Strooker 1998

Methods	Randomised, split‐mouth study. Patients cannot be blind, outcome assessor not blind. No withdrawals.
Participants	Adults. Four mandibular implants splinted with a bar supporting an overdenture. 16 enrolled and results given for 16.
Interventions	2 groups. Monthly 35% phosphoric etching gel (pH1) for 1 minute versus supra‐ and sub‐gingival mechanical debridement using carbon fibre curettes and rubber cup. Maintenance every month. Study duration: 5 months.
Outcomes	Plaque index by Silness and Loe, calculus index by Bjorby and Loe, a modification of the gingival index by Loe and Silness, change in probing pocket depth (mm), microbiological sampling, postoperative pain and treatment time at 1 and 5 months. 5‐month data used.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Test and control therapy was randomly assigned to left and right sides of the mandible". Written to the author: "We used a cobblestone, even numbers was the right test side of the patient, uneven the left".
Allocation concealment (selection bias)	Unclear risk	"Test and control therapy was randomly assigned to left and right sides of the mandible".
Blinding (performance bias and detection bias) All outcomes	High risk	"Test and control site were then revealed to the same examiner, who performed the assigned treatments and follow‐up measurements throughout the study".
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No withdrawals. No reply from author.
Selective reporting (reporting bias)	Unclear risk	No reply from author.
Other bias	Unclear risk	No reply from author.
Blinding of the patient?	Low risk	Not possible. This was judged to be irrelevant for the risk assessment.