Objective: To estimate the frequency and possible predictors of patient-mediated intentional alterations in antiretroviral medication regimens in direct response to symptoms associated with antiretroviral therapy use.
Design: Cross-sectional survey of a population-based dynamic cohort of antiretroviral recipients in a province-wide HIV drug treatment program, the only source of free-of-charge antiretroviral medications in the province of British Columbia.
Methods: Program participants voluntarily complete program surveys on an annual basis. Study subjects were those who responded to the annual treatment program survey between January 1 and November 1, 2001. Patients reported on the occurrence and severity of symptoms of 42 side effects of antiretroviral agents. Symptoms were classified into four subgroups based on whether they were considered subjective or objective and whether they would or would not prompt clinical action. For each of the four symptom categories, patients reported what their physician recommended in response to symptoms in that group and what the patient actually did in response to these same symptoms. Intentional nonadherence was defined as reporting either skipping or altering dosages of selective regimen components or temporary cessation of therapy that was not recommended by the physician in response to adverse drug effects in the past year.
Results: Of 638 study subjects, 70 (11%) reported intentional nonadherence with between 4% and 7.4% reporting this activity over the preceding year depending on the symptom group. Multivariate analysis revealed that a plasma viral load of <400 copies/mL (adjusted odds ratio [AOR], 0.35; 95% CI, 0.21-0.61) and completion of high school (AOR, 0.43; 95% CI, 0.24-0.78) were both inversely associated with intentional nonadherence. Those subjects reporting at least one severe symptom were more than twice as likely to report intentional nonadherence (AOR, 2.24; 95% CI, 1.16-4.33). Similarly, each additional symptom considered to be objective and to require clinical action was associated with a 25% increase in the risk of intentional nonadherence (AOR, 1.25; 95% CI, 1.10-1.43).
Conclusion: Intentional nonadherence to antiretroviral therapy is common among persons experiencing therapy-related side effects. Although the type and severity of adverse effects impact intentional nonadherence, this activity occurs in relation to symptoms regardless of their strict clinical relevance.