Background: Conditional survival estimates take into account the time that a patient has remained alive following diagnosis to provide a realistic perspective on the probability of longer term survival. Such estimates are scarce for childhood cancer, particularly by age at diagnosis or stage of cancer.

Methods: De-identified population-based data were obtained from the Australian Paediatric Cancer Registry for children aged 0–14 years diagnosed with cancer between 1983 and 2010. Mortality status was followed up to the end of 2011. The hybrid period method was used to calculate relative survival estimates for those who were at risk during the period 2002–2011. Conditional survival stratified by diagnostic group or subgroup, age and stage at diagnosis was then obtained from the ratio of the relative survival estimates at different time points.

Results: A total of 13,537 children were eligible for inclusion. Five-year survival for all childhood cancers combined improved from 82% at diagnosis (95% confidence interval = 81–83%) to 89% (88–90%) conditional on surviving one year, and 97% (97–98%) conditional on surviving five years after diagnosis. Conditional survival reached 95% within five years of diagnosis for nearly all types of cancer, regardless of a child's age or stage at diagnosis.

Conclusion: Most children diagnosed with cancer who are alive five years after diagnosis can anticipate similar survival to children in the general population. This information may help alleviate some of the distress associated with childhood cancer, particularly for those with an initially poor prognosis.