Phan 1993.

Methods	2‐arm RCT undertaken in Belgium
Participants	People with stage II, III and IV or recurrent head and neck cancer selected for extensive surgery (with or without radical neck dissection and flap reconstruction) Exclusion criteria: undergoing simple laryngectomy, partial glossectomy or pharyngoplasty. In total 207 participants randomised; 102 to receive the standard gauze (86 evaluated) and 105 to the ointment group (93 evaluated).
Interventions	Group A (n = 86): standard gauze dressing (not named). Changed twice daily with wound cleaning using alcoholic chlorhexidine solution Group B (n = 93): surgical wound ointment with pure Vaseline (Qualifar) without gauze dressing. Vaseline was removed twice a day using sterile gauze, followed by cleaning of the wound with alcoholic chlorhexidine solution before application of a new cover with pure Vaseline. Duration for which the dressings remained in place was not recorded.
Outcomes	Primary review outcome: SSI (defined as a clinically documented infection localised at the surgical site and presenting with a purulent discharge with a severe inflammatory reaction > 5 cm of erythema and induration) Secondary review outcomes: not reported
Notes	Trial outcome data: seeTable 12 32 participants in each group received antibiotic treatment. Follow‐up: 20 days
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The study was prospective and randomised". Comment: limited detail provided to assess
Allocation concealment (selection bias)	Unclear risk	Quote: "Randomization was performed using sealed envelopes" Comment: not clear whether envelopes were numbered, or another method was employed to ensure concealment, so judged as unclear.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	No mention of blinding in the trial report.
Incomplete outcome data (attrition bias) All outcomes	High risk	26 participants excluded because lower GI surgery took place, surgery cancelled, protocol violation, participants could not be evaluated due to death or other circumstances.
Selective reporting (reporting bias)	Low risk	Given the information presented in the paper, all prespecified outcomes reported.
Other bias	Unclear risk	Differences at baseline: more stage IV cases in gauze group compared with Vaseline group (54% vs 39%) ‐ possibly due to exclusions?