Adherence, Medication, Atrial Fibrillation, Cardiovascular Diseases, Coronary Artery Disease, Diabetes Mellitus, Hyperlipidemias, Hypertension, Medication Adherence
Conditions
Keywords
cardiovascular disease, telemedicine, medication adherence, behavioral health
Brief summary
The study plans to learn if sending different text messages, serving as reminders or encouragement, may help patients take their medication more often if they have had trouble keeping up with their medicines.
Detailed description
Background: Up to fifty percent of patients do not take their cardiovascular medications as prescribed resulting in increased morbidity, mortality, and healthcare costs. Mobile and digital technologies for health promotion and disease self-management offer an intriguing and as of yet untested opportunity to adapt behavioral 'nudges' using ubiquitous cell phone technology to facilitate medication adherence. Objectives: Aim 1: Conduct a pragmatic patient-level randomized intervention across three health care systems (HCS) to improve adherence to chronic CV medications. The primary outcome will be medication adherence defined by the proportion of days covered (PDC) using pharmacy refill data. Secondary outcomes include clinical events (e.g., event times for stroke, MI, mortality), utilization of care (e.g., hospitalizations or clinic visits for CV-related reasons), and costs of healthcare utilization. Aim 2: Evaluate the intervention using a mixed methods approach and applying the RE-AIM (reach, effectiveness, adoption, implementation, and maintenance) framework. In addition, assess the context and implementation processes to inform local tailoring, adaptations and modifications, and eventual expansion of the intervention within the 3 HCS more broadly and nationally. Setting: The study will be conducted within three HCS in metro Denver: VA Eastern Colorado Health Care System (VA), Denver Health and Hospital Authority, and UCHealth.
Interventions
BEHAVIORALNudge
Interventions will include a variety of text messages aimed at improving medication adherence.
Sponsors
National Heart, Lung, and Blood Institute (NHLBI)
University of Colorado, Denver
Study design
Allocation
RANDOMIZED
Intervention model
PARALLEL
Primary purpose
PREVENTION
Masking
NONE
Intervention model description
The study will be a pragmatic, randomized controlled study with four treatment arms. Once patients are identified through pharmacy refill data to have a 7-day gap in any prescribed CV medication refills, they will be randomized to one of four arms, described in Intervention below. Randomization will be stratified within each of the clinics, and within strata defined by number of other CV medication classes that are prescribed at randomization (1-4), using blocks of 4 patients to ensure balance within clinics over time. Thus, within each clinic and number of other medication stratum, each set of 4 consecutively enrolled subjects will be randomized to the four study arms. Treatments will be initiated immediately upon randomization, in response to the 7-day gap.
Eligibility
Sex/Gender
ALL
Age
18 Years to 89 Years
Healthy volunteers
Yes
Inclusion criteria
* Patients with the following cardiovascular conditions and respective medication classes: * Hypertension (Beta-blockers \[B-blockers)\], Calcium Channel Blocker \[CCB\], Angiotensin converting enzyme inhibitors (ACEi), Angiotensin Receptor Blockers \[ARB\], or Thiazide diuretic) * Hyperlipidemia (HMG CoA reductase inhibitor \[Statins\]) * Diabetes (Alpha-glucosidase inhibitors, Biguanides, DPP-4 inhibitors, Sodium glucose transport inhibitor, Meglitinides, Sulfonylureas, Thiazolidinediones, or statins Coronary artery disease P2Y12 inhibitor \[Clopidogrel, Ticagrelor, Prasugrel, Ticlopidine\], B-blockers, ACEi or ARB or statins) * Atrial fibrillation (Direct oral anticoagulants, B-blockers, CCB)
Exclusion criteria
* Patients who do not have a mailing address listed in EHR; * Patients who do not have a landline or cellphone listed in EHR; * Currently pregnant if denoted in the EHR at the time of the data pull; * Patients with a mailing address outside of the state of Colorado; * Patients that do not speak either English or Spanish.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Medication Adherence | 12 months after intervention | The primary outcome was medication adherence defined by the proportion of days covered (PDC) using pharmacy refill data adjusted difference from usual care. |
| Mean Proportion of Days Covered Relative to Usual Care | 12 months | We analyzed this longitudinal data (up to 12 observations per patient) and estimated absolute differences in PDC between treatment arms and usual care using a Generalized Estimating Equation (GEE) model with an identity link and independence with unequal variances for the covariance structure of the 12 observations .A multistage gatekeeper approach was used to account for multiple treatment comparisons by comparing each of the 3 treatment arms to the control in stage 1 using significance levels of 0.05/3, and if any test was significant, a significance level of (R/3)\*(0.05/3) using the Holm method was used for the 3 pairwise comparisons, where R is the number of significant stage 1 tests. |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Hospitalization | one year | estimated probability of hospitalization at one year |
| Death | one year | estimated probability of death at one year |
| Emergency Department Visit | one year | estimated probability of emergency department visit at one year |
Countries
United States
Participant flow
Participants by arm
| Arm | Count |
|---|---|
| Generic Nudge A generic reminder text will be delivered to patients to refill their medication at days 1, 3, 5, 7 and 10 after they been labeled as non-adherent.
