Key Points
Question
Can a primary care–based multicomponent intervention that includes automated identification of patients with uncontrolled hypertension and medication nonadherence combined with team-based care improve medication fill adherence?
Findings
In this pragmatic cluster randomized clinical trial of 1726 patients seen in 10 primary care clinics, the intervention was not associated with improvement in proportion of days covered for antihypertensive medications or blood pressure compared with usual care.
Meaning
While the intervention was able to automatically identify patients with medication nonadherence in clinical practice, it did not lead to improved medication adherence for patients with uncontrolled hypertension.
Abstract
Importance
Medication nonadherence is present in nearly half of patients with hypertension but is underrecognized in clinical care. Data linkages between electronic health records and pharmacies have created opportunities for scalable assessment of medication adherence at the point of care.
Objective
To test the effectiveness of a multicomponent intervention that identified patients with uncontrolled hypertension and medication nonadherence using linked electronic health record–pharmacy data combined with team-based care to address adherence barriers.
Design, Setting, and Participants
TEAMLET (Leveraging Electronic Health Record Technology and Team Care to Address Medication Adherence) was a pragmatic, 2-arm, cluster randomized clinical trial conducted between October 2022 and November 2024 in 10 primary care sites in New York. The study included adults with uncontrolled hypertension and low medication adherence, defined as proportion of days covered (PDC) less than 80%. Data analysis was performed from November 2024 to January 2025.
Intervention
The intervention consisted of the following: (1) automated identification of patients with medication nonadherence at the time of the visit; (2) prompting of medical assistants to screen for barriers to adherence; (3) clinical decision support alerting the primary care physicians and nurse practitioners to barriers to adherence; and (4) adherence discussion between the primary care physician or nurse practitioner and the patient. The comparator was usual care.
Main Outcomes and Measures
The primary outcome was change in PDC from baseline to 12 months.
Results
Among 1726 patients (mean [SD] age, 67.2 [13.9] years; 887 [51.4%] female), the mean (SD) baseline PDC was 33.2% (30.5%) overall (32.4% [30.4%] in the intervention group and 34.0% [30.6%] in the control group). The mean (SD) PDC at 12 months was 51.1% (39.5%) for the intervention group and 53.1% (39.6%) for the control group. No difference was found in the change in PDC from baseline to 12 months between the intervention and control groups (mean [SD] absolute change in PDC, 18.5 [41.1] vs 18.2 [40.9] percentage points, respectively; adjusted difference, −0.15 percentage point; 95% CI, −4.06 to 3.76 percentage points). Change in systolic blood pressure and patients who became adherent (PDC ≥80%) at 12 months were also similar between groups.
Conclusions and Relevance
In this pragmatic trial, an intervention that combined team-based primary care with automated identification of patients with antihypertensive medication nonadherence did not lead to improvements in adherence or blood pressure.
Trial Registration
ClinicalTrials.gov Identifier: NCT05349422
This cluster randomized clinical trial evaluates whether a multicomponent intervention that includes automated identification of patients with uncontrolled hypertension and medication nonadherence combined with team-based care improves medication fill adherence.
Introduction
Medication nonadherence is present in nearly half of patients with hypertension and is a major contributor to inadequate blood pressure control.1,2,3,4 Poor medication adherence has been associated with increased rates of cardiovascular events, hospitalizations, and mortality in patients with hypertension.5,6 For instance, one study estimated that increasing antihypertensive medication adherence by 15% would reduce the hazards of stroke by 9% and mortality by 7%.5 Although physician counseling can improve medication adherence,2,7 physicians may be unaware of gaps in their patients’ adherence to antihypertensive therapy more than half the time.8,9,10,11,12 Therefore, clinicians rarely address medication adherence in practice. Even when medication adherence is discussed, clinicians have limited time for high-level counseling and shared decision-making.11,13,14
Data linkages between electronic health records (EHRs) and pharmacies have led to the capability to assess medication adherence at the point of care. Specifically, linked EHR-pharmacy data can track whether a patient is regularly filling their medications, a common method to assess adherence.15,16 This information can be used to alert clinicians to adherence gaps that may otherwise be missed.
The Leveraging EHR Technology and Team Care to Address Medication Adherence (TEAMLET) intervention used linked EHR-pharmacy data to alert primary care physicians (PCPs) and nurse practitioners (NPs) of medication nonadherence among patients with uncontrolled blood pressure at the point of care.17 To address limitations in clinician time, the intervention also included an EHR-based workflow enhancement that incorporated team-based care to address medication adherence.18,19 The purpose of this study was to test whether the multicomponent intervention could improve medication adherence for patients with hypertension compared with usual care.
