Persons receiving care in safety-net settings, with complex diseases and chronic conditions often have significant health burdens including complicated treatment regimens, limited access to health information and inadequate health information technology to support their care. Thus, applying the chronic care model (CCM) may improve the care for these patients and applying this model may offer important new research opportunities. Clinical information systems (CIS) are a key element in the CCM and have three important roles: reminder systems; feedback mechanisms; and registries. While CIS have focused on the provider as the recipient of data, we believe that a CIS might just as importantly provide data to patients (the consumers of information and healthcare). The personal health record (PHR) is a recent, increasingly common application acting as a CIS for patients. Patients can use PHRs to receive and understand information and respond to research questions, in essence self-managing their health. Adherence is critically important for patients to achieve the maximal benefits of antiretroviral medications, but an adherence intervention trial must rely on good measures of adherence to assess the efficacy of the intervention. Typical measures used to assess adherence, including self-report, pill counts and medication event monitoring systems, are limited, and a independent "biomarker" of drug adherence would be useful. There is no gold standard or even optimal method for the assessment of drug exposure; however, our group has recently shown that the strongest independent predictor for virologic response is hair concentrations of the anchor antiretroviral medication. Hair concentrations were stronger predictors of response than self-reported adherence. In this project, we will work directly with underserved patients in a safety-net setting to conduct a randomized clinical trial of standard-of-care adherence counseling versus a novel adherence intervention embedded in an existing PHR that uses cellular phone short message service (SMS) reminders and patient responses to the reminders. We will compare adherence rates in both arms as assessed by antiretroviral medication concentrations in hair samples and self-report; clinical outcomes will also be compared in the two arms. The project will examine medication adherence for antiretroviral medications and treatments of common conditions such as depression, hypertension, diabetes and hypercholesterolemia. Our hypotheses are that persons receiving private and individualized automated reminders for medication adherence from trusted personal health records (the self-management intervention) will have improved rates of adherence, contributing to better clinical outcomes.

Primary care for persons in the safety net setting and specifically for persons with HIV infection relies on complex sources of information. There are now thirty-one approved anti-HIV agents from five classes of medications, and these medications substantially reduce mortality and disease progression but have significant toxicities. Our project will develop an application to promote medication adherence using wireless technology. Specifically the project will examine the medication adherence for antiretroviral medications and common medical and mental health conditions such as depression, hypertension, diabetes and hypercholesterolemia.