Gill 2007.

Methods	2‐arm parallel randomised controlled trial
Participants	Numbers recruited: 60 (authors reported gender of participants who completed the study: 32 females and 25 males) Mean (± SD) age: Group 1: 13.72 ± 2.5 years, Group 2: 13.37 ± 1.42 years Inclusion criteria: people treated with upper and lower pre‐adjusted edgewise appliances (both extraction and non‐extraction cases) Exclusion criteria: hypodontia where a Hawley‐type retainer was required for space maintenance cases where rapid maxillary expansion was used or surgically assisted expansion non‐orthognathic cases where significant movement of the lower incisors required a bonded retainer people with reduced periodontal support Setting: hospital department in the UK, participants treated free of charge
Interventions	Comparison: part‐time wear of upper and lower thermoplastic (vacuum‐formed) retainer vs. upper and lower full‐time wear of thermoplastic (vacuum‐formed) retainer Group 1: part‐time wear of vacuum‐formed retainers (sleeping hours with minimum of 8 hours of wear) Group 2: full‐time wear of vacuum‐formed retainers The vacuum‐formed retainers were full‐arch coverage and made from 1 mm thick polycarbonate sheets
Outcomes	Outcome relevant to review: stability Stability assessed using Little's Irregularity Index, intercanine width and intermolar width in both arches, and overjet and overbite. PAR index reported in a graph, but data not sufficiently clear for further analysis Outcomes were measured after 6 months
Notes	Authors reported the type of initial malocclusion, whether the participants were treated by extraction or non‐extraction and the time in treatment. Despite randomisation, there were more Class II participants in the part‐time group
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The authors confirmed the randomisation was generated by rolling a die and allocating the group according to odd or even numbers
Allocation concealment (selection bias)	Low risk	"Identical opaque sealed envelopes containing a slip for group allocations were used for the purpose of random allocation"
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessor was blind to the treatment group
Incomplete outcome data (attrition bias) All outcomes	Low risk	Loss of follow‐up data were reported in Consort diagram. Effect was not analysed, but as drop‐out was low, it is unlikely to have a significant effect
Selective reporting (reporting bias)	Low risk	Authors confirmed they did not collect any other outcomes and there was no selective reporting
Other bias	Low risk	No areas of risk of bias were identified