Methods
We conducted a pragmatic, 2-arm, cluster randomized clinical trial to assess the effectiveness of a multicomponent intervention on medication adherence and blood pressure control in patients with hypertension. Details of the study design were previously described.17 The trial protocol is available in Supplement 1. The trial was conducted across 10 primary care practices within the NYU Langone Health network, a large academic health system that serves New York, New York, and surrounding areas and that uses a single EHR (Epic; Epic Systems Corp). We implemented a staggered practice enrollment approach between October 6, 2022, and March 16, 2023. The primary data source was the EHR, which also included information on pharmacy medication fills through linkage to Surescripts (Surescripts, LLC).15,16
Patients were included if they were aged 18 years or older and had a clinical encounter at 1 of the 10 study sites during the first 8 months of that practice’s trial enrollment. Other inclusion criteria were a diagnosis of hypertension, an active order of at least 1 antihypertensive medication, uncontrolled blood pressure, and low medication adherence. To ensure specificity, we defined uncontrolled blood pressure based on 2 encounters: 140/90 mm Hg or higher on the day of the visit and 140/90 mm Hg or higher at the previous outpatient visit; in both cases, the presence of either elevated systolic or elevated diastolic blood pressure was sufficient. We defined low medication adherence based on the proportion of days covered (PDC), calculated as the percentage of days on which a patient possessed antihypertensive medications during the prior 180 days, being less than 80%.20,21 For study inclusion, we used PDC that was calculated within the EHR. All patients who met the inclusion criteria were automatically enrolled in the study. This study was approved by the NYU Institutional Review Board, which granted a waiver of informed consent for patient data collected as part of routine clinical care. Data analysis was performed from November 2024 to January 2025. The study follows the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.
Intervention
Intervention development was guided by the capability-opportunity-motivation behavior framework, which posits that interaction between the 3 components is necessary for successful behavior change.22 The intervention consisted of 4 key components17 (eFigure in Supplement 2). First, patients presenting for a clinical visit were automatically screened, using an EHR algorithm, for uncontrolled hypertension, ie, 2 recent blood pressure measurements of 140/90 mm Hg or higher, and low medication adherence using linked EHR-pharmacy data. If patients met these criteria, the remainder of the workflow was initiated. Second, medical assistants (MAs) received an interruptive alert within the EHR for all included patients. The MAs were prompted to administer a brief questionnaire, derived from a validated measure,23 about 6 common reasons for medication nonadherence. MAs documented the patient response using predefined options of the 6 barriers, self-reported adherence, or other. Third, PCPs were notified of low adherence and the questionnaire results through 2 methods: MAs were prompted to send a message with the results to the PCPs, and PCPs received a passive clinical decision support (CDS) alert listing the barrier and a link to an order set. Fourth, PCPs were encouraged by the alert to discuss medication adherence with the patient. The CDS order set included 3 orders: automated documentation of a discussion of barriers into the progress note, addition of visit diagnostic codes related to medication adherence and hypertension, and inclusion of tailored patient education materials targeting the patient’s specific adherence barrier in the after-visit summary. CDS recommendations were tailored to the barrier identified. For example, if financial concerns were identified, the CDS recommended exploring cost-effective options, such as switching to generic medications or using pharmacy assistance programs.
PCPs and MAs in the intervention group received both initial and ongoing training on the intervention components, general information on medication adherence, and health coaching best practices to address nonadherence.
Control
Patients in the control group received usual care. Additionally, PCPs and MAs received an email at the beginning of the study with information on medication adherence and an overview of health coaching to support behavior change. The email also contained information on an active tool in the EHR that presented PDC for each active medication during a clinical encounter.
Clinical Outcome Measures
The primary study outcome was change in PDC from baseline to 12 months. For outcomes, PDC was calculated over the prior 180 days using medication order information from the EHR and medication fill information available in the linked EHR-pharmacy data at the time of incident encounter visit (baseline) and at 12 months’ follow-up.
The secondary clinical outcome was change in systolic blood pressure from baseline to 12 months, measured as a continuous variable. Baseline blood pressure was collected at the incident encounter visit. Systolic blood pressure at 12 months was based on the mean of the last 2 outpatient systolic blood pressure values recorded in the EHR within 12 months of enrollment. We also assessed the percentage of patients considered adherent as a binary outcome, based on a PDC of 80% or higher, at the end of the study period.
Other Measures
Baseline variables included age, sex, race and ethnicity, number of active antihypertensive medications at time of the incident clinical encounter, insurance, comorbidities, PDC for antihypertensive medications, and systolic and diastolic blood pressure. Race and ethnicity were based on self-report in the EHR and were categorized as Hispanic, non-Hispanic Black or African American, non-Hispanic White, non-Hispanic other (including non-Hispanic American Indian or Alaska Native and non-Hispanic Asian; collapsed due to small numbers), or unknown. Insurance was based on primary insurance for the specific visit and was categorized as Medicare, Medicaid, commercial, or missing. Comorbidities were based on the Elixhauser definitions plus Medicare definition for ischemic heart disease.24,25
To understand clinical care intensity for patients in both the intervention and control groups, the number of individual patient follow-up clinic visits across the 10 practices during the 12-month follow-up period was also collected.
