Abstract 1565P
Background
In the era of precision medicine, precision symptom management is also important to advance. Symptom interventions delivered by a combination of electronic and clinician interaction allows for differential delivery of cancer symptom treatment elements. Both prognostic and predictive relationships could form the basis of tailored symptom care.
Methods
387 cancer patients receiving chemotherapy were randomized to four different symptom treatment approaches using a combination of automated self-management coaching, nurse practitioner-delivered care and decision support algorithms. A total symptom burden score, comprised of 11 common physical and psychosocial chemotherapy-related symptoms was assessed daily for six months. Prognostic and predictive relationships were evaluated between this daily symptom score and pre-randomization covariates derived from PROMIS-29 assessments of anxiety, depression, fatigue, pain, physical function, social function, and sleep disturbance. A prognostic analysis examined the symptom burden trajectory as a function of the pre-randomization covariate battery, while a predictive analysis investigated how burden trajectories for each of the 4 treatment interventions varied as a function of covariate levels. Both analyses used mixed effects linear models with quadratic time effects and random intercepts estimated under maximum likelihood.
Results
Participants were on average 59 years old, female (57.6%) with a variety of cancer diagnoses and stage 3 or 4 disease (72%). Demographic and disease covariates were not prognostic of symptom burden; however, significant (p<.001) prognostic relationships emerged between subsequent symptom burden trajectories and anxiety, social function, and sleep disturbance. Anxiety, depression, pain, physical function, social function, sleep disturbance (all p<=.001) and fatigue (p=.009) predicted differential intervention change trajectories in the predictive analysis.
Conclusions
Pre-treatment psychosocial covariates predict subsequent symptom burden and can identify in advance those most likely to benefit from individual intervention components that reduce symptom burden.
Clinical trial identification
NCT02779725.
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
U.S. National Cancer Institute.
Disclosure
K. Mooney: Financial Interests, Personal, Advisory Board: Reimagine Care. All other authors have declared no conflicts of interest.