Medication Adherence, Hypertension
Conditions
Brief summary
Medicare Advantage beneficiaries ages 65-79 and insured by Humana with at least two unique fills of hypertension medication within the 2023 calendar year and adherence level between 60 and 85% will be identified using Humana Medicare Advantage claims data. Individuals meeting these inclusion criteria will be included and, with an institutional review board approved waiver of informed consent, will be randomized to one of 6 mailed messages or control (no message). The messages will be sent by Humana and use different behavioral techniques (social norms, messenger effects, and/or processing fluency) providing their medication refill score. Humana will send a second message within 60 days of the first message noting any changes in the refill score. The primary outcome will be the average end-of-year adherence in each arm. A secondary outcome will be the proportion of study participants with end-of-year adherence greater than or equal to 80%. The study team's hypothesis is that messages using dynamic social norms, messenger effects, and processing fluency in combination will more effectively increase average end-of-year adherence level compared to usual care. Humana will conduct all study participant outreach and data analyses, which will be performed using routinely collected insurance claims data. Regulatory oversight is conducted using Humana's centralized institutional review board (IRB) of record. The work completed by Humana study staff is funded by Humana, Inc. Dr. Choudhry and his colleagues (including subaward recipients ideas42 and Tuck School of Business at Dartmouth) will provide technical and subject matter expertise related to study research design and implementation, protocol design, statistical analysis, publication (abstract, poster, manuscript) preparation and/or review, and assistance throughout the peer review process including revisions and additional analyses if necessary for this project. The work completed by study staff at Brigham and Women's Hospital, ideas42, and Tuck School of Business at Dartmouth is funded by NIA.
Interventions
BEHAVIORALScorecard
Participants will be mailed a scorecard from Humana reporting patients' medication adherence using a refill score.
BEHAVIORALSocial norms
Participants will be mailed a dynamic social norms messaging (noting the proportion of Humana members improving their medication refill scores).
BEHAVIORALMessenger effects
Participants will be mailed a scorecard coming from the trusted messenger of a Humana-identified pharmacist taking the same medication.
BEHAVIORALProcessing fluency
Participants will be mailed a modified scorecard increasing processing fluency through a visual metaphor of closing the ring.
Sponsors
Humana Inc.
National Institute on Aging (NIA)
Brigham and Women's Hospital
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
HEALTH_SERVICES_RESEARCH
Masking
DOUBLE (Investigator, Outcomes Assessor)
Eligibility
Sex/Gender
ALL
Age
65 Years to 79 Years
Healthy volunteers
No
Inclusion criteria
* Medicare Advantage beneficiary insured by Humana * Between the ages of 65 and 79 * Having at least two unique fills of any class of hypertension medication within the calendar year (2023). * Adherence level (as measured by the proportion of days covered \[PDC\] metric) between 60% and 85%
Exclusion criteria
* Beneficiaries in plans which, per contractual agreements, privacy policies and rules, or legal requirements, do not participate in research
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Average End-of-year Adherence Level | approximately 4 months | The primary outcome will be the average end-of-year adherence level in each arm (as measured by the Proportion of Days Covered metric \[PDC\]) stratified by refill cycle and adjusting for age, sex, and race. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Proportion of Participants With End-of-year PDC≥80% | approximately 4 months | As measured by the Proportion of Days Covered metric \[PDC\] |
Countries
United States
Participant flow
Recruitment details
We randomized 65,177 study participants before our study launched with recruitment starting in August 2023. We recruited in two waves (on August 18, 2023 and August 24, 2023) to ensure a sufficient sample size after adjusting for dropouts due to loss of eligibility or changes in membership.
Participants by arm
| Arm | Count |
|---|---|
| Baseline Mailing (Scorecard) Participants randomized to this arm will receive a scorecard from Humana reporting patients' medication adherence using a refill score. Two messages will be sent during the intervention: The first message will contain the baseline refill score. The second message will note any changes in the refill score.
Scorecard: Participants will be mailed a scorecard from Humana reporting patients' medication adherence using a refill score. | 9,198 |
| Baseline Mailing + Dynamic Social Norms Participants randomized to this arm will receive the Arm 1 scorecard plus dynamic social norms messaging (noting the proportion of Humana members improving their medication refill scores). Two messages will be sent during the intervention: The first message will contain the baseline refill score. The second message will note any changes in the refill score.
Scorecard: Participants will be mailed a scorecard from Humana reporting patients' medication adherence using a refill score.
Social norms: Participants will be mailed a dynamic social norms messaging (noting the proportion of Humana members improving their medication refill scores). | 9,180 |
| Baseline Mailing + Messenger Effects Participants randomized to this arm will receive the Arm 1 scorecard, coming from the trusted messenger of a Humana-identified pharmacist taking the same medication. Two messages will be sent during the intervention: The first message will contain the baseline refill score. The second message will note any changes in the refill score.
