Deductible Plan with Free Chronic Disease Medications Mary Reed, - PowerPoint PPT Presentation

Value-Based Insurance Designs: Medication Adherence after Switching to a Deductible Plan with Free Chronic Disease Medications Mary Reed, DrPH Division of Research Kaiser Permanente Northern California AcademyHealth Annual Research Meeting June 2016 Kaiser Permanente Research

Study Team Mary Reed, DrPH 1 E. Margaret Warton, MPH 1 Eileen Kim, MD 2 Matthew Solomon, MD, PhD 1,2 Andrew Karter, PhD 1 1 Kaiser Permanente Division of Research 2 The Permanente Medical Group, Northern California Research funded by: Kaiser Permanente Research Health Policy and Disparities Grant No Conflicts of Interest to Declare 2

Background  US health insurance enrollees are increasingly facing health plan deductibles with high out-of-pocket costs  Higher out-of-pocket medication costs are associated with greater cost-related medication underuse  Patients with lower socio-economic status (SES) have reported more cost-based reductions in medication adherence  Combining value-based insurance designs (VBID) with a high-deductible plan can limit out-of-pocket costs for high value services 3

Research Questions Among patients switching to a plan with a deductible, how does having a VBID pharmacy benefit (free chronic disease medications) impact:  Overall patient medication adherence?  Adherence in patient groups? – Lower baseline medication adherence – Lower SES – Higher medication burden 4

Study Setting and Benefit Plans  Integrated Health Care Delivery System (IDS): Kaiser Permanente Northern California (KPNC)  Health plan deductibles apply to care until spending reaches the deductible: Medical services (office visits, ED visits, etc.) and most medications Spending Reaches Deductible Amount Full Cost Fixed Copayment VBID eligible medications: Diabetes, hypertension, lipid control, asthma, etc. Free ($0) Free ($0) 5

Study Design  Design: Difference-in-differences 2013-2014  Population:  Adult: age 18-64  Employer-sponsored insurance plan  Switch from no deductible to deductible plan (in 2014)  Taking a VBID-eligible medication at baseline 2013 Plan 2014 Study Group Switch VBID No Deductible Deductible + VBID No VBID No Deductible Deductible 6

Measures and Analysis  Exposure: Switch into a VBID pharmacy plan linked to a deductible plan (vs. no VBID)  Outcomes: Medication adherence among VBID-eligible medications – Medications for diabetes (oral medications), lipid control, or hypertension – Proportion of days covered (PDC) – percent time with adequate medication  Analysis: Random effects linear regression with interaction between time period and VBID exposure – Covariates: • Patient age, sex, race/ethnicity, language preference, complex chronic conditions, and medical center – Stratified analysis: • Baseline adherence (<80%) • Neighborhood SES • Medication burden (count, $) – Sensitivity analyses • Deductible $1000+ • Adherence outcome defined as 80%+ (logistic regression) – Percentages based on population-level estimates 7

Table 1: Baseline Patient Characteristics: VBID and Non-VBID Groups Total VBID No VBID (N=2,482) (N=1,458) (N=1,024) Characteristics p-value % % % Female 45.9 48.4 42.4 0.003 Age (yr): 18-44 18.4 16.9 20.6 0.018 45-64 81.6 83.1 79.4 Race: Black 12.6 11.5 14.2 White 44.5 43.1 46.5 Asian 19.3 22.8 14.3 <0.0001 Hispanic 15.2 13.8 17.2 Other 8.5 8.9 7.9 English Speaker 96.1 93.8 99.4 <0.0001 Lower SES 19.2 21.8 15.4 <0.0001 Chronic Conditions: Diabetes 23.8 23.7 23.9 0.91 Coronary Artery Disease 4.0 3.8 4.3 0.57 Heart Failure 1.7 1.9 1.3 0.21 Hypertension 57.5 58.0 56.9 0.61 Count of Chronic Conditions: 0 28.4 28.0 28.6 1 52.7 53.4 51.7 0.69 2 16.6 16.5 16.9 ≥ 3 2.5 2.2 2.8 Lower SES = patient lives in a census block group with ≥ 25% of adult residents earning less than a high school education or 8 ≥20% of households with annual income below the federal poverty line

