Comparison of Population Pharmacokinetics Based on Steady-State Assumption Versus Electronically Monitored Adherence to Lopinavir, Atazanavir, Efavirenz, and Etravirine: A Retrospective Study.

BACKGROUND: Population pharmacokinetic (PopPK) analyses often rely on steady state and full adherence to prescribed dosage regimen assumptions from data gathered during therapeutic drug monitoring (TDM). Nonadherence is common in chronic diseases such as HIV. This study evaluates the impact of adherence measurement by electronic monitoring on PopPK parameter estimation and individual concentration profile predictions, and also the influence of adherence issues on the clinical interpretation of a concentration measurement. METHODS: Published PopPK models for lopinavir, atazanavir, efavirenz, and etravirine were applied to estimate PK parameters and individual concentrations in 140 HIV patients taking part in a medication adherence program using 2 dosing data sets. The first set included the last dose reported by the patient with steady-state and full adherence assumptions; the second set used detailed electronic dosing history. PopPK parameter estimates and individual predictions were compared between the 2 dosing entries. RESULTS: Clearance estimates and likewise predicted concentrations did not markedly differ between the 2 dosing histories. However, certain patterns of nonadherence such as sparse missed doses or consecutive missed doses lead to suboptimal drug exposure. The interpretation based on self-reported information would have concluded on a wrongly appropriate individual exposure. CONCLUSIONS: PopPK analysis assuming steady state with full adherence produced similar results to those based on detailed electronic dosing history reconciled with patients' allegations. Self-reported last dose intake appeared reliable for concentration predictions and therapeutic drug monitoring interpretation for most patients followed at the medication adherence program. Yet, clinicians should be aware that concentration predictions based on self-reported last dose intake might be overestimated in case of undetected patterns of nonadherence, increasing the risk of forthcoming therapeutic failure.

Development of an algorithm for analysing the electronic measurement of medication adherence in routine HIV care.

Background Medication adherence is crucial for successful treatment. Various methods exist for measuring adherence, including electronic drug monitoring, pharmacy refills, pill count, and interviews. These methods are not equivalent, and no method can be considered as the gold standard. A combination of methods is therefore recommended. Objective To develop an algorithm for the management of routinely collected adherence data and to compare persistence and implementation curves using post-algorithm data (reconciled data) versus raw electronic drug monitoring data. Setting A community pharmacy located within a university medical outpatient clinic in Lausanne, Switzerland. Methods The algorithm was developed to take advantage of the strengths of each available adherence measurement method, with electronic drug monitoring as a cornerstone to capture the dynamics of patient behaviour, pill count as a complementary objective method to detect any discrepancy between the number of openings measured by electronic monitoring and the number of pills ingested per opening, and annotated interviews to interpret the discrepancy. The algorithm was tested using data from patients taking lopinavir/r and having participated in an adherence-enhancing programme for more than 3 months. Main outcome measure Adherence was calculated as the percentage of persistent patients (persistence) and the proportion of days with correct dosing over time (implementation) from inclusion to the end of the median follow-up period. Results A 10-step algorithm was established. Among 2041 analysed inter-visit periods, 496 (24 %) were classified as inaccurate, among which 372 (75 %) could be reconciled. The average implementation values were 85 % (raw data) and 91 % (reconciled data) (p < 0.0001). At day 544, persistence values were 68 % (raw) and 82 % (reconciled) (p = 0.11), and adherence values were 74 % (raw) and 82 % (reconciled) (p < 0.0001). Conclusion Combining electronic drug monitoring, pill count and patient interviews is possible within the setting of a medication adherence clinic. Electronic drug monitoring underestimates medication adherence, affecting subsequent analysis of routinely collected adherence data. To ensure a set of reliable electronic drug monitoring data, structured and timely electronic drug monitoring management should be reinforced.

Association of disclosure of HIV status with medication adherence.

OBJECTIVE: Disclosure may affect adherence to antiretroviral treatment. In a medication adherence program, this cross-sectional study describes disclosure, perceived reaction after disclosure, living situations, and the relationship of disclosure with antiretroviral adherence. METHODS: A combination of a questionnaire to measure disclosure and longitudinal electronic monitoring of medication adherence was used. RESULTS: A total of 103 out of 159 eligible patients gave informed consent. The characteristics differed between participants and nonparticipants (race, education, sexual orientation, medication adherence). Thirteen participants did not disclose their HIV status. Seventy-three (81%) participants judged the reaction after disclosure positive. Among the 62 participants cohabiting, 52% disclosed to all co-residents. Adherence was high (median 100%). HIV disclosure was negatively associated with adherence, when disclosing to the mother (OR=2.46, p-value=0.086) and to siblings (OR=2.89, p-value=0.029). Living alone was associated to a lower adherence than cohabitation (Rate Ratio=1.42, p-value=0.007). CONCLUSION: HIV disclosure and adherence are sensitive issues, which may explain the reason for refusal. Nonparticipants may be those with the most difficulties disclosing. PRACTICE IMPLICATIONS: An unbiased collection of sensitive information, as HIV disclosure, is a difficult task. A cohort design, with research data collected systematically by a trusted healthcare provider, may better describe the association between adherence and disclosure.