Implementation Outcomes
Using EHR data, we identified adoption of various components of the CDS tools in the intervention group, including the percentage of times the intervention was initiated for eligible patients, MAs completed the barrier questionnaire, PCPs interacted with the CDS vs dismissed it immediately, and PCPs ordered individual components of the CDS order set.
We also evaluated implementation fidelity through performing medical record review of the incident clinical encounters. We randomly sampled 20% of all enrolled patients, including both the intervention and control groups, and reviewed progress notes for presence of documentation of hypertension, medication adherence, and whether the elevated blood pressure was addressed during the encounter. Two reviewers independently reviewed one-third of these medical records with high agreement (mean κ statistic = 0.90). Any discrepancies were discussed until consensus was reached. The remaining medical records were reviewed by a single reviewer.
Sample Size
Our sample size estimate was based on a primary outcome of change in PDC from baseline to 12 months. We calculated that a sample size of 1694 patients across 10 sites would provide 80% power to detect at least a 20% difference between study groups, assuming a type I error rate of 5%, attrition of 30% from baseline to 12 months, and an intraclass correlation coefficient of 0.15 between sites.
Statistical Analysis
Descriptive statistics were used to summarize baseline variables and outcomes. To compare the clinical outcomes of change in PDC and change in systolic blood pressure between the intervention and control groups, we independently fit generalized linear mixed-effects models26,27 with PDC and blood pressure each serving as the dependent variable. Each model included a binary variable for the randomized study group, a binary time variable of baseline vs 12 months, and an interaction between group and time. Models were adjusted for covariates imbalanced at baseline. Each model also included practice-level random effects to account for clustering of patients within clinics.
We also performed exploratory stratified analyses for subgroups of age (≥65 vs <65 years), sex, and race and ethnicity. Additionally, stratified analyses by number of clinic visits during the 12-month follow-up were categorized by quantile. For stratified analyses of variables with multiple categories, we created a binary variable of each category of interest vs all other categories.
We calculated sensitivity of the intervention trigger as the number of patients for whom the MA questionnaire alert was triggered over the total number of patients who met inclusion criteria. We used χ2 tests to evaluate differences in implementation outcomes between the control and intervention groups.
All tests were 2-sided, with a type I error rate of .05. Analyses were performed using R version 4.3.0 statistical software (R Project for Statistical Computing).
Results
A total of 1726 patients from 94 PCPs or NPs were enrolled in the study, with 919 patients in the intervention group and 807 in the control group (Figure). The mean (SD) age of the cohort was 67.2 (13.9) years, and 887 (51.4%) were female (Table 1). Patients in the intervention group were more likely than those in the control group to be female (53.9% vs 48.6%) and to have a race and ethnicity of non-Hispanic Black or African American (24.2% vs 18.3%). At baseline, the mean (SD) PDC score was 33.2% (30.5%) overall, 32.4% (30.4%) in the intervention patients, and 34.0% (30.6%) in the control patients.
Figure. Study Flow Diagram.
TEAMLET indicates Leveraging Electronic Health Record Technology and Team Care to Address Medication Adherence.
Table 1. Patient-Level Baseline Characteristics.
| Characteristic | Intervention (n = 919) | Control (n = 807) |
|---|---|---|
| Age, mean (SD), y | 67.6 (14.3) | 66.6 (13.5) |
| Sex, No. (%) | ||
| Female | 495 (53.9) | 392 (48.6) |
| Male | 424 (46.1) | 415 (51.4) |
| Race and ethnicitya | ||
| Hispanic | 100 (10.9) | 67 (8.3) |
| Non-Hispanic Black or African American | 222 (24.2) | 148 (18.3) |
| Non-Hispanic White | 468 (50.9) | 475 (58.9) |
| Non-Hispanic otherb | 42 (4.6) | 36 (4.5) |
| Unknown | 87 (9.5) | 81 (10.0) |
| Active medication count, mean (SD), No. | 2.08 (0.97) | 2.09 (1.00) |
| Insurance, No. (%) | ||
| Medicare | 484 (52.7) | 410 (50.8) |
| Commercial | 365 (39.7) | 331 (41.0) |
| Medicaid | 65 (7.1) | 59 (7.3) |
| Missing | 5 (0.5) | 7 (0.9) |
| Comorbidities | ||
| Cerebrovascular disease | 70 (7.6) | 50 (6.2) |
| Heart failure | 52 (5.7) | 30 (3.7) |
| Dementia | 13 (1.4) | 7 (0.9) |
| Depression | 60 (6.5) | 46 (5.7) |
| Diabetes | 226 (24.6) | 243 (30.1) |
| Obesity | 201 (21.9) | 161 (20.0) |
| Peripheral vascular disease | 78 (8.5) | 56 (6.9) |
| Kidney disease and failure | 103 (11.2) | 83 (10.3) |
| Alcohol abuse | 11 (1.2) | 10 (1.2) |
| Cardiovascular disease | 129 (14.0) | 94 (11.6) |
| PDC, mean (SD), % | 32.4 (30.4) | 34.0 (30.6) |
| Blood pressure, mean (SD), mm Hg | ||
| Systolic | 149.0 (14.3) | 148.0 (12.3) |
| Diastolic | 84.7 (11.6) | 84.4 (10.5) |
Abbreviation: PDC, proportion of days covered.