Nudge: Interventions will include a variety of text messages aimed at improving medication adherence. | 2,324 |
| Optimized Nudge An optimized nudge text will be delivered to patients to remind them to refill their medications at days 1, 3, 5, 7 and 10 after they have been labeled as non-adherent.
Nudge: Interventions will include a variety of text messages aimed at improving medication adherence. | 2,305 |
| Optimized Nudge Plus AI Chat Bot An optimized nudge text will be delivered to patients to remind them to refill their medications at days 1 and 3 after they have been labeled as non-adherent. If the patient has not filled their medication on days 5 and 7, in addition to receiving an optimized nudge text, an AI will conduct interactive chat via a chat bot to assess barriers filling the medication as described in Aim 1 above. If they still have not filled the medication, they will receive another message on day 10.
Nudge: Interventions will include a variety of text messages aimed at improving medication adherence. | 2,319 |
| Usual Care This group will not receive an intervention. We have included a usual care group to demonstrate the impact of the text messaging interventions above and beyond usual care given that many prior medication adherence interventions have demonstrated small to negligible effects. | 2,321 |
| Total | 9,269 |
Withdrawals & dropouts
| Period | Reason | FG000 | FG001 | FG002 | FG003 |
|---|---|---|---|---|---|
| Overall Study | Lost to Follow-up | 47 | 63 | 51 | 54 |
| Overall Study | opted out | 6 | 5 | 5 | 1 |
Baseline characteristics
| Characteristic | Generic Nudge | Usual Care | Optimized Nudge Plus AI Chat Bot | Optimized Nudge | Total |
|---|---|---|---|---|---|
| Age, Continuous | 59.9 years STANDARD_DEVIATION 12.5 | 60.1 years STANDARD_DEVIATION 12.6 | 60.1 years STANDARD_DEVIATION 12.7 | 60 years STANDARD_DEVIATION 12.9 | 60.0 years STANDARD_DEVIATION 12.7 |
| Baseline Medication Classes 1 Active Class | 597 Participants | 557 Participants | 567 Participants | 564 Participants | 2285 Participants |
| Baseline Medication Classes 2 Active Classes | 551 Participants | 591 Participants | 584 Participants | 572 Participants | 2298 Participants |
| Baseline Medication Classes >=3 Active Classes | 1176 Participants | 1173 Participants | 1168 Participants | 1169 Participants | 4686 Participants |
| Ethnicity (NIH/OMB) Hispanic or Latino | 1100 Participants | 1149 Participants | 1168 Participants | 1147 Participants | 4564 Participants |
| Ethnicity (NIH/OMB) Not Hispanic or Latino | 1204 Participants | 1150 Participants | 1134 Participants | 1141 Participants | 4629 Participants |
| Ethnicity (NIH/OMB) Unknown or Not Reported | 20 Participants | 22 Participants | 17 Participants | 17 Participants | 76 Participants |
| Insurance Commercial | 463 Participants | 441 Participants | 500 Participants | 471 Participants | 1875 Participants |
| Insurance Medicaid | 659 Participants | 665 Participants | 629 Participants | 632 Participants | 2585 Participants |
| Insurance Medicare | 853 Participants | 889 Participants | 860 Participants | 878 Participants | 3480 Participants |
| Insurance None | 218 Participants | 202 Participants | 192 Participants | 221 Participants | 833 Participants |
| Insurance Unknown | 123 Participants | 112 Participants | 130 Participants | 96 Participants | 461 Participants |
| Insurance VHA | 8 Participants | 12 Participants | 8 Participants | 7 Participants | 35 Participants |
| Intervention Delivery Interactive Voice Response Telephone Messages | 198 Participants | 0 Participants | 202 Participants | 216 Participants | 616 Participants |
| Intervention Delivery No Text or IVR | 0 Participants | 2321 Participants | 0 Participants | 0 Participants | 2321 Participants |
