Kandavel 2018.

Study characteristics
Methods	Randomised controlled trial, 2 groups, setting not described (study group is from the United Kingdom) November 2015 to September 2017 (as described in trial registry) Number of participants randomised: unknown Number of participants analysed: 109
Participants	Inclusion criteria: women aged 18 to 42 years with recurrent miscarriage; written informed consent; actively trying to get pregnant (as described in trial registry) Exclusion criteria: "no active treatment in pregnancy"; inherited or acquired thrombophilia; medical conditions (diabetes, hypertension, thyroid disorders); inability to tolerate internal examinations; uterine anomalies; previous entry or randomisation in the present trial (as described in trial registry)
Interventions	Intervention group: endometrial scratch in the luteal phase Control group: a sham procedure in the luteal phase. The sham procedure consists of cleaning the cervix with saline using a cotton tip (as described in trial registry) Degree of endometrial injury: Wallace catheter (as described in trial registry) Timing of endometrial injury: in the luteal phase Study length: not described Type of conception: not described
Outcomes	Reported in the abstract: Pain graded as mild/moderate/severe Bleeding Author correspondence confirmed that pregnancy outcomes were also recorded, but we were not able to obtain these data, as we did not receive a response to follow‐up emails
Notes	Only a conference abstract was available Funding source: not reported Conflicts of interest: not reported Trial registration number: NCT02681627 (retrospectively registered) Author correspondence was undertaken, as pregnancy outcomes were not reported in the abstract. Author correspondence confirmed that pregnancy outcomes were recorded, but the corresponding author did not specify which pregnancy outcomes and did not respond to follow‐up emails
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	In the abstract, study authors stated that women were randomised but did not describe the randomisation method
Allocation concealment (selection bias)	Unclear risk	Abstract did not report whether allocation concealment was performed
Blinding of participants (performance bias)	Unclear risk	Abstract did not report blinding of participants. A sham procedure was performed, but it is unclear whether participants were successfully blinded by the sham procedure
Blinding of personnel (performance bias)	Unclear risk	Abstract did not report any blinding of personnel
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Abstract did not report blinding of outcome assessors. The study recorded patient‐reported outcomes (pain and bleeding after the procedure), which could be influenced if participants were not blinded. It is unclear whether participants were successfully blinded by the sham procedure
Incomplete outcome data (attrition bias) All outcomes	High risk	Outcomes were recorded by a questionnaire, which was filled in by 68 out of 109 randomised women (response rate 62.38%), of which 33 women were in the intervention group and 35 in the control group. It is unknown how many women were in the intervention and control groups, as only the total number of randomised women was given. Non‐responders were not included in the analysis. The large proportion of missing data is likely to introduce attrition bias
Selective reporting (reporting bias)	Unclear risk	Trial was registered retrospectively
Other bias	Unclear risk	Information is insufficient to show whether an important risk of bias exists