Gupta 2018.

Study characteristics
Methods	Randomised controlled trial, 2 groups, set in infertility clinic of Guru Teg Bahadur Hospital, Delhi, India December 2013 to April 2015 Number of participants randomised: 240 Number of participants analysed: 205
Participants	Inclusion criteria: women aged ≤ 35 years with ≥ 1 previous intrauterine insemination (IUI) failure and 1 of the following: (a) unexplained infertility (documented ovulation, patent tubes, and normal semen analysis); (b) minimal endometriosis with patent tubes; (c) mild male factor infertility (total motile sperm count > 10 million); (d) unilateral patent tube (IUI after confirmed ovulation on the side of the patent tube) Exclusion criteria: bilateral tubal blockage; acute pelvic inflammatory disease and/or vaginal infection; submucous myomas/endometrial polyps or anovulation in stimulated cycles Cause of infertility: unexplained infertility, mild endometriosis with patent tubes, mild male factor infertility, unilateral patent tube
Interventions	Intervention group: endometrial scratching in the cycle preceding the IUI cycle on Cycle day 20 to 22 (women with a cycle of 28 to 30 days) or on postovulatory Day 6 to 8 (women with prolonged cycles) in which ovulation was confirmed by ultrasonography Control group: no endometrial scratching Both: IUI was performed for all patients after controlled ovarian stimulation with gonadotropins (human menopausal gonadotropin (hMG)) as per standard protocol (not further described). Luteal support was provided with micronised progesterone for 15 days Degree of endometrial injury: pipelle Timing of endometrial injury: in the luteal phase of the cycle preceding the IUI cycle (between Cycle day 20 and 22 in women with a cycle duration of 28 to 30 days, and in women with "prolonged cycles", scratching was performed 6 to 8 days after ultrasonographically confirmed ovulation) Study length: 1 cycle Type of conception: IUI If the IUI cycle was cancelled, participants underwent endometrial scratching for a second time for tissue analysis (except those without a dominant follicle). These patients were considered not pregnant and were excluded from the analysis (confirmed by author correspondence)
Outcomes	Reported in the paper: Clinical pregnancy (definition provided by author correspondence: "ultrasonographic documentation cardiac activity") Ongoing pregnancy (definition provided by author correspondence: "when pregnancy had completed 20 week period of gestation") Abortion (definition provided by author correspondence: "spontaneous loss of pregnancy before 20 weeks of gestation") Ectopic pregnancy (definition provided by author correspondence: "ultrasound diagnosis of ectopic pregnancy along with beta hCG co‐relation")
Notes	Funding source: not reported Conflicts of interest: study authors declared no conflict of interest Trial registration: not found Author correspondence was undertaken
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "randomisation was done using computer generated random number table"
Allocation concealment (selection bias)	High risk	Quote from author correspondence: "the random number allocation table was provided to us by department of statistics. The person performing the randomisation could see the table. Blinding was not done" Although an adequate method of randomisation was used, as the assignment could be foreseen, there is high risk of selection bias. We noticed baseline imbalance in prognostic factors, which is a sign that allocation may not be random
Blinding of participants (performance bias)	High risk	Author correspondence confirmed that blinding was not performed; lack of participant blinding is anticipated to introduce performance bias
Blinding of personnel (performance bias)	High risk	Author correspondence confirmed that blinding was not performed
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Author correspondence confirmed that blinding was not performed; however, outcomes were unlikely to be influenced by lack of blinding (no patient‐reported outcomes)
Incomplete outcome data (attrition bias) All outcomes	Low risk	20 women in the scratch group and 15 in the control group were excluded from the analysis (respectively, 7 vs 6 due to semen sample < 0.5 mL, 6 vs 4 with unruptured follicle, 7 vs 5 husband not available on day of IUI). Missing outcome data are balanced in numbers and reasons across intervention groups; therefore risk of attrition bias is low
Selective reporting (reporting bias)	Low risk	Author correspondence confirmed that the trial was not registered. A study protocol (in Word file, made 11 November 2013, last modified 1 January 2012) was provided by the corresponding author, in which the primary outcome (pregnancy rate) was pre‐specified; therefore risk of reporting bias was rated as low
Other bias	Low risk	We did not identify any other potential sources of bias