as other personalized strategies to maintain engagement among these vulnerable patients.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.