ID #1074 Site-Level Variation in Biospecimen Collection in International Multisite Pediatric Clinical Trial
Anjali Singh, Kate Swartz, Molly Clarke, Giles Robinson, Amar Gajjar, Anna VinitskyAbstract
Background
Efficient biospecimen collection is essential for pediatric clinical trials to support study objectives and improve outcomes. However, variability in site practices can compromise data completeness. We evaluated consent and collection outcomes for parental and patient blood samples, as well as patient cerebrospinal fluid (CSF) samples in a multi-site pediatric trial to identify gaps and inform improvement.
Method
Trial was opened at 22 sites (1 sponsor, 12 domestic and 9 international). Optional parental and patient blood samples and patient CSF samples were collected for exploratory analyses. Consented and collection data were extracted from the electronic data capture system. Study team interviews were conducted to identify potential barriers. Descriptive statistics were used for analysis.
Results
A total of 660 patients were enrolled over 9 years (site range 2-49; sponsor site 259) along with 614 parents for blood samples collection (351 maternal, 263 paternal). Parental blood sample consents varied widely (median 33%; range 0-78%). with 27% of sites achieving ≥50% consent and 23% obtaining no parental blood sample consent. Among consented participants, the actual parental blood sample acquisition was low (median 50%, range 0-75%). Patient blood consent was consistently high (median 98%; range 64-100%), but collection was more variable (median 82%; range 41-100%). Of 11 sites achieving 100% patient blood consent, only 5 (45%) completed full sample collection. CSF collection further demonstrated post-consent attrition: baseline compliance ranged from ∼21-100%, improved at mid-therapy timepoints (>90% at post-radiotherapy and pre-cycle 3 for many sites), and declined at end-of-therapy (most sites collected 60-75% of consented CSF samples). The sponsor site (St. Jude) consistently outperformed other sites with 99% patient blood consent (100% collected) and 96% CSF consent (92% baseline, 92% post-radiotherapy, 88% pre-cycle 3, 73% end-of-therapy), substantially influencing overall study metrics for patient’s blood and CSF samples acquisition. Key barriers included site-level operational and logistical constraints, and suboptimal sponsor monitoring.
Conclusion
Substantial site-level variability and post-consent attrition limited biospecimen completeness across specimen types. Standardized workflows, optimized logistics, and ongoing site engagement are critical to maximize biospecimen yield, feasibility, and analytical validity in future multi-site pediatric trials.