Bloomfield 1987.

Study characteristics
Methods	Randomised controlled trial Sample size calculation not stated.
Participants	Setting: single‐centre study at Cincinnati General Hospital ‐ time frame not given Inclusion criteria: hospitalised women with moderate of severe uterine cramp pain within 48 hours of an uncomplicated vaginal birth Exclusion criteria: women were excluded with known hypersensitivity to aspirin or NSAIDs, if they had been given other analgesia or were breastfeeding their babies
Interventions	Following initial pain assessment women were randomly allocated to 1 of 2 treatment groups and given appropriate study medication on demand. Randomisation was stratified by initial pain intensity and by 1 of 3 nurse observers Naproxen sodium 550 mg (2 capsules 275 mg naproxen sodium) (N = 30) Placebo (2 capsules placebo) (N = 30)
Outcomes	Adequate pain relief as assessed by the woman: women were interviewed by 1 of 3 trained nurse observers before drug administration and ½ hour post‐treatment and then hourly for 6 hours Pain intensity was measured on a 4‐point ordinal scale and pain relief on a 5‐point ordinal scale (not described). Women were asked to give a global rating of the medication at the final interview on a scale of 0 (worst) to 10 (best pain reliever ever taken). Pain intensity difference, SPID and mean total pain relief scores. Reported SPID scores were used to calculate 'adequate pain relief as assessed by the woman' (estimated over 6 hours) Need for additional analgesia: women requiring greater pain relief were removed from the study and given medication as needed, they were not interviewed further. Data for these women were included in the analysis. Maternal side effects: women were questioned about side effects at each interview without a checklist or leading questions.
Notes	Additional study arms: this study included an additional 2 arms of anirolac 50 mg and 100 mg. This medication is no longer available, so these arms were not included Dates of study: not stated. Funding sources: Sytnex Research ‐ manufacturer of anirolac Declarations of interest: none stated
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Women were assigned using a "predetermined balanced randomized schedule". "On enrolment, patients underwent two‐way stratification: first according to clinical nurse‐observer (morning or afternoon shift) and second, according to initial pain intensity (moderate or severe). Within each of these 6 strata, patients were allocated to 1 of 6 treatment groups according to a predetermined, balanced, randomization schedule that assured that all groups were of equal size and matched with respect to initial intensity of pain and nurse‐observers." (Only 2 of the 6 strata have been included in this meta‐analysis).  Unclear exactly how random sequence was generated.
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Study medications were "packaged in code‐numbered individual dose containers". "All capsules identical in taste and appearance."
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated
Incomplete outcome data (attrition bias) All outcomes	Low risk	Data on all women randomised have been reported. There were 30 women allocated to each of the groups. One women in the placebo group was disqualified as she was inadvertently given a medication contraindicated with study medications. One women in the naproxen group had received analgesia prior to enrolling in the study and was also disqualified. These women were replaced
Selective reporting (reporting bias)	Unclear risk	No protocol published or trial registration available
Other bias	Low risk	No other risk of bias identified