ESTRO 2024 - Abstract Book

S2412

Clinical - Urology

ESTRO 2024

The 6 other questions in the CRF had a smaller number of events but displayed similar results as fractures and flexible sigmoidoscopy showing a mismatch between recording of events in the CHHiP CRF and HES datasets. Certain questions were represented poorly with HES data. For example, the question of whether a patient had required a long-term urinary catheter was very difficult to pick out in HES data as there is no specific OPCS or ICD code relating to this.

Conclusion:

There was a marked difference in the number of co-morbidity events recorded on CRF compared to healthcare systems data in long-term trial follow up. In general, the healthcare systems data recorded a higher number of events than the trial CRF suggesting that reliance on hospital-based trial follow-up 5 or more years after treatment has completed may result in under-reporting. In this case CRF data was collected retrospectively at the 10 year timepoint which may be part of the reason for the incompleteness of data over the time period. This analysis shows that it is possible to use HES data to identify specifically defined long term co-morbidities in large clinical trials. To improve the efficiency of collecting long-term outcomes, future trials should consider whether late side effects can be measured using specific ICD and OPCS codes since this will ensure the validity of HES data in picking up outcomes of interest.

Keywords: Data linkage, toxicity, urology

References:

Mc Cord, K.A., et al., Routinely collected data for randomized trials: promises, barriers, and implications. Trials, 2018. 19(1): p. 29

1230

Digital Poster

PSMA-PET-staged prostate cancer managed by hypofractionated radiotherapy : preliminary results