Introduction: Though ideal models of survivorship care are not definitively established, it has been suggested that childhood acute lymphoblastic leukemia (ALL) survivors can be cared for by properly informed primary care physicians (PCP - e.g. family physicians, community pediatricians) given low risks of late effects. PCP-driven models of care are dependent on the willingness of families to re-engage with their PCPs after a prolonged period of treatment delivered by pediatric oncologists during which PCP involvement may be minimal. We thus aimed to identify rates and predictors of PCP visits both during and after treatment among a population-based cohort of children with ALL.
Methods: We identified all children <18 years at diagnosis of ALL at a pediatric center in Ontario, Canada between 2002-2012. Patients were linked to healthcare data and matched to population-controls by age, sex, and geography (1:5 ratio). PCP physicians at the time of diagnosis were identified through validated algorithms using primary care billing codes and patient rosters. PCP visit rates during treatment were determined and compared between cases and controls, with cases censored at the time of relapse, stem cell transplant, or death. Post-treatment PCP visit rates were calculated among those completing frontline therapy until relapse, death, or the end of the study period. Predictors included demographic- (e.g. age, sex, socioeconomic status, distance from treatment center), disease-related (e.g. risk status, lineage), and PCP-related variables (e.g. pediatrician vs. non-pediatrician).
Results: Of 801 children with ALL, 751 (93.8%) had an identified PCP at the time of diagnosis. Excluding a further 8 (1.0%) patients who did not have treatment information available resulted in 743 cases and 3,112 controls. The median age of children with ALL was 4.0 years [interquartile range (IQR) 3.0-8.0], 409 (55%) of whom were male. Nearly half of patients (361, 48.6%) did not visit their PCP during treatment. The rate of PCP visits during treatment was 0.64 per person per year (PPPY) compared to 1.4 PPPY among controls. Adjusting for age, sex, and socioeconomic status resulted in an adjusted rate ratio (aRR) of 0.47 [95th confidence interval (95CI) 0.40-0.54; p<0.0001]. In multivariable analyses, no disease-related factors were associated with PCP visit rates. Infants had lower PCP visit rates during treatment (RR 0.09 vs. age 1-4 years, 95CI 0.01-0.6, p=0.01) while patients living at greatest distance from their treatment centre had higher rates (RR 1.6, 95CI 1.1-2.3, p=0.01). PCP type (pediatrician vs. other) did not have an impact on visit rates during treatment.
Excluding 32 (4.3%) patients who relapsed, died, or underwent stem cell transplant prior to completing frontline therapy yielded 711 cases (survivors) and 2,973 controls available for analyses of post-treatment PCP visits. The median time of follow up after the end of treatment among survivors was 6.0 years (IQR 4.0-8.5). Though 287 (40.4%) of survivors did not visit their PCP during the post-treatment period, survivors overall still experienced greater post-treatment PCP visit rates compared to controls (aRR 1.4, 95CI 1.2-1.6; p<0.0001). This was true throughout the post treatment period, with the greatest increase in visit rates compared to controls seen 10 years from the end of treatment and beyond (aRR 3.6, 95CI 1.7-7.5, p=0.0007). In multivariable analyses, survivors who had seen their PCP during active treatment had post-treatment visit rates twice as high as those who had not (aRR 2.0, 95CI 1.6-2.5; p<0.0001). Survivors with pediatricians as PCPs also had higher post-treatment visit rates compared to survivors who did not (aRR 1.4, 95CI 1.1-1.8; p=0.003).
Conclusions: Only a portion of children with ALL see their PCPs during treatment and return to PCP care following the completion of leukemia treatment, indicating that PCP-led survivorship care is feasible only for a subset of this population. The rate of PCP visits among survivors however continues to increase relative to general-population controls beyond 10 years after end of treatment. Post-treatment engagement with PCPs may be improved by PCP involvement during treatment, as well as in some cases the involvement of community pediatricians.
Gupta: Jazz Pharmaceuticals: Consultancy, Membership on an entity's Board of Directors or advisory committees.