OBJECTIVES: In 2015 the National Institute for Health and Care Excellence (NICE) recommended dexamethasone intravitreal implant (DEX) for the treatment of diabetic macular oedema (DMO) where non-corticosteroid therapy is unsuitable or insufficiently responsive in pseudophakic eyes (TA349). However, DEX was not considered cost effective compared with ‘watch and wait’ (w&w) in phakic eyes. In 2019, NICE acknowledged that anti-vascular endothelial growth factors (anti-VEGFs) and laser treatments are frequently administered to insufficient responders given a lack of recommended alternatives (TA613). We assessed the cost effectiveness of DEX in phakic eyes compared with continued use of anti-VEGFs/laser despite unresponsiveness.

METHODS: The economic model previously submitted to NICE (TA349) – which assumed a comparator of w&w incurring no costs and included the cost of surgery for the proportion of phakic patients experiencing cataract (based on the ITT population of the MEAD studies, consistent with published observational studies) – was updated using comparator cost assumptions agreed in TA613: a composite comparator of 28% laser, 63% ranibizumab and 9% bevacizumab. TA613 comparator costs were applied to fellow eyes in the DEX arm and both eyes for w&w. Threshold analysis explored the impact of the confidential discount for ranibizumab. Consistent with TA613, continued use of anti-VEGF/laser was assumed to result in the same efficacy as w&w.

RESULTS: Applying TA613 comparator costs with ranibizumab at list price decreased incremental costs from £5,347 (TA349) to -£1,326, while incremental quality-adjusted life years remained consistent (0.0419). DEX is therefore dominant versus continued use of anti-VEGF/laser. DEX remains cost saving compared with continued use of anti-VEGF/laser up to a discount of 48% to the list price of ranibizumab.

CONCLUSIONS: Ensuring modelled cost estimates reflect current UK clinical practice as agreed by NICE in 2019, DEX becomes cost saving and is therefore dominant in the DMO phakic population where non-corticosteroid therapy is insufficiently responsive.