Long-Term Follow-up or Lost to Follow-up? Driving Distance and Continuity of Follow-up Care after Allogeneic Transplantation

Abstract

Introduction Long-term follow-up (LTFU) clinics overcome any potential survival disadvantage from long driving distances in hematopoietic stem cell transplant (HCT) recipients (Ragon et al, 2014; Khera et al, 2017). However, omitting pts who are lost to follow-up (f/u) – defined as no phone/electronic communication or appointment requests in 12+ mos – may introduce selection bias. We hypothesized that long driving distances lead to higher ‘lost' rates in HCT survivors. Methods We conducted a landmark analysis of HCT survivors > 12 mos after allogeneic HCT seen at least once in our clinic between 1/1/02 & 6/30/18. We identified each pt's last point of contact with our institution (including admissions & death notifications). For pts not known to have died, we calculated ‘no-contact' intervals since last points of contact. No-contact intervals of 12+ mos constituted being lost to f/u. We used the 75th percentile of driving distance (as done by Ragon et al) to delineate closer-living vs. furthest-living groups, calculating percentiles within f/u cohorts (1-4.9 yrs, 5-9.9 yrs, or 10+ yrs after HCT) to account for catchment area changes over time. Results We analyzed 437 pts, of whom 22 (5%) had received HCTs at other centers. Compared to the closer-living group (median distance 24.1 mi, interquartile range [IQR] 13.9-33.3), the furthest-living group (median distance 78.4 mi, IQR 66.2-117.0) was more likely to be white but was otherwise comparable by age, gender, insurance, HCT indication, & GVHD distribution. The two groups had similar overall survival (median 32.0 vs. 31.0 mos, p = 0.94). Among patients not known to have died (n = 272), lost to f/u rates were almost 2x as high (17% vs. 9%) in the furthest-living group; however, this difference did not reach significance (p = 0.13). Only 1/4 of lost pts (n = 30) had documented reasons for ceasing contact, e.g. preference for local care or transfer to another center. The furthest-living group had significantly longer no-contact intervals than the closer-living group (p Discussion Our sample size was relatively small, and we were unable to assess HCT recipients not seen in our ambulatory LTFU clinic. Nevertheless, our single-center study suggests that long-distance pts may disproportionately become lost to f/u. LTFU clinics should ensure adequate inclusion, e.g. via telehealth, of long-distance HCT survivors in clinical care and research.