More than 50% of adults treated for diabetes, hypertension, or lipid disorders have suboptimal medication adherence, a prominent barrier to continued improvement in chronic disease care in the United States. Primary care providers (PCPs) often fail to identify medication nonadherence and/or have insufficient time and training to address underlying reasons for it. In this project, we propose a patient-centered and technology-driven strategy to identify patients with adherence issues and apply a team approach to help them achieve evidence-based personalized goals for glucose, blood pressure, or lipids. This intervention extends the use of a widely available clinical decision support (CDS) infrastructure to support a model of care that, for the first time outside of a fully integrated care environment, will integrate pharmacists within the primary care team. The intervention relies on a continuous health informatics loop to do the following: (a) identify high-risk patients with adherence problems at the point of care by expanding the capability of an electronic medical record (EMR)-linked CDS to identify poor adherence to medications; (b) establish and maintain an auto-populating up-to-date registry of patients identified for proactive pharmacist outreach; (c) implement a pharmacist outreach strategy based on an information-motivation-behavioral (IMB) framework recommended by the World Health Organization (WHO) with demonstrated ability to influence adherence across a variety of clinical applications; and (d) coordinate care and adherence information by incorporating pharmacist assessment and action plans into CDS at subsequent office encounters.

Patients are randomly allocated 1:1 through a computer-generated program to either control or intervention.

Control: All control patients will continue to receive the basic Electronic Medical Record (EMR)-linked Clinical Decision Support (CDS) tool for cardiovascular (CV) risk factor management. This CDS includes algorithmically derived identification of high CV risk patients and prioritized treatment suggestions for lipids, Blood Pressure (BP), glycemic control, weight, tobacco, and aspirin use based on distance from goal, current medications, labs, allergies, and safety considerations. The basic CDS does not include any information on medication adherence. Patients will receive basic CDS plus usual care.

Intervention: For intervention patients, the basic CDS system for CV risk-factor control is enhanced to support a team-based care model that identifies risk of non-adherence, computes adherence information and incorporates it in the CDS, creates a registry to direct proactive pharmacist outreach, and coordinates action plans. To do this, the CDS Web service will combine EMR-identified medications with Epic Medication Adherence score (e-PDC) data. Algorithms will identify poor adherence using e-PDC scores available within the Epic EMR at patient encounters. When medication adherence issues are identified, alert messages will appear on the CDS tools for patients and providers. Patients with index visits will be followed-up for 6 months as they continue to receive CDS at subsequent encounters. At 6 months, patient e-PDC as well as clinical outcomes will be reevaluated. Patients with persistent adherence issues identified are added to a registry that is used by the pharmacists to conduct outreach. Pharmacist outreach will be conducted primarily by phone, but in-person arrangements are also an option. Pharmacists conducting outreach will identify themselves to patients as part of the care team working with the primary care provider (PCP). Examples of specific action plans that may be recommended includes education, recommending lower cost alternative medications or combination medications, addressing side effects, using pill boxes, modifying pill-taking schedules and/or using reminder systems, or referrals to medication therapy management (MTM) services or health educators. Pharmacists will be guided by a script template for the phone outreach that walks them through the IMB intervention and data collection. The pharmacists involved will have full read/write access to the EMR in full compliance with HIPAA regulations. They can make medication changes through established care protocols and/or communicate with the prescriber through secure messaging or phone consultation. The date of the pharmacist outreach and actions that result from the IMB intervention will be documented in the EMR and incorporated into subsequent CDS tools and registries to reflect the new patient state.

Inclusion Criteria:

• One or more of the following clinical criteria:

A. In the 12 months prior to the index visit, most recent hemoglobin A1C ≥8% AND have one or more active non-insulin glycemic medications on their electronic health record (EHR) medication list AND a potential adherence issue for one or more of these medications based on the Epic Medication Adherence score (e-PDC <80%, moderate or high confidence).

B. Two consecutive encounters with BP values ≥140/90 mm Hg AND one or more BP medications on their EHR medication list AND a potential adherence issue identified (e-PDC <80%, moderate/high confidence).

C. Meet the American College of Cardiology/American Heart Association (ACC/AHA) criteria listed below for moderate or high-intensity statin use AND a statin medication on their EHR medication list AND a potential statin adherence issue identified (e-PDC <80%, moderate/high confidence):

1. Age >21 with atherosclerotic cardiovascular disease (ASCVD) identified by a cardiovascular disease (CVD) diagnosis on the problem list or two or more International Classification of Diseases (ICD)-10 diagnostic codes in the last 2 years
2. Age >21 and LDL >190 mg/dL
3. Aged 40 to 75 AND diabetes identified by the diagnosis on the problem list or two or more ICD-10 diagnostic codes in the last 2 years
4. Aged 40 to 75 with 10-year CV Risk Score >7.5% based on the ACC/AHA 10-year ASCVD risk equation.

Exclusion Criteria:

An individual who meets any of the following criteria will be excluded from the study analysis:

1. Patients enrolled in hospice,
2. Patients with active cancer or undergoing chemotherapy
3. Patients with pregnancy in the last year
4. Patients without HealthPartners insurance coverage for at least 11 of the 12 months before the index visit will be excluded from cost analysis.
5. For Statin cohort, ≥1 LDL result <100 mg/dl within 2 years

Safety monitoring started at a patients index date and continued for an average of 11.6 months after index date.

The number of participants at Risk (safety analysis population) includes participants that were screened for eligibility and determined not eligible.

Because this study was conducted at a single, relatively high-performing medical group, generalizability of results to other care delivery systems or patient populations is uncertain.