| Intervention Delivery Text Messages | 2126 Participants | 0 Participants | 2117 Participants | 2089 Participants | 6332 Participants |
| Marital Status Divorced/Widowed | 434 Participants | 483 Participants | 452 Participants | 464 Participants | 1833 Participants |
| Marital Status Married Married | 994 Participants | 950 Participants | 980 Participants | 940 Participants | 3864 Participants |
| Marital Status Single | 883 Participants | 874 Participants | 870 Participants | 883 Participants | 3510 Participants |
| Marital Status Unknown | 13 Participants | 14 Participants | 17 Participants | 18 Participants | 62 Participants |
| Medical History Cerebrovascular Disease | 142 participants | 134 participants | 125 participants | 145 participants | 546 participants |
| Medical History Chronic Kidney Disease | 191 participants | 189 participants | 203 participants | 202 participants | 785 participants |
| Medical History Depression | 421 participants | 416 participants | 442 participants | 463 participants | 1742 participants |
| Medical History Heart Failure | 160 participants | 171 participants | 163 participants | 199 participants | 693 participants |
| Medical History Posttraumatic Stress Disorder | 118 participants | 108 participants | 102 participants | 114 participants | 442 participants |
| Medical History Prior Coronary Revascularization | 63 participants | 58 participants | 70 participants | 56 participants | 247 participants |
| Medical History Prior Myocardial Infarction | 95 participants | 98 participants | 121 participants | 109 participants | 423 participants |
| Medical History Substance Abuse | 89 participants | 88 participants | 116 participants | 101 participants | 394 participants |
| Qualifying Condition >1 Qualifying Condition | 1406 participants | 1438 participants | 1410 participants | 1390 participants | 5644 participants |
| Qualifying Condition Atrial Fibrillation | 132 participants | 134 participants | 130 participants | 152 participants | 548 participants |
| Qualifying Condition Coronary Artery Disease | 305 participants | 328 participants | 352 participants | 325 participants | 1310 participants |
| Qualifying Condition Diabetes | 1148 participants | 1149 participants | 1162 participants | 1164 participants | 4623 participants |
| Qualifying Condition Hyperlipidemia | 1072 participants | 1054 participants | 1089 participants | 1052 participants | 4267 participants |
| Qualifying Condition Hypertension | 1837 participants | 1864 participants | 1821 participants | 1829 participants | 7351 participants |
| Race (NIH/OMB) American Indian or Alaska Native | 22 Participants | 35 Participants | 23 Participants | 27 Participants | 107 Participants |
| Race (NIH/OMB) Asian | 29 Participants | 29 Participants | 21 Participants | 31 Participants | 110 Participants |
| Race (NIH/OMB) Black or African American | 391 Participants | 392 Participants | 356 Participants | 378 Participants | 1517 Participants |
| Race (NIH/OMB) More than one race | 10 Participants | 9 Participants | 16 Participants | 14 Participants | 49 Participants |
| Race (NIH/OMB) Native Hawaiian or Other Pacific Islander | 3 Participants | 3 Participants | 6 Participants | 2 Participants | 14 Participants |
| Race (NIH/OMB) Unknown or Not Reported | 268 Participants | 255 Participants | 251 Participants | 238 Participants | 1012 Participants |
| Race (NIH/OMB) White | 1601 Participants | 1598 Participants | 1646 Participants | 1615 Participants | 6460 Participants |
| Sex: Female, Male Female | 1087 Participants | 1088 Participants | 1101 Participants | 1075 Participants | 4351 Participants |
| Sex: Female, Male Male | 1237 Participants | 1233 Participants | 1218 Participants | 1230 Participants | 4918 Participants |
Adverse events
| Event type | EG000 affected / at risk | EG001 affected / at risk | EG002 affected / at risk | EG003 affected / at risk |
