Research Objective: Many primary care practices have purchased electronic health records (EHRs) and accompanying patient portals. The role online portals may play in quality and outcome improvement will depend not only on who is using such technology, but how it is used. We evaluate the characteristics of patients using an online portal in comparison to those not using, and examine the portal features and functionalities accessed by users.Study Design: Observational, cohort study for which data were obtained from EHR and health system administrative data. Patient-level data (including demographic information, service use, and portal activation and use) were joined with information characterizing clinics in which patients received care (e.g., medical teaching on site, size, and urban/suburban location). The primary study outcome, portal use, was defined by the initiation of at least one online session. Among users, user-initiated clicks were used to determine specific features accessed. Logistic regression models with random effects were fit using the PROC GLIMMIX procedure (SAS software, Version 9.4) to test the role of clinic- and patient-level variables on patient portal activation. Subjects were blocked by physician, nested within clinic, and the Laplace method was used for likelihood approximation.Population Studied: Study eligible patients were aged 18 years and older with an office visit between 4/1/2013 and 3/31/2014 to a primary care physician practicing in one of the 26 primary care clinics of an integrated delivery system serving Detroit, Michigan and the surrounding suburban areas (N=20,282 patients).Principal Findings: As implemented in December 2012, the online portal enabled users to securely schedule appointments, receive appointment reminders, pay bills online, view lab and other test results, manage information about their health, and communicate with care teams via a secure messaging system. Cohort patients were on average 68.7 years of age (SD=14.7), predominately white (65%) or black (30%) race, and 60% female. Within 18 months of implementation, 33% had an activated account, with African Americans (OR=0.50, 95% CI 0.46-0.56), Hispanics (OR=0.63, 95% CI 0.47-0.84), those over aged 70 years (OR=0.48, 95% CI 0.44-0.52), and those preferring a language other than English (OR=0.43, 95% CI 0.31-0.59) less likely to be a portal user. Patients who were married (OR=0.55, 95% CI 1.44-1.67) and more connected with the clinic, as measured by visit frequency and health maintenance visit use, were more likely to be portal users (OR=1.08, 95% CI 1.05-1.10 and OR=1.39, 95% CI 1.27-1.52, respectively). Among users, the medical record access and management feature (95.9%) was most commonly accessed, most often to obtain laboratory testing results (91.7%). The majority of users also accessed appointment management (76.6%) and messaging (59.1%) functionalitiesConclusions: While the diversity of functions accessed by those with a portal account bodes well for the ability of portals to engage patients, without purposeful intervention enhancements to care delivery afforded by portals may be inaccessible to many, including racial/ethnic minorities and those less connected to traditional care services.Implications for Policy or Practice: Online portals have the potential to extend care beyond the confines of traditional office visits, but inattention to who uses portals may exacerbate known disparities in health care access and outcomes.