Based on self-report in the electronic health record.
Includes includes non-Hispanic American Indian or Alaska Native and non-Hispanic Asian. Categories were collapsed owing to small counts.
Mean (SD) PDC at 12 months was 51.1% (39.5%) for the intervention group and 53.1% (39.6%) for the control group. The primary outcome of mean (SD) absolute change in PDC was an increase of 18.5 (41.1) percentage points among intervention patients and 18.2 (40.9) percentage points among control patients (Table 2), with a difference of −0.17 (95% CI, −4.08 to 3.74) percentage points after accounting for clustering by site. After adjusting for baseline age, race and ethnicity, and diabetes status, the difference in change in PDC between groups was −0.15 (95% CI, −4.06 to 3.76) percentage points.
Table 2. Change in Medication Adherence as Measured by Proportion of Days Covered (PDC) and Blood Pressure Over 12-Month Study Period.
| Outcome | Intervention (n = 919) | Control (n = 807) | Adjusted | |
|---|---|---|---|---|
| Estimate or odds ratio (95% CI) | P value | |||
| PDC change from baseline to 12 mo, mean (SD), percentage points | 18.5 (41.1) | 18.2 (40.9) | −0.15 (−4.06 to 3.76)a | .94 |
| Patients with PDC ≥80%, No. (%) | 323 (35.1) | 300 (37.2) | 1.03 (0.63 to 1.69)b | .89 |
| Systolic blood pressure change from baseline to 12 mo, mean (SD), mm Hg | −11.6 (17.8) | −12.2 (16.8) | 0.75 (−0.94 to 2.44)a | .38 |
Expressed as estimate.
Expressed as odds ratio.
Secondary outcomes were also similar between groups. The proportion of patients considered adherent, ie, those with a PDC of 80% or higher, at 12 months was 35.1% in the intervention group and 37.2% in the control group (Table 2). After adjustment for covariates imbalanced at baseline and clustering by site, the odds ratio for intervention vs control patients who became adherent at 12 months was 1.03 (95% CI, 0.63 to 1.69). The mean (SD) change in systolic blood pressure from baseline to 12 months was −11.6 (17.8) mm Hg for the intervention patients and −12.2 (16.8) mm Hg for the control patients, with no difference between the intervention and control groups after adjustment (adjusted estimate, 0.75 [95% CI, −0.94 to 2.44]).
We found no difference between the intervention and control patients for change in PDC among subgroups of age and sex (eTable 1 in Supplement 2). The intervention was associated with reduced change in PDC from baseline to 12 months among patients who self-identified as Hispanic compared with their control counterparts (mean [SD] absolute PDC change, 9.6 [40.5] vs 25.6 [35.9] percentage points; adjusted estimate, −17.97 [95% CI, −31.33 to −4.61] percentage points). Among patients with no follow-up PCP visits after the incident clinical encounter, those in the intervention group had a greater increase in PDC compared with those in the control group (mean [SD] absolute PDC change, 17.1 [44.0] vs 3.7 [42.8] percentage points; adjusted estimate, 13.95 [95% CI, 3.42 to 24.48] percentage points) (eTable 1 in Supplement 2). There was no difference between groups among patients who had 1, 2, or 3 or more follow-up visits.
The MA barrier questionnaire was triggered for 889 of 919 intervention patients, for a sensitivity of intervention initiation of 96.7% (Table 3). This alert was also triggered for 16 patients who did not meet inclusion criteria for the intervention group, for a positive predictive value of 98.2%. The barrier questionnaire was completed by MAs for 724 intervention patients (78.8%). The most common response to the questionnaire was “I always take my medications” (50.3%), followed by “other” (25.3%) and “I forgot” (11.7%) (eTable 2 in Supplement 2).
