Barton 2002.

Study characteristics
Methods	Multicenter, double‐blinded, placebo‐ and active controlled. Participants evaluated up to 24 h post‐dose or until receiving rescue analgesia. Interventions administered on postoperative day 1 after participants reported a pain intensity of ≥ 45 mm on a 0 ‐ 100 VAS and a categorical pain intensity of moderate or severe, within 6 h after discontinuation of patient‐controlled analgesia
Participants	Type of surgery: total abdominal hysterectomy (95 ‐ 100% in each group) or myomectomy (0 ‐ 5% in each group). Between 38 ‐ 41% of participants in each treatment group had severe pain at baseline, with mean VAS pain intensity ranging from 66.3 to 69.4 Ketorolac group Entered/completing: 41/41 Age (mean, SD): 40.8 (range 27 ‐ 52) Sex (male, %): 0 Placebo group Entered/completing: 42/39 Age (mean, SD): 41.0 (29 ‐ 63) Sex (male, %): 0 Parecoxib group Entered/completing: 38/38 Age (mean, SD): 42.0 (29 ‐ 65) Sex (male, %): 0 Morphine group Entered/completing: 42/42 Age (mean, SD): 40.7 (25 ‐ 61) Sex (male, %): 0 Setting: Salt Lake City, Utah, USA Dates Conducted: not specified
Interventions	Ketorolac: 30 mg single dose at report of moderate or severe pain; administration details not specified Placebo: as with ketorolac; solution and administration details not specified Parecoxib: 40 mg as with ketorolac; administration details not specified Morphine: 4 mg as with ketorolac; administration details not specified
Outcomes	Primary (as specified in study): not specified Efficacy outcomes assessed until 24 h or rescue analgesia; AEs assessed for duration of the study Time to onset of perceptible and meaningful analgesia Pain relief and TOTPAR Pain intensity difference and SPID Time to and number of participants needing rescue medication Global evaluation of intervention Adverse events by observation and indirect questioning; physical examination changes; vital signs; clinical laboratory values
Source of funding	Supported by Pharmacia Corporation, Skokie, Illinois. Author COIs not reported
Were treatment groups comparable at baseline?	Yes: demographic (age, weight) and clinical (surgical procedure, baseline pain intensity). Baseline pain intensity was ranked moderate in 59% to 62% across groups; 38% to 41% severe. Average VAS was 66 to 69 at baseline
Notes	2 strengths of parecoxib assessed: 20 mg and 40 mg. 40 mg data used here. After surgery, morphine or meperidine PCA was initiated until no later than 12:00 PM on the first postsurgical day. Any participant that had VAS ≥ 45 and a categorical pain intensity of moderate to severe within 6 hours of PCA discontinuation was randomized
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated
Allocation concealment (selection bias)	Unclear risk	Not mentioned
Blinding of participants, personnel and outcome assessors	Unclear risk	Quote: “All participants were blinded to the identity of the treatments until all study data had been collated in a database”. Comment: No mention of interventions appearing identical
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	LOCF used for missing data as a result of participants taking rescue medication or withdrawing from the study. Isolated missing data imputed by linear interpolation
Selective reporting (reporting bias)	Low risk	All outcomes described in Methods reported in full in Results. Pain relief and PID data presented graphically only
Sample size	High risk	< 50 participants in each arm