AbstractBackgroundTreatment‐related late effects can worsen over time among cancer survivors. Such worsening health states may trigger changes in internal standards, values, or conceptualization of quality‐of‐life (QOL). This “response‐shift” phenomenon can jeopardize the validity of QOL assessment, and misrepresent QOL comparisons over time. This study tested response‐shift effects in reporting future‐health concerns among childhood cancer survivors who experienced progression in chronic health conditions (CHCs).Methods2310 adult survivors of childhood cancer from St. Jude Lifetime Cohort Study completed a survey and clinical assessment at two or more timepoints. Based on 190 individual CHCs graded for adverse‐event severity, global CHC burden was classified as “progression” or “non‐progression”. QOL was assessed using the SF‐36TM eight domains and physical‐ and mental‐component summary scores (PCS, MCS). A single global item measured concerns about future health. Random‐effects models comparing survivors with and without progressive global CHC burden (progressors vs. non‐progressors) evaluated response‐shift effects (recalibration, reprioritization, reconceptualization) in reporting future‐health concerns.ResultsCompared with non‐progressors, progressors were more likely to de‐emphasize (or downplay) overall physical and mental health in evaluating future‐health concerns (p‐values<0.05), indicating recalibration response‐shift, and more likely to de‐emphasize physical health earlier rather than later in follow‐up (p‐value<0.05), indicating reprioritization response‐shift. There was evidence for a reconceptualization response‐shift with progressor classification associated with worse‐than‐expected future‐health concerns and physical health, and better‐than‐expected pain and role‐emotional functioning (p‐values<0.05).ConclusionWe identified three types of response‐shift phenomena in reporting concerns about future health among childhood cancer survivors. Survivorship care or research should consider response‐shift effects when interpreting changes in QOL over time.