Thin 2015.

Methods	Randomised parallel‐arm Blinding: Investigator Follow‐up: 6 months Setting: two centres, London, UK Intention‐to‐treat analysis: yes
Participants	Enrolled: 40 Age: 59 years (mean age) Gender: male (1); female (39). Duration of symptoms: not stated Inclusion criteria: Age > 18 years Meets NICE criteria (2007) for symptom severity and failure of previous conservative therapy Exclusion criteria: Inability to provide informed consent for the research study Severe concomitant medical condition precluding randomization to operative treatment Neurological diseases, such as diabetic neuropathy, multiple sclerosis and progressing Parkinson’s disease Other medical conditions precluding stimulation, such as bleeding disorders, certain cardiac pacemakers, peripheral vascular disease Congenital anorectal anomalies or absence of native rectum as a result of surgery Present evidence of external full‐thickness rectal prolapse. Previous rectal surgery (rectopexy/resection) < 12 months ago Stoma in situ Chronic bowel disease, such as inflammatory bowel disease, chronic uncontrolled diarrhoea Anatomical limitations that would prevent successful placement of electrodes Pregnancy or intention to become pregnant. Previous experience of SNS or PTNS
Interventions	SNS vs percutaenous tibial nerve stimulation (PTNS)
Outcomes	FI episodes/week (total, urge and passive); symptom severity scoring with Cleveland Clinic Incontinence Score (CCIS); quality of life measurements using EQ‐5D, FIQL and SF‐36.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomization sequence generated remotely by a statistician uninvolved in recruitment, and requests were made and actioned by e‐mail.
Allocation concealment (selection bias)	Low risk	Randomization sequence generated remotely by a statistician uninvolved in recruitment, and requests were made and actioned by e‐mail.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Investigator blinded but unclear whether patients were blinded.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Investigator blinded but unclear whether patients were blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Out of 23 participants randomised to the SNS group, 4 withdrew prior to intervention (intercurrent disease n = 1; symptoms resolved n = 1; declined further participation n = 2). Then a further 3 participants failed temporary SNS. Another participant who passed the temporary SNS withdrew from treatment owing to urgent investigation for new pararectal cyst. Moreover, in the PTNS group 1 participant withdrew before intervention (symptoms resolved).
Selective reporting (reporting bias)	Low risk	Pre‐specified outcomes in study protocol reported.
Other bias	Low risk	None identified