HIV drug resistance testing is recommended as routine part of clinical practice in HIV/AIDS treatment and care. Our objective is to assess the determinants of accessing HIV drug resistance testing and examine the factors associated with resistance testing prior to or after starting highly active antiretroviral therapy (HAART) in a setting where access to HIV care is free and universal. The Longitudinal Investigation into Supportive and Ancillary health services (LISA) study is an open prospective cohort of HIV-positive persons on HAART in British Columbia (BC), Canada. Non-clinical data were collected through an interviewer-administered survey and clinical data were obtained through the BC Centre for Excellence in HIV/AIDS Drug Treatment Program. Independent associations between key explanatory variables and resistance testing were analyzed using logistic regression. We restricted our post-HAART analyses to those patients who met the criteria for resistance testing after HAART initiation. Of 359 LISA participants who started HAART after 2000 and at a time when resistance testing was available free of charge, almost half did not receive a baseline resistance test. Post-HAART initiation, 165 of 359 study subjects met the criteria for resistance testing based on current therapeutic guidelines due to virological failure. About 37.6% of them remain untested for resistance. Multivariable analyses show that baseline testing was less likely performed for persons of Aboriginal ethnicity and more likely performed for patients initiating HAART in 2004 or after. Additionally, persons initiating HAART in 2004 or after were less likely to have received a resistance test after virologic failure. Our results show that despite existing clinical guidelines, resistance testing is underused, even in an environment where the service is available free of charge. Further, resistance testing is particularly underutilized among vulnerable populations. Urgent efforts are needed to ensure the optimal use of resistance testing at baseline and at the time of virologic failure as recommended by current guidelines.