Table 2

Key model inputs and assumptions

Time period	Model input	Combination therapy	Ranibizumab monotherapy	Laser monotherapy	Source
Year 1	BCVA progression	RESTORE trial data, adjusted for drop-out rates	RESTORE trial data, adjusted for drop-out rates	RESTORE trial data, adjusted for drop-out rates	RESTORE data on file
	Treatment frequency	7 injections + 2 laser sessions, as in RESTORE; drop-outs continue in standard care (ie, laser therapy)	7 injections; drop-outs continue in standard care (ie, laser therapy)	2 laser sessions; drop-outs continue in standard care (ie, laser therapy)	RESTORE data on file
	Monitoring visits	12	12	4	SmPC and expert interview (data on file)
	Adverse events	Negligible	Negligible	Negligible	RESTORE data on file
Year 2	BCVA progression	Equal rates of improvement and worsening (3% in 3 months)	Equal rates of improvement and worsening (3% in 3 months)	Equal rates of improvement and worsening (3% in 3 months)	Supported by DRCR.net and RESTORE data on file
	Treatment frequency	2 injections + 1 laser session	3 injections (no laser)	1 laser session	Supported by DRCR.net
	Monitoring visits	8	10	4	Assumption
Year 3	BCVA progression	Constant rates of change of BCVA with a majority of patients having a decline in BCVA	Constant rates of change of BCVA with a majority of patients having a decline in BCVA	Constant rates of change of BCVA with a majority of patients having a decline in BCVA	Calibrated to WESDR data (Supplementary Methods)
	Treatment frequency	No additional ranibizumab; laser therapy as required	No additional ranibizumab; laser therapy as required	Laser therapy as required	Assumption
	Monitoring visits	4	4	4	Assumption
Any year	Cost of blindness	When BCVA ≤35 letters is reached in better-seeing eye	When BCVA ≤35 letters is reached in better-seeing eye	When BCVA ≤35 letters is reached in better-seeing eye	Adapted from costing approach by Meads et al16 (Supplementary table 2)

Patient-level changes in BCVA data in RESTORE were used to derive transition probabilities in year 1.
After year 1, long-term changes in BCVA were simulated assuming categorisation into one of three possible outcomes: ≥10 letters improvement within 3 months (one health state up), ≥10 letters worsening within 3 months (one health state down) or no change exceeding 10 letters within 3 months.
BCVA, best corrected visual acuity; DRCR, Diabetic Retinopathy Clinical Research Network; SmPC, summary of product characteristics; WESDR, Wisconsin Epidemiologic Study of Diabetic Retinopathy.