Scorecard: Participants will be mailed a scorecard from Humana reporting patients' medication adherence using a refill score.
Messenger effects: Participants will be mailed a scorecard coming from the trusted messenger of a Humana-identified pharmacist taking the same medication. | 9,180 |
| Baseline Mailing + Processing Fluency Participants randomized to this arm will receive a modified scorecard increasing processing fluency through a visual metaphor of closing the ring. Two messages will be sent during the intervention: The first message will contain the baseline refill score. The second message will note any changes in the refill score.
Processing fluency: Participants will be mailed a modified scorecard increasing processing fluency through a visual metaphor of closing the ring. | 9,193 |
| Processing Fluency + Dynamic Social Norms Participants randomized to this arm will receive the Arm 4 processing fluency scorecard plus dynamic social norms messaging. Two messages will be sent during the intervention: The first message will contain the baseline refill score. The second message will note any changes in the refill score.
Social norms: Participants will be mailed a dynamic social norms messaging (noting the proportion of Humana members improving their medication refill scores).
Processing fluency: Participants will be mailed a modified scorecard increasing processing fluency through a visual metaphor of closing the ring. | 9,179 |
| Processing Fluency + Messenger Effects Participants randomized to this arm will receive the Arm 4 processing fluency scorecard coming from the trusted messenger of a Humana-identified pharmacist taking the same medication. Two messages will be sent during the intervention: The first message will contain the baseline refill score. The second message will note any changes in the refill score.
Messenger effects: Participants will be mailed a scorecard coming from the trusted messenger of a Humana-identified pharmacist taking the same medication.
Processing fluency: Participants will be mailed a modified scorecard increasing processing fluency through a visual metaphor of closing the ring. | 9,177 |
| No Mailing (Usual Care) Participants randomized to this arm will not receive any mailed message. | 9,183 |
| Total | 64,290 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 | FG002 | FG003 | FG004 | FG005 | FG006 |
|---|---|---|---|---|---|---|---|---|
| Enrollment Wave 1 (August 18, 2023) | Excluded due to membership changes | 107 | 122 | 128 | 110 | 125 | 123 | 123 |
| Enrollment Wave 2 (August 24, 2023) | Excluded due to membership changes | 6 | 9 | 3 | 8 | 7 | 11 | 5 |
Baseline characteristics
| Characteristic | Baseline Mailing + Dynamic Social Norms | Total | No Mailing (Usual Care) | Baseline Mailing (Scorecard) | Processing Fluency + Messenger Effects | Processing Fluency + Dynamic Social Norms | Baseline Mailing + Processing Fluency | Baseline Mailing + Messenger Effects |
|---|---|---|---|---|---|---|---|---|
| Age, Continuous | 71.44 years STANDARD_DEVIATION 4.09 | 71.4 years STANDARD_DEVIATION 4.1 | 71.50 years STANDARD_DEVIATION 4.07 | 71.36 years STANDARD_DEVIATION 4.11 | 71.44 years STANDARD_DEVIATION 4.07 | 71.49 years STANDARD_DEVIATION 4.1 | 71.33 years STANDARD_DEVIATION 4.07 | 71.44 years STANDARD_DEVIATION 4.09 |
| Hypertension PDC ≥ 80% as of selection | 3615 Participants | 25792 Participants | 3722 Participants | 3690 Participants | 3650 Participants | 3690 Participants | 3720 Participants | 3705 Participants |
| Hypertension proportion of days covered (PDC) as of selection | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 | 0.77 proportion of days covered (PDC) STANDARD_DEVIATION 0.06 |
| Low-Income Subsidy (LIS)/Dual Eligibility Status LIS and Dual Eligible | 2364 Participants | 16376 Participants | 2296 Participants | 2407 Participants | 2294 Participants | 2330 Participants | 2396 Participants | 2289 Participants |
| Low-Income Subsidy (LIS)/Dual Eligibility Status Low-Income Subsidy (LIS) Only | 518 Participants | 3722 Participants | 566 Participants | 497 Participants | 567 Participants | 524 Participants | 514 Participants | 536 Participants |
| Low-Income Subsidy (LIS)/Dual Eligibility Status Medicaid Eligible only | 7 Participants | 72 Participants | 10 Participants | 9 Participants | 8 Participants | 16 Participants | 12 Participants | 10 Participants |
| Low-Income Subsidy (LIS)/Dual Eligibility Status No LIS/Dual Eligibility | 6291 Participants | 44120 Participants | 6311 Participants | 6285 Participants | 6308 Participants | 6309 Participants | 6271 Participants | 6345 Participants |