Table 3. Penetrance and Adoption of TEAMLET Intervention Among Intervention Group Patients.
| Intervention | Intervention patients, No. (%) (n = 919) |
|---|---|
| MA received barrier questionnaire | 889 (96.7) |
| MA completed barrier questionnaire | 724 (78.8) |
| PCP opened order set | 109 (11.9) |
| Documentation added to PCP progress note from order set | 96 (10.4) |
| Patient education handout added to after-visit summary from order set | 85 (9.2) |
Abbreviations: MA, medical assistant; PCP, primary care physician; TEAMLET, Leveraging Electronic Health Record Technology and Team Care to Address Medication Adherence.
Among all patients in the intervention group, 109 (11.9%) had their PCP or NP open the order set from the PCP alert. Once the order set was opened, use was high: the clinician added the automated documentation related to medication adherence for 96 of the 109 patients (88.1%) and patient education materials for 85 (78.0%). We found that the rate of clinician opening the order set varied across sites, with a range of 2.6% to 25.4%.
In a 20% random sample, we found in medical record review that 73.9% of patients had mention of hypertension in the PCP or NP progress note of the visit, with similar rates for intervention and control patients (Table 4). There was documentation that clinicians addressed patients’ elevated blood pressure in 36.3% of intervention patient encounters and 32.9% of control patient encounters (P = .59; Table 4). Medication adherence was documented in the progress note for 26.4% of intervention patients and 17.1% of control patients, although this difference did not reach statistical significance (P = .05).
Table 4. Electronic Health Record Documentation of Intervention Adoption in a Random Sample of 20% of Patients.
| Intervention | Patients, No. (%) | P value | |
|---|---|---|---|
| Intervention (n = 193) | Control (n = 152) | ||
| Hypertension indicated in progress note | 142 (73.6) | 113 (74.3) | .97 |
| Addressed elevated blood pressurea | 70 (36.3) | 50 (32.9) | .59 |
| New medication added | 23 (11.9) | 12 (7.9) | NA |
| Medication dose increased | 16 (8.3) | 13 (8.6) | NA |
| Medication changed | 9 (4.7) | 2 (1.3) | NA |
| Lifestyle recommendations | 28 (14.5) | 31 (20.4) | NA |
| Patient refused treatment | 3 (1.6) | 1 (0.7) | NA |
| Adherence indicated in progress note | 51 (26.4) | 26 (17.1) | .05 |
Abbreviation: NA, not applicable.
Percentages in the subcategories are not mutually exclusive, meaning that a single patient encounter could be included in multiple subcategories.
Discussion
In this pragmatic, cluster randomized trial of 1726 patients with persistent high blood pressure and low medication adherence, a practice-based multicomponent intervention was not associated with improvements in adherence or blood pressure. At baseline, included patients had antihypertensive medications available on only approximately one-third of days. While medication adherence, as measured by PDC, increased by 18.5% in the intervention group, this improvement was similar to that in the control group (18.2%) and still represented inadequate adherence at follow-up.
The intervention was developed in consideration of 2 key strategies to address gaps in medication adherence for hypertension. First was the use of linked EHR-pharmacy data to automatically identify patients with low medication adherence and high blood pressure at the point of care. In general, we found that we were successful in this approach: we estimated that the intervention captured 96.7% of patients who met inclusion criteria and was appropriately triggered 98.2% of the time. A related strength was the use of the same linked EHR-pharmacy data for assessment of outcomes. This pragmatic strategy of using clinical data for intervention triggering and evaluation creates opportunities for future scalable interventions to address medication adherence in real-time clinical care.
The second intervention strategy was the incorporation of team-based care to address medication nonadherence. Our results suggested that the implementation of team-based care had mixed success. Adoption by MAs was high, as approximately three-quarters of MAs submitted a response to the patient barrier questionnaire, compared with typical alert uptake of 10% to 40%.28,29 However, approximately three-quarters of patients identified as having low adherence by PDC responded with either “I always take my medications” or “other,” suggesting that MAs may not have always asked the barrier questionnaire as intended or that patients may not always have answered accurately.30 Additionally, we found that only approximately 1 in 8 PCPs opened the order set. Nonetheless, the intervention was associated with a 9% increase in PCP documentation of medication adherence in a sample of patients. In total, these findings suggest opportunity to improve delivery of the current and future interventions through improvement in implementation of team-based care, including engagement of MAs beyond a single questionnaire; audit and feedback approaches targeting increased uptake by PCPs; and possibly task-shifting discussions to pharmacists, who have been successful in improving medication adherence for hypertension.2 We are conducting qualitative interviews with PCPs, MAs, and patients to better understand barriers and facilitators to intervention adoption and fidelity.
Our observation of improvement in PDC and significant reduction in blood pressure in the control group may have been due to regression to the mean.31 It is also possible that our initial trainings increased PCP awareness of PDC measurements available in the EHR, thus increasing recognition of medication nonadherence in the control group. PCPs may have been aware that they were being measured for the study, and contamination across groups may have occurred.
