Smith 2005.

Methods	Participants were allocated using computer-generated randomisation codes prepared in advance. Randomisation and the ordering of spectacles were performed by a principal investigator who had no contact with participants during the study. Masking: Participant: yes Provider: no; Outcome: yes. Exclusions after randomisation: None reported. Losses to follow-up: 10 in the custom prism group, 6 in the standard prism group, and 2 in the placebo group
Participants	Country: UK Number randomised: 243 individuals with AMD at Manchester Royal Eye Hospital,England. Age: mean 81 years Sex: women 65% Inclusion criteria: English-speaking. Exclusion criteria: illiterate, resident in a hospital or a nursing home. Study duration: 3 months follow-up during the period July 15, 2001, through March31, 2003
Interventions	Treatment: 3 types of test spectacles: custom, incorporating bilateral prisms to match participants' preferred power and base direction; standard, incorporating standard bilateral prisms (6 prism diopters base in for logMAR VA of 0.48 - 1.00 and 10 prism diopters base in for logMAR VA of 1.02 - 1.68); placebo, consisting of spectacles matched in weight and thickness to prism spectacles but without the prism. The spectacles prescribed to each group included the optimal refractive correction for distance and near vision
Outcomes	logMAR visual acuity with ETDRS chart. Reading speed and critical print size with an MNREAD chart. National Eye Institute Visual Functioning Questionnaire (NEI-VFQ-25). Melbourne Low-Vision ADL (Activities of Daily Living) Index (MLVAI), part 1 consisting of the performance of 16 typical ADL dependent on vision assessed for speed, accuracy, and independence of performance, and part 2 consisting of a questionnaire. Manchester Low VisionQuestionnaire23 (MLVQ) with itemsmeasuring helpfulness and use of test spectacles
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were allocated to a group using computer-generated randomisation codes prepared in advance
Allocation concealment (selection bias)	Low risk	Randomisation and the ordering of spectacles were performed by a principal investigator who had no contact with participants during the study
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low number of losses to follow-up: 10 in the custom prism group, 6 in the standard prism group, and 2 in the placebo group
Selective reporting (reporting bias)	Low risk	No protocol available, but primary outcome reported
Other bias	Low risk	No other bias identified