Figure 1. Differential change in medication adherence after switching to deductible plan with and without VBID pharmacy plan No VBID VBID 78 76.1 76 Adherence (%) 74.3 74.1 74 73.8 VBID Differential: 2.1% (0.4,3.9) 72 70 No Deductible (2013) Deductible (2014) 9

Figure 2. Differential change in medication adherence after switching to deductible plan with and without VBID pharmacy plan by baseline adherence No VBID (Adherent) VBID (Adherent) No VBID (Non-Adherent) VBID (Non-Adherent) 100 93.1 89.7 Adherent Patients: 93.4 90 VBID Differential 89.3 -0.1% (NS) 80 Adherence (%) 70 60 Non Adherent 53.6 Patients: 50.7 50 50.3 VBID Differential 48.8 5.2% (1.8, 9.6) 40 No Deductible (2013) Deductible (2014) Non-adherent: N=1024 (41.2%) 10

Figure 3. Differential change in medication adherence after switching to deductible plan with and without VBID pharmacy plan by SES No VBID (Lower SES) VBID (Lower SES) No VBID (Higher SES) VBID (Higher SES) 78 77 76 75.4 Higher SES: 75.2 Adherence (%) VBID Differential 74.6 74 2.5% (0.6, 4.5) 72 72 71.4 70.2 70 Lower SES: VBID Differential 69.4 0.2% (NS) 68 No Deductible (2013) Deductible (2014) Lower SES N=476 (19.2%) 11

Figure 4. Differential change in medication adherence after switching to deductible plan with and without VBID pharmacy plan by medication burden No VBID (Low <5) VBID (Low <5) No VBID (High 5+) VBID (High 5+) High (5+): 84.2 84 83.6 VBID Differential 83.2 -2.9% (NS) Adherence (%) 80.2 80 76 75.2 72.8 73.2 Low (<5): 72 72 VBID Differential 2.8% (0.8, 4.8) 68 No Deductible (2013) Deductible (2014) High medication burden (5+ VBID medications): N=339 (13.7%) 12

Limitations  Single delivery system  Non-experimental allocation of benefit plans  High baseline medication adherence  Widespread low-cost generic medication use  Adherence measures based on pharmacy data 13

Conclusions  Medication adherence remained steady when switching to a deductible plan with a VBID pharmacy plan  Compared to declines in adherence after switching to a non- VBID deductible plan  Patients with lower adherence improved their adherence with a VBID pharmacy plan, even when switching to a deductible for other services  Patients with higher SES or taking fewer medications improved their adherence under the VBID pharmacy plan  No improvements in adherence for patients with lower SES or greater medication burden 14

Implications  VBID plans can offset reductions in medication adherence associated with switching to a deductible plan  VBID pharmacy plans can be particularly beneficial in patients with low medication adherence  Patients with additional clinical complexity or vulnerable populations with lower SES may need additional engagement or education about VBID provisions 15

Thanks! Mary Reed mary.e.reed@kp.org 16

17

18

Extras 19

Table 1: Baseline Patient Characteristics: VBID Exposed and Unexposed Groups (2013) Characteristics Total VBID No VBID p-value N=2,482 % N=1,458 % N=1,024 % Female 1,140 45.9 706 48.4 434 42.4 0.003 Age (yr) 18-44 457 18.4 246 16.9 211 20.6 0.018 45-64 2,025 81.6 1,212 83.1 813 79.4 Race Black 313 12.6 168 11.5 145 14.2 White 1,104 44.5 628 43.1 476 46.5 Asian 478 19.3 332 22.8 146 14.3 <0.0001 Hispanic 377 15.2 201 13.8 176 17.2 Other 210 8.5 129 8.9 81 7.9 English Speaker 2,385 96.1 1,367 93.8 1,018 99.4 <0.0001 Low SES 476 19.2 318 21.8 158 15.4 <0.0001 Chronic Conditions Diabetes 591 23.8 346 23.7 245 23.9 0.91 Coronary Artery Disease 100 4.0 56 3.8 44 4.3 0.57 Heart Failure 41 1.7 28 1.9 13 1.3 0.21 Hypertension 1,428 57.5 845 58.0 583 56.9 0.61 Count of Chronic Conditions 0 701 28.4 408 28.0 293 28.6 1 1,307 52.7 778 53.4 529 51.7 0.69 20 2 413 16.6 240 16.5 173 16.9 ≥3 61 2.5 32 2.2 29 2.8