The intervention was associated with a significant increase in pharmacy fill adherence among a subgroup of patients who did not have primary care follow-up in the subsequent year. Although this finding was exploratory, it may have reflected a greater impact of the intervention on patients who experience challenges in seeing their physician. Prior observational studies suggested that physician follow-up was associated with increased cardiovascular medication adherence,32,33,34 a finding that may have been associated with increased opportunities for clinicians to screen for and address medication adherence during a visit.32 Such opportunities may not be available for patients with infrequent follow-up.
In general, addressing adherence is challenging, and most prior interventions have had limited success in improving medication adherence.30,35 For instance, in 5 studies in which clinicians received feedback about their patients’ medication adherence, none were associated with improved adherence.36 Unlike our intervention, these studies were unable to provide real-time patient data at the point of care. Furthermore, although our intervention included team-based care, an approach that is guideline recommended for hypertension,37 we found potential limitations in its implementation. These findings should not imply lack of effectiveness for team-based care more generally. Given prior success with hypertension management, inclusion of a pharmacist may have been helpful, although this may have also affected the cost of the intervention.2,37
Limitations
Our results should be interpreted in the context of potential limitations. First, our study took place at practices within a single health system that is linked by 1 EHR. Nonetheless, Epic is the largest EHR vendor in the United States,38,39 and pharmacy fill data are available with other EHR vendors.40 Second, our PDC measurement relied on accurate recording of medication fills in the Surescripts data, which may not receive data from all pharmacies or pharmacy benefit managers. Nonetheless, we previously found the Surescripts database captured more than 90% of fills observed in commercial insurance claims, which are commonly used for the PDC measurement, and contained many fills unavailable in insurance data.15 Third, measurement of adherence was dependent on accurate documentation of medications prescribed in the EHR. If patients were advised verbally to take medications differently than prescribed, that may have affected the PDC. Fourth, our measurement of blood pressure was dependent on values obtained during an encounter so did not generally occur exactly at 12 months. We took the mean of the last 2 blood pressure measurements to improve measurement reliability.41
Conclusions
In this pragmatic trial of patients with uncontrolled hypertension seen in primary care, we implemented an intervention that used linked EHR-pharmacy data to automatically identify patients with low medication adherence at the point of care. The intervention, which also incorporated team-based care, was not associated with improvements in adherence or blood pressure. Our findings suggest that the intervention, which was effective at identifying patients with high blood pressure and low medication adherence, could be improved by strengthening implementation of team-based care, an effective strategy for management of hypertension.42
Trial Protocol
eFigure. Intervention In-Clinic Workflow
eTable 1. Mean (SD) Change in Proportion of Days Covered (PDC) From Baseline to 12 Months, in Exploratory Analysis of Patient Subgroups
eTable 2. Responses to Questionnaire From Medical Assistants (MAs)
Data Sharing Statement
References
- 1.Chang TE, Ritchey MD, Park S, et al. National rates of nonadherence to antihypertensive medications among insured adults with hypertension, 2015. Hypertension. 2019;74(6):1324-1332. doi: 10.1161/HYPERTENSIONAHA.119.13616 [DOI] [PubMed] [Google Scholar]
- 2.Kini V, Ho PM. Interventions to improve medication adherence: a review. JAMA. 2018;320(23):2461-2473. doi: 10.1001/jama.2018.19271 [DOI] [PubMed] [Google Scholar]
- 3.Naderi SH, Bestwick JP, Wald DS. Adherence to drugs that prevent cardiovascular disease: meta-analysis on 376,162 patients. Am J Med. 2012;125(9):882-7.e1. doi: 10.1016/j.amjmed.2011.12.013 [DOI] [PubMed] [Google Scholar]
- 4.Neutel JM, Smith DH. Improving patient compliance: a major goal in the management of hypertension. J Clin Hypertens (Greenwich). 2003;5(2):127-132. doi: 10.1111/j.1524-6175.2003.00495.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Bailey JE, Wan JY, Tang J, Ghani MA, Cushman WC. Antihypertensive medication adherence, ambulatory visits, and risk of stroke and death. J Gen Intern Med. 2010;25(6):495-503. doi: 10.1007/s11606-009-1240-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ho PM, Magid DJ, Shetterly SM, et al. Medication nonadherence is associated with a broad range of adverse outcomes in patients with coronary artery disease. Am Heart J. 2008;155(4):772-779. doi: 10.1016/j.ahj.2007.12.011 [DOI] [PubMed] [Google Scholar]
