This project tests a model of chronic disease medication management in which the decision to initiate or adjust medical therapy is directly linked to a sequence of subsequent clinical actions (e.g. monitoring for adverse drug events, assessing response to therapy, changing medication dose) performed independently of the office visit. The investigators hypothesize that establishing a visit-independent, health information technology (IT) supported cycle of laboratory monitoring and iterative medication dose adjustment will result in more effective chronic disease care.

To implement a new model of chronic disease management, we will build on our existing electronic health record and integrated data systems to develop an advanced health IT application called the "Medication Metronome" to enable providers to schedule future laboratory tests related to a specific set of medications (for glycemic, cholesterol, and blood pressure management). As these lab test dates become due, the Medication Metronome system will remind patients via letter and inform providers when the tests remain "missing." The goal of this intervention is to implement an efficient, visit-independent system to ensure that patients are rapidly and safely brought to evidence-based treatment goals and to prevent delays in planned laboratory monitoring. This study has the following aims:

Aim 1: To develop the Medication Metronome system. This work involves health IT development and evaluation of design prototypes to create a system that supports timely medication intensification, improves safety, and meets both patient and provider needs.

Aim 2: To conduct a randomized controlled trial of the Medication Metronome system. We will use three target chronic conditions to test different elements of the system. We hypothesize that use of the Medication Metronome system will lead to:

H2a. More effective HbA1c control among patients with type 2 diabetes prescribed hypoglycemic medicines; H2b. Safer medication management among patients with hypertension prescribed thiazide diuretics, angiotensin converting enzyme inhibitors, or angiotensin II receptor blockers; H2c. Both more effective LDL-cholesterol control and safer monitoring for hepatitis among patients with hyperlipidemia prescribed HMG-CoA reductase inhibitors.

Aim 3: To evaluate the impact of the Medication Metronome visit-independent care model on the content of office-based visits. Time spent addressing different clinical care domains will be assessed using audiotape-based content analysis in a subset of selected office visits.

Summary: We will implement, and rigorously evaluate a health IT-supported model of visit-independent medication management designed to enable safer and more effective chronic disease care. We will also carefully investigate the impact of this system on primary care visits. The broader goal of this work is to support health delivery redesign that fosters patient-centered primary care by combining visit-independent medication management with more productive visit-based patient-provider interactions.

Inclusion Criteria:

• All primary care physicians from participating practices will be eligible to participate in the study.

  • Patients Eligible for Analysis: The primary unit of analysis will be prescribed medicine, grouped with patient and within prescribing PCP. Three potentially overlapping medication-based cohorts will be defined: 1) Patients prescribed any hypoglycemic agents, 2) Patients prescribed thiazide diuretics, ACE-Is, or ARBs, and 3) Patients prescribed statins. Based on this design, individual patients may contribute to more than one medication analytic cohort.

Exclusion Criteria:

• Patients Excluded from Analysis: Patients who are subsequently identified as having died during the course of the study intervention using the Social Security Death Index, to have left the MGH system, or to have changed PCPs.