Table 4.

The Association of the Coverage Gap with Medication Adherence Among Aged Beneficiaries With Heart Failure and/or Diabetes

Health Conditions	Study & Comparison Groups	Unadjusted Data*		Diff-in-Diff Coverage Gap Effects†		% Change, Diff- in-Diff Effects/Pre-Gap Values‡
		Pre-Gap	Within- Gap	Estimate	95% CI	%	95% CI
Medication Possession Ratio
Heart failure	No-coverage	0.868	0.829	−0.03	(−0.04, −0.03)	−3.6	(−4.2, −2.9)
	Generic-only	0.873	0.863	−0.01	(−0.02, 0.00)	−1.1	(−1.8, −0.4)
	LIS	0.853	0.850	Ref
Diabetes	No-coverage	0.763	0.656	−0.08	(−0.09, −0.07)	−10.3	(−11.3, −9.4)
	Generic-only	0.770	0.696	−0.05	(−0.06, −0.05)	−7.1	(−8.2, −6.1)
	LIS	0.754	0.729	Ref

The sample sizes for the subsample of beneficiaries with heart failure are 11,542 in the no-coverage group, 4,578 in the generic-only group and 16,786 in the LIS group.

The sample sizes for the subsample of beneficiaries with diabetes are 12,349 in the no-coverage group, 5,308 in the generic-only group and 15,797 in the LIS group.

^*These are average numbers in the pre-gap and within-gap periods, but they are unadjusted and unweighted raw data.

^†“Diff-in-Diff Coverage Gap Effects” are adjusted estimates from the difference-in-difference model with the inverse of propensity score as a weight. The estimates measure changes in outcomes between within-gap and pre-gap periods in each study group, relative to the changes in outcomes in the comparison group (“Ref”). Propensity score model was conducted separately for each subpopulation.

^‡“% Change” is calculated using “Diff-in-Diff Coverage Gap Effects” divided by pre-gap values.

Abbreviations: LIS = low-income-subsidies; this is the comparison group. Ref = reference group.