- 7.Schoenthaler A, Chaplin WF, Allegrante JP, et al. Provider communication effects medication adherence in hypertensive African Americans. Patient Educ Couns. 2009;75(2):185-191. doi: 10.1016/j.pec.2008.09.018 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Heeb RM, Kreuzberg V, Grossmann V. Physicians’ assessment of medication adherence: a systematic review. J Pharm Care Health Syst. 2019;6(1):1-7. doi: 10.4172/2376-0419.1000202 [DOI] [Google Scholar]
- 9.Sidorkiewicz S, Tran VT, Cousyn C, Perrodeau E, Ravaud P. Discordance between drug adherence as reported by patients and drug importance as assessed by physicians. Ann Fam Med. 2016;14(5):415-421. doi: 10.1370/afm.1965 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Zeller A, Taegtmeyer A, Martina B, Battegay E, Tschudi P. Physicians’ ability to predict patients’ adherence to antihypertensive medication in primary care. Hypertens Res. 2008;31(9):1765-1771. doi: 10.1291/hypres.31.1765 [DOI] [PubMed] [Google Scholar]
- 11.Bokhour BG, Berlowitz DR, Long JA, Kressin NR. How do providers assess antihypertensive medication adherence in medical encounters? J Gen Intern Med. 2006;21(6):577-583. doi: 10.1111/j.1525-1497.2006.00397.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Meddings J, Kerr EA, Heisler M, Hofer TP. Physician assessments of medication adherence and decisions to intensify medications for patients with uncontrolled blood pressure: still no better than a coin toss. BMC Health Serv Res. 2012;12(1):270. doi: 10.1186/1472-6963-12-270 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Tarn DM, Mattimore TJ, Bell DS, Kravitz RL, Wenger NS. Provider views about responsibility for medication adherence and content of physician-older patient discussions. J Am Geriatr Soc. 2012;60(6):1019-1026. doi: 10.1111/j.1532-5415.2012.03969.x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bosworth HB, Fortmann SP, Kuntz J, et al. Recommendations for providers on person-centered approaches to assess and improve medication adherence. J Gen Intern Med. 2017;32(1):93-100. doi: 10.1007/s11606-016-3851-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Blecker S, Adhikari S, Zhang H, et al. Validation of EHR medication fill data obtained through electronic linkage with pharmacies. J Manag Care Spec Pharm. 2021;27(10):1482-1487. doi: 10.18553/jmcp.2021.27.10.1482 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Blecker S, Zhao Y, Li X, et al. Approach to estimating adherence to heart failure medications using linked electronic health record and pharmacy data. J Gen Intern Med. 2025;40(4):811-817. doi: 10.1007/s11606-024-09216-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Blecker S, Schoenthaler A, Martinez TR, et al. Leveraging electronic health record technology and team care to address medication adherence: protocol for a cluster randomized controlled trial. JMIR Res Protoc. 2023;12:e47930. doi: 10.2196/47930 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hopkins K, Sinsky CA. Team-based care: saving time and improving efficiency. Fam Pract Manag. 2014;21(6):23-29. [PubMed] [Google Scholar]
- 19.Ho PM, Lambert-Kerzner A, Carey EP, et al. Multifaceted intervention to improve medication adherence and secondary prevention measures after acute coronary syndrome hospital discharge: a randomized clinical trial. JAMA Intern Med. 2014;174(2):186-193. doi: 10.1001/jamainternmed.2013.12944 [DOI] [PubMed] [Google Scholar]
- 20.Choudhry NK, Shrank WH, Levin RL, et al. Measuring concurrent adherence to multiple related medications. Am J Manag Care. 2009;15(7):457-464. [PMC free article] [PubMed] [Google Scholar]
- 21.Martin BC, Wiley-Exley EK, Richards S, Domino ME, Carey TS, Sleath BL. Contrasting measures of adherence with simple drug use, medication switching, and therapeutic duplication. Ann Pharmacother. 2009;43(1):36-44. doi: 10.1345/aph.1K671 [DOI] [PubMed] [Google Scholar]
- 22.Michie S, van Stralen MM, West R. The behaviour change wheel: a new method for characterising and designing behaviour change interventions. Implement Sci. 2011;6(1):42. doi: 10.1186/1748-5908-6-42 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Voils CI, Maciejewski ML, Hoyle RH, et al. Initial validation of a self-report measure of the extent of and reasons for medication nonadherence. Med Care. 2012;50(12):1013-1019. doi: 10.1097/MLR.0b013e318269e121 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Agency for Healthcare Research and Quality . Comorbidity Software: ICD-10-CM/PCS. Healthcare Cost and Utilization Project (HCUP). Accessed December 6, 2024. https://hcup-us.ahrq.gov/toolssoftware/comorbidityicd10/comorbidity_icd10.jsp#archive [PubMed]
- 25.Chronic Conditions Data Warehouse . 30 CCW chronic conditions algorithms. Accessed July 1, 2023. https://www2.ccwdata.org/web/guest/condition-categories-chronic
