The aim of this study was to use electronic health care records (EHRs) to examine retrospectively the incidence of and attributes associated with dental implant failures necessitating implant removal in a large cohort of patients treated in the student clinics of a U.S. dental school over three and a half years. EHRs were searched for all patients who received dental implants between July 1, 2011, and December 31, 2014. Characteristics of patients and implants that were actively removed due to irrevocable failure of any etiology ("failure cohort") during this period were compared to those of all other patients who received dental implants during the same time frame ("reference cohort"). Differences in the frequency distribution of various characteristics between the failure and reference cohorts were compared. Of a total 6,129 implants placed in 2,127 patients during the study period, 179 implants (2.9%) in 120 patients (5.6%) were removed. In the multivariate analysis, presence of a removable (OR=2.86) or fixed temporary prosthesis (OR=3.71) was statistically significantly associated with increased risk for implant failure. In contrast, antibiotic coverage (pre- and post-surgery OR=0.16; post-surgery only OR=0.38) and implants of certain manufacturers were associated with lower risk of implant failure. In this sizeable cohort of patients receiving care in dental student clinics, the review of EHRs facilitated identification of multiple variables associated with implant failure resulting in removal; however, these findings do not suggest causative relationships. The adopted analytical approach can enhance quality assurance measures and may contribute to the identification of true risk factors for dental implant failure.