Obtaining uncontaminated urine samples from infants, children and young people for suspected urinary tract infection (UTI) is clinically challenging. Non‑invasive clean catch urine (CCU) collection can be slow and may yield contaminated samples, while invasive methods such as transurethral bladder catheterisation (TUBC) and suprapubic aspiration (SPA) are more invasive but potentially more accurate. The FROG study assessed whether randomisation between these approaches is feasible to inform a definitive trial.
FROG was a multicentre, randomised feasibility trial incorporating a mixed‑methods perspectives study, a health‑economic analysis and a stakeholder consensus meeting. Eligible participants for the feasibility trial were children under 16 years who required urine testing for suspected UTI and could not provide a midstream sample. Parents, children and healthcare professionals were eligible to participate in the perspectives study and the consensus meeting.
Recruitment, consent and allocation procedures were assessed alongside practical barriers to receiving the allocated sampling technique. The perspectives component used parent interviews, parent questionnaires and staff focus groups/interviews to capture acceptability and decision factors. A stakeholder consensus meeting was convened to determine age ranges, comparator arms and priority outcomes for a potential definitive trial.
Screening and recruitment
Of 703 children screened across participating centres, 170 were offered the study and 99 were recruited. Overall, 64/170 (37.6%) consented to randomisation, which exceeded the prespecified feasibility threshold of 33%.
Randomisation and receipt of allocated method
Among those consenting to randomisation, 32 participants were allocated to CCU and 32 to TUBC. The allocated urine collection method was actually received by 46/64 (71.9%). Reasons for not receiving the allocated method included logistical delays, unsuccessful attempts at collection and distress in the child.
Microbiology and contamination
For participants with culture results available, contamination was reported in 2/12 (16.7%) allocated to CCU and in 0/6 allocated to TUBC. The trial report notes that no participants consented to randomisation involving SPA, so no data were generated for that method in the randomised comparison.
The perspectives study enrolled 14 parent interviews, 89 parent questionnaires and 28 staff participants across five focus groups and one staff interview. Both parents and staff regarded CCU and TUBC as acceptable options, but acceptance involved trade‑offs between the speed and accuracy of collection and the potential for pain and distress in the child. Availability and acceptability of SPA were limited among stakeholders and potential participants, which contributed to the absence of consent for SPA randomisation.
Nineteen stakeholders attended the consensus meeting. There was strong support for focusing a definitive trial on younger children: 94% supported recruiting children aged under 18 months. All attendees (100%) supported comparing CCU with TUBC and excluding SPA from the randomised comparison. Participants ranked diagnostic accuracy as the highest‑priority outcome for the definitive trial.
The FROG feasibility study indicates that a definitive randomised trial comparing a CCU‑first versus TUBC‑first approach in children aged under 18 months is feasible. The authors recommend the primary outcomes of such a trial should reflect diagnostic accuracy and the clinical consequences of contamination. Key secondary outcomes should include successful sample collection, time to collection, and measures of pain and distress. The trial encountered practical barriers that affected receipt of the allocated method, which would need addressing in trial design and delivery.
The study reports ethical approval from North East, Newcastle & North Tyneside 1 Research Ethics Committee (Ref: 24/NE/0222). The trial is registered with ISRCTN (ISRCTN84676764). Participant‑level datasets have not been deposited publicly because of re‑identification risk; de‑identified data may be made available on reasonable request in line with consent, ethical and sponsor approvals, and a suitable data‑sharing agreement.