Persistence on prostaglandin ocular hypotensive therapy: an assessment using medication possession and days covered on therapy

Table 3 Persistence and days covered estimates

Method of imputing days supply	Bimatoprost	Latanoprost	Travoprost	Total	P value*
No exclusions applied due to early failure or changes in therapy (base model)
	% or mean (SD)
Medication possession
Unadjusted days supply	23%	31%	23%	28%	< 0.001
2.0-2.1 Variable Factor^†	43%	50%	39%	47%	< 0.001
2.0 Constant Factor/45
Minimum³	43%	52%	41%	48%	< 0.001
Days covered
Unadjusted days supply	119 (80)	141 (89)	108 (76)	131 (86)	< 0.001
2.0-2.1 Variable Factor^†	218 (115)	239 (118)	197 (116)	228 (118)	< 0.001
2.0 Constant Factor/45
Minimum^‡	220 (113)	249 (115)	209 (114)	236 (116)	< 0.001
Exclusions applied due to early failure or changes in therapy (alternative model)
	% or mean (SD)
Medication possession
Unadjusted days supply	34%	44%	35%	41%	< 0.001
2.0-2.1 Variable Factor^†	60%	66%	56%	64%	< 0.001
2.0 Constant Factor/45
Minimum^‡	60%	67%	58%	65%	< 0.001
Days covered
Unadjusted days supply	183 (78)	198 (78)	167 (73)	191 (78)	< 0.001
2.0-2.1 Variable Factor^†	294 (83)	305 (79)	280 (88)	299 (82)	< 0.001
2.0 Constant Factor/45
Minimum^‡	294 (82)	310 (75)	287 (83)	304 (78)	< 0.001

*Significance from chi-square test (%) or F test (mean). Comparison is between bimatoprost, latanoprost, and travoprost.
^†Imputed days supply is unadjusted days supply multiplied by 2.0 (latanoprost, travoprost) or 2.1 (bimatoprost).
^‡Imputed days supply is unadjusted days supply multiplied by 2.0 (all prostaglandins), with a minimum imputed estimate of 45 days.
SD = standard deviation.

ISSN: 1471-2415