Background A diagnosis of childhood cancer results in new parent-child communication challenges. Little is known about how communication changes over time after diagnosis or relapse. The objective of this study was to determine the effect of time since diagnosis and relapse on quality of parent-child communication. We hypothesized that there would be a positive correlation between time and quality of parent-child communication. Methods Cross-sectional study in children (7 to 17 years) with relapsed/refractory cancer and their caregivers, who spoke English, were not cognitively impaired, and had internet access. Parents were recruited through Facebook ads. Parents and children completed the Parent-Adolescent Communication Scale (PACS), a 20-item measure of communication quality, with openness and problem subscales. Spearman’s Rho coefficients assessed correlations between PACS scores and time since diagnosis/relapse. Results There was a statistically significant negative correlation between parent PACS scores and time since child’s cancer diagnosis (Spearman’s Rho = - 0.21, p = 0.02), indicating a tendency for overall worsening communication as time since diagnosis increased. There was a positive correlation between the parent PACS problem scores and time since diagnosis (Spearman’s Rho = + 0.22, p = 0.01), indicating more problematic communication as time since diagnosis increased. Correlations of time since relapse and PACS scores were small and not statistically significant. Conclusion Parent-child communication worsens over time following a child’s cancer diagnosis with more communication problems, contrary to our hypothesis. Future studies are needed to evaluate intervention timing to best support parent-child communication beyond the new diagnosis period.