- 26.Bell ML, Rabe BA. The mixed model for repeated measures for cluster randomized trials: a simulation study investigating bias and type I error with missing continuous data. Trials. 2020;21(1):148. doi: 10.1186/s13063-020-4114-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Offorha BC, Walters SJ, Jacques RM. Analysing cluster randomised controlled trials using GLMM, GEE1, GEE2, and QIF: results from four case studies. BMC Med Res Methodol. 2023;23(1):293. doi: 10.1186/s12874-023-02107-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Kouri A, Yamada J, Lam Shin Cheung J, Van de Velde S, Gupta S. Do providers use computerized clinical decision support systems? a systematic review and meta-regression of clinical decision support uptake. Implement Sci. 2022;17(1):21. doi: 10.1186/s13012-022-01199-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kwan JL, Lo L, Ferguson J, et al. Computerised clinical decision support systems and absolute improvements in care: meta-analysis of controlled clinical trials. BMJ. 2020;370:m3216. doi: 10.1136/bmj.m3216 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Choudhry NK, Kronish IM, Vongpatanasin W, et al. ; American Heart Association Council on Hypertension; Council on Cardiovascular and Stroke Nursing; Council on Clinical Cardiology . Medication adherence and blood pressure control: a scientific statement from the American Heart Association. Hypertension. 2022;79(1):e1-e14. doi: 10.1161/HYP.0000000000000203 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Umair M, Khan M, Olivier J. Accounting for regression to the mean under the bivariate t-distribution. Stat Methods Med Res. 2024;33(9):1624-1636. doi: 10.1177/09622802241267808 [DOI] [PubMed] [Google Scholar]
- 32.Brookhart MA, Patrick AR, Schneeweiss S, et al. Physician follow-up and provider continuity are associated with long-term medication adherence: a study of the dynamics of statin use. Arch Intern Med. 2007;167(8):847-852. doi: 10.1001/archinte.167.8.847 [DOI] [PubMed] [Google Scholar]
- 33.Hamo CE, Mukhopadhyay A, Li X, et al. Association between visit frequency, continuity of care, and pharmacy fill adherence in heart failure patients. Am Heart J. 2024;273:53-60. doi: 10.1016/j.ahj.2024.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Faridi KF, Peterson ED, McCoy LA, Thomas L, Enriquez J, Wang TY. Timing of first postdischarge follow-up and medication adherence after acute myocardial infarction. JAMA Cardiol. 2016;1(2):147-155. doi: 10.1001/jamacardio.2016.0001 [DOI] [PubMed] [Google Scholar]
- 35.Nieuwlaat R, Wilczynski N, Navarro T, et al. Interventions for enhancing medication adherence. Cochrane Database Syst Rev. 2014;2014(11):CD000011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Zaugg V, Korb-Savoldelli V, Durieux P, Sabatier B. Providing physicians with feedback on medication adherence for people with chronic diseases taking long-term medication. Cochrane Database Syst Rev. 2018;1(1):CD012042. doi: 10.1002/14651858.CD012042.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Whelton PK, Carey RM, Aronow WS, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/ASH/ASPC/NMA/PCNA guideline for the prevention, detection, evaluation, and management of high blood pressure in adults: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Hypertension. 2018;71(6):e13-e115. [DOI] [PubMed] [Google Scholar]
- 38.Al Ani M, Garas G, Hollingshead J, Cheetham D, Athanasiou T, Patel V. Which electronic health record system should we use? a systematic review. Med Princ Pract. 2022;31(4):342-351. doi: 10.1159/000525135 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Insights H. Hospitals and health systems implementing EPIC in 2024. Accessed January 21, 2025. https://www.definitivehc.com/resources/healthcare-insights/hospitals-and-health-systems-implementing-epic
- 40.Blecker S, Gannon M, De Leon S, et al. Practice facilitation for scale up of clinical decision support for hypertension management: study protocol for a cluster randomized control trial. Contemp Clin Trials. 2023;129:107177. doi: 10.1016/j.cct.2023.107177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Powers BJ, Olsen MK, Smith VA, Woolson RF, Bosworth HB, Oddone EZ. Measuring blood pressure for decision making and quality reporting: where and how many measures? Ann Intern Med. 2011;154(12):781-788, W-289-W-290. doi: 10.7326/0003-4819-154-12-201106210-00005 [DOI] [PubMed] [Google Scholar]
- 42.Mills KT, Obst KM, Shen W, et al. Comparative effectiveness of implementation strategies for blood pressure control in hypertensive patients: a systematic review and meta-analysis. Ann Intern Med. 2018;168(2):110-120. doi: 10.7326/M17-1805 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Trial Protocol
eFigure. Intervention In-Clinic Workflow
eTable 1. Mean (SD) Change in Proportion of Days Covered (PDC) From Baseline to 12 Months, in Exploratory Analysis of Patient Subgroups
eTable 2. Responses to Questionnaire From Medical Assistants (MAs)
Data Sharing Statement
