1 Department of Psychology and Biobehavioral Sciences, St. Jude Children's Research Hospital, Memphis, TN, USA.
2 Comprehensive Cancer Center, St. Jude Children's Research Hospital, Memphis, TN, USA.
3 Department of Biostatistics, St. Jude Children's Research Hospital, Memphis, TN, USA.
4 Phastar Inc., Cambridge, MA, USA.
5 Division of Behavioral Medicine and Clinical Psychology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH, USA.
6 Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH, USA.
7 Department of Epidemiology and Cancer Control, St. Jude Children's Research Hospital, Memphis, TN, USA.
8 Department of Psychology, Concordia University, Montreal, QC, Canada.
9 Virginia Commonwealth University Massey Comprehensive Cancer Center, Richmond, VA, USA.
10 University of Minnesota Medical School, Minneapolis, MI, USA.
11 The University of Texas MD Anderson Cancer Center, Houston, TX, USA.

Background: Childhood cancer survivors face physical, psychological, and neurological symptoms that contribute to risky health behaviors and increased healthcare utilization. Traditional survivorship care models overlook risk associated with this symptom burden. The current study examined symptoms phenotypes to identify high-risk groups.

Methods: Five-year survivors (N = 17,231; Mean [standard deviation] age = 27.4 [5.98]; 80% non-Hispanic White; 48% female) from the Childhood Cancer Survivor Study (NCT01120353) self-reported symptoms and risky behavior at baseline and first follow-up (original cohort data collection: baseline 1994-1998 and follow-up 2002-2004; expansion cohort: baseline 2008-2010 and follow-up 2014-2016). Medical records were extracted through chart review. Chronic health conditions (CHCs) were graded according to common terminology criteria for adverse events criteria. Latent class analysis derived symptom phenotypes.

Findings: Five phenotypes emerged: 1) Low Burden (63.1%); 2) Cardio-Pulmonary-Pain (5.3%) 3); Neurologic-Pain (10.6%); 4) Psychological Distress-Pain (13.3%); 5) Global burden (7.7%). Compared to survivors with Low Burden, those in other symptom phenotypes were older, female, had lower education, no health insurance, smoked cigarettes, were physically inactive, and had = grade 3 CHC (all ps < 0.05). Survivors in symptom phenotypes were at-risk for future emergency room use (all ps < 0.05). Risk for future physical inactivity was higher in Cardio-Pulmonary-Pain (OR = 1.19, CI = 1.09, 1.31), Global (OR = 1.12, CI = 1.02, 1.22), and Neurologic-Pain (OR = 1.18, CI = 1.10, 1.27) phenotypes. Cigarette use was higher in Cardio-Pulmonary-Pain (OR = 1.62, CI = 1.08, 2.42) and (Global OR = 1.65, CI = 1.17, 2.31) phenotypes.

Interpretation: Symptom phenotyping identified groups at-risk for future risky health behaviors, which was not explained alone by diagnosis or CHCs. Integrating symptom assessments may guide interventions to improve health outcomes.

Funding: The work was supported by the National Cancer Institute (U24 CA055727, PI: GT Armstrong). Support to St. Jude Children's Research Hospital was also provided by the National Cancer Institute Cancer Center Support grant (P30 CA021765, PI: CWM Roberts) and by the American Lebanese Syrian Associated Charities.