|---|---|---|---|---|
| deaths Total, all-cause mortality | 63 / 2,324 | 53 / 2,305 | 55 / 2,319 | 68 / 2,321 |
| other Total, other adverse events | 612 / 2,324 | 594 / 2,305 | 616 / 2,319 | 639 / 2,321 |
| serious Total, serious adverse events | 267 / 2,324 | 283 / 2,305 | 283 / 2,319 | 303 / 2,321 |
Outcome results
Primary
Mean Proportion of Days Covered Relative to Usual Care
We analyzed this longitudinal data (up to 12 observations per patient) and estimated absolute differences in PDC between treatment arms and usual care using a Generalized Estimating Equation (GEE) model with an identity link and independence with unequal variances for the covariance structure of the 12 observations .A multistage gatekeeper approach was used to account for multiple treatment comparisons by comparing each of the 3 treatment arms to the control in stage 1 using significance levels of 0.05/3, and if any test was significant, a significance level of (R/3)\*(0.05/3) using the Holm method was used for the 3 pairwise comparisons, where R is the number of significant stage 1 tests.
Time frame: 12 months
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Generic Nudge | Mean Proportion of Days Covered Relative to Usual Care | 0.022 proportion of days covered |
| Optimized Nudge | Mean Proportion of Days Covered Relative to Usual Care | 0.020 proportion of days covered |
| Optimized Nudge Plus AI Chat Bot | Mean Proportion of Days Covered Relative to Usual Care | 0.023 proportion of days covered |
p-value: <0.05GEE
Primary
Medication Adherence
The primary outcome was medication adherence defined by the proportion of days covered (PDC) using pharmacy refill data adjusted difference from usual care.
Time frame: 12 months after intervention
Population: mean pdc at 12 months
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Generic Nudge | Medication Adherence | .606 mean PDC |
| Optimized Nudge | Medication Adherence | .620 mean PDC |
| Optimized Nudge Plus AI Chat Bot | Medication Adherence | .623 mean PDC |
| Usual Care | Medication Adherence | .630 mean PDC |
p-value: <0.05GEE
Secondary
Death
estimated probability of death at one year
Time frame: one year
Population: 1-year Kaplan-Meier rate estimates with 95% confidence intervals.
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Generic Nudge | Death | .031 estimated probability of death |
| Optimized Nudge | Death | .026 estimated probability of death |
| Optimized Nudge Plus AI Chat Bot | Death | .027 estimated probability of death |
| Usual Care | Death | .033 estimated probability of death |
Secondary
Emergency Department Visit
estimated probability of emergency department visit at one year
Time frame: one year
Population: 1-year Kaplan-Meier rate estimates with 95% confidence intervals.
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Generic Nudge | Emergency Department Visit | .292 estimated probability of ED visit |
| Optimized Nudge | Emergency Department Visit | .285 estimated probability of ED visit |
| Optimized Nudge Plus AI Chat Bot | Emergency Department Visit | .294 estimated probability of ED visit |
| Usual Care | Emergency Department Visit | .301 estimated probability of ED visit |
Secondary
Hospitalization
estimated probability of hospitalization at one year
Time frame: one year
Population: 1-year Kaplan-Meier rate estimates with 95% confidence intervals.
| Arm | Measure | Value (MEAN) |
|---|---|---|
| Generic Nudge | Hospitalization | .13 estimated probability of hospitalization |
| Optimized Nudge | Hospitalization | .136 estimated probability of hospitalization |
| Optimized Nudge Plus AI Chat Bot | Hospitalization | .136 estimated probability of hospitalization |
| Usual Care | Hospitalization | .143 estimated probability of hospitalization |