Skip to main content

Table 3 Regression-adjusted estimates of expenditures, prescription drug use, and adherence to medications stratified by type of medicare part D coverage (n = 344,817)

From: Patterns of prescription drug expenditures and medication adherence among medicare part D beneficiaries with and without the low-income supplement

 

LIS N = 25,226

Non-LIS/GC N = 29,289

Non-LIS/Non-GC N = 290,302

Expenditures

  Total expenditures

$1,887a,b

$1,360a

$1,341

(1,864-1,910)

(1,344-1,375)

(1,336-1,347)

  Out-of-pocket expenditures

$148a,b

$546a

$570

(146–150)

(539–552)

(567–572)

  Plan expendituresc

$1,708a,b

$822a

$776

(1,687-1,729)

(811–833)

(772–780)

Expenditures on brand name medications

   

  Total expenditures

$1,325a,b

$926a

$898

 

(1,305-1,346)

(911–941)

(893–903)

  Out-of-pocket expenditures

$96a,b

$374a

$369

(94–98)

(368–381)

(367–371)

  Plan expendituresc

$1,221a,b

$560a

$537

(1,202-1,240)

(549–571)

(533–540)

Expenditures exceeding the gap threshold

27.6%a,b

17.4%a

16.9%

(27.2-28.1)

(17.0-17.8)

(16.8-17.0)

Prescriptions

   

  Total number of prescriptions

38.1a,b

25.1a

26.5

(37.8-38.4)

(24.9-25.3)

(26.4-26.5)

  Total number of brand name prescriptions

10.7a,b

6.4a

6.9

(10.5-10.8)

(6.3-6.5)

(6.87-6.92)

Adherence to medications

   

  Diabetes drug adherence

62.5%a,b

57.7%

57.4%

(61.0-63.9)

(56.0-59.4)

(56.9-58.0)

  Hypertension drug adherence

65.6%a,b

64.2%a

62.4%

(64.9-66.3)

(63.5-64.9)

(62.2-62.6)

  Lipid-lowering drug adherence

59.6%a,b

57.0%a

55.6%

(58.5-60.6)

(56.0-58.0)

(55.3-55.9)

  1. LIS: low-income subsidy beneficiaries; GC: gap coverage; PDC, proportion of days covered, which was calculated from dispensing data within each class of drugs; Adherence: defined as having a PDC ≥80% in the year or month for the entire regimen, allowing drug supply to carry over from month to month.
  2. Each model of medication adherence included only beneficiaries with the condition of interest. Estimates were adjusted for the individual beneficiary characteristics and clinical co-morbidities listed in Table 1 and for the residential census characteristics listed in Table 2.
  3. aSignificantly different from the Non-LIS/Non-GC group (p <0.05).
  4. bSignificantly different from the Non-LIS/GC group (p <0.05).
  5. cMeasures of plan cost for LIS beneficiaries is comprised of the cost to the plan plus subsidies.