| Race/Ethnicity, Customized Black | 2140 Participants | 15237 Participants | 2168 Participants | 2229 Participants | 2143 Participants | 2114 Participants | 2218 Participants | 2225 Participants |
| Race/Ethnicity, Customized Other | 820 Participants | 5736 Participants | 829 Participants | 778 Participants | 845 Participants | 842 Participants | 814 Participants | 808 Participants |
| Race/Ethnicity, Customized Unknown | 306 Participants | 2180 Participants | 307 Participants | 309 Participants | 289 Participants | 320 Participants | 324 Participants | 325 Participants |
| Race/Ethnicity, Customized White | 5914 Participants | 41137 Participants | 5879 Participants | 5882 Participants | 5900 Participants | 5903 Participants | 5837 Participants | 5822 Participants |
| Refill cycle 30-day refill cycle | 1962 Participants | 13786 Participants | 1975 Participants | 1965 Participants | 1972 Participants | 1974 Participants | 1977 Participants | 1961 Participants |
| Refill cycle 90-day refill cycle | 7218 Participants | 50504 Participants | 7208 Participants | 7233 Participants | 7205 Participants | 7205 Participants | 7216 Participants | 7219 Participants |
| Sex: Female, Male Female | 5005 Participants | 34811 Participants | 4939 Participants | 4881 Participants | 4967 Participants | 5040 Participants | 4938 Participants | 5041 Participants |
| Sex: Female, Male Male | 4175 Participants | 29479 Participants | 4244 Participants | 4317 Participants | 4210 Participants | 4139 Participants | 4255 Participants | 4139 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk | EG002 affected / at risk | EG003 affected / at risk | EG004 affected / at risk | EG005 affected / at risk | EG006 affected / at risk |
|---|---|---|---|---|---|---|---|
| deaths Total, all-cause mortality | 0 / 9,198 | 0 / 9,180 | 0 / 9,180 | 0 / 9,193 | 0 / 9,179 | 0 / 9,177 | 0 / 9,183 |
| other Total, other adverse events | 0 / 9,198 | 0 / 9,180 | 0 / 9,180 | 0 / 9,193 | 0 / 9,179 | 0 / 9,177 | 0 / 9,183 |
| serious Total, serious adverse events | 0 / 9,198 | 0 / 9,180 | 0 / 9,180 | 0 / 9,193 | 0 / 9,179 | 0 / 9,177 | 0 / 9,183 |
Outcome results
Primary
Average End-of-year Adherence Level
The primary outcome will be the average end-of-year adherence level in each arm (as measured by the Proportion of Days Covered metric \[PDC\]) stratified by refill cycle and adjusting for age, sex, and race.
Time frame: approximately 4 months
| Arm | Measure | Value (MEAN) | Dispersion |
|---|---|---|---|
| Baseline Mailing (Scorecard) | Average End-of-year Adherence Level | 0.81 proportion of days covered | Standard Deviation 0.12 |
| Baseline Mailing + Dynamic Social Norms | Average End-of-year Adherence Level | 0.81 proportion of days covered | Standard Deviation 0.12 |
| Baseline Mailing + Messenger Effects | Average End-of-year Adherence Level | 0.81 proportion of days covered | Standard Deviation 0.12 |
| Baseline Mailing + Processing Fluency | Average End-of-year Adherence Level | 0.81 proportion of days covered | Standard Deviation 0.12 |
| Processing Fluency + Dynamic Social Norms | Average End-of-year Adherence Level | 0.81 proportion of days covered | Standard Deviation 0.12 |
| Processing Fluency + Messenger Effects | Average End-of-year Adherence Level | 0.81 proportion of days covered | Standard Deviation 0.12 |
| No Mailing (Usual Care) | Average End-of-year Adherence Level | 0.81 proportion of days covered | Standard Deviation 0.12 |
Secondary
Proportion of Participants With End-of-year PDC≥80%
As measured by the Proportion of Days Covered metric \[PDC\]
Time frame: approximately 4 months
| Arm | Measure | Value (COUNT_OF_PARTICIPANTS) |
|---|---|---|
| Baseline Mailing (Scorecard) | Proportion of Participants With End-of-year PDC≥80% | 6518 Participants |
| Baseline Mailing + Dynamic Social Norms | Proportion of Participants With End-of-year PDC≥80% | 6590 Participants |
| Baseline Mailing + Messenger Effects | Proportion of Participants With End-of-year PDC≥80% | 6467 Participants |
| Baseline Mailing + Processing Fluency | Proportion of Participants With End-of-year PDC≥80% | 6515 Participants |
| Processing Fluency + Dynamic Social Norms | Proportion of Participants With End-of-year PDC≥80% | 6558 Participants |
| Processing Fluency + Messenger Effects | Proportion of Participants With End-of-year PDC≥80% | 6598 Participants |
| No Mailing (Usual Care) | Proportion of Participants With End-of-year PDC≥80% | 6587 Participants |