Deschamps, Ann E MSN; Wijngaerden, Eric Van; Denhaerynck, Kris; Geest, Sabina De; Vandamme, Anne-Mieke
Department of Internal Medicine, University Hospitals of Leuven, Leuven, Belgium
Institute of Nursing Science, University of Basel, Basel, Switzerland
Rega Institute, Katholieke Universiteit Leuven, Leuven, Belgium
Centre for Health Services and Nursing Research, School of Public Health, Faculty of Medicine, University of Leuven, Leuven, Belgium
To the Editor:
This study was partly sponsored by the European Commission, DG Sanco, agreement SPC2002334, and Bristol-Myers Squibb Belgium.
Nonadherence to antiretroviral therapy jeopardizes successful clinical outcome in HIV patients.1 Assessment of nonadherence is a crucial first step in adequate behavioral management of HIV patients. Electronic monitoring (EM) has been shown to be the most sensitive and reliable measurement method of nonadherence in HIV, as demonstrated in cross-validation studies comparing EM with self-report, collateral report, prescription refill, and (unannounced) pill count.2-5 Issues with EM measurement remain in view of (1) privacy and disclosure problems; (2) lack of availability of guidelines concerning the correct use of EM, and (3) the potential presence of a hypothesized intervention effect when EM is introduced into a patient's daily life. These issues may jeopardize the reliability of EM data.6,7 It is therefore imperative that patients receive clear oral and written instructions for correct use of the EM device. Privacy and disclosure issues can be resolved by allowing patients to use a diary (admittedly an intervention) to note the date and time of actual medication intake outside of the home.6,8 Reliability can further be enhanced by assessment of adherence to the EM guidelines as part of each study protocol. Moreover, the presence of a hypothesized intervention effect when EM is introduced in a patient's daily life should be explored. This can be done by assessing a self-reported intervention effect or by using statistical techniques to explore the intervention effect in the EM data. Assuming that the intervention effect is strongest immediately at the start of the study and wanes over time, we can consider adherence behavior at the end of the study to be closer to the prestudy adherence behavior. Thus, the time during which nonadherence increases reflects the time over which the intervention effect is active. The difference between the adherence level at the beginning of the study and at the time when the intervention effect reaches its plateau can be an indirect measure of the size of the intervention effect. An adherence study in renal transplant recipients recently demonstrated that the duration of this intervention effect was 35 days (Denhaerynck k, et al, unpublished data). In our previous work,6 26% of our patients reported an intervention effect attributable to EM primarily at the start of the EM, yet its duration was not assessed.
We used data of a prospective, descriptive, single-center study statistically to explore the possible presence and duration of an intervention effect on adherence attributable to introducing EM in the daily life of HIV patients. Patients had to be adult HIV-infected persons in follow-up at the University Hospital of Leuven (Belgium); able to read and write in Dutch, French, or English; and willing to give informed consent. Patients with visual, hearing, or cognitive impairment preventing adequate communication or patients unable to manage their medication intake independently were excluded. Nonadherence was assessed by EM for 1 antiretroviral drug during 3 months. Patients were asked to note the date and time of medication intake in a diary if privacy or disclosure issues prevented them from taking the EM device outside of the house. Reliability of EM data was assessed by a structured interview at completion of the study, and a decision on the reliability of EM data was achieved using an algorithm developed for this purpose (Denhaerynck k, et al, unpublished data). More specifically, EM data were considered reliable if the patient self-reported always having used the EM device or if the EM device was sometimes not used, yet notes on times and dates of intake were available from the diary. These data were integrated in the EM data set. If the patient self-reported never or sometimes having used the EM device and there were no notes, the EM data were considered unreliable and excluded from analysis. Self-reported impact of EM use on medication intake was also assessed. Patients stated if they had experienced a neutral, positive, or negative effect. We used binary sequences of EM data in view of dose taken (1) or not (0) and doses taken at the correct time (1) or not (0). A correct timing interval was defined as ±25% range of the dosing interval. A random-intercepts logistic regression analysis, accounting for the repeated measurements structure of the data, was performed with a binary nonadherence sequence as a dependent variable, time since EM start and a patient's self-reported intervention effect as fixed variables, and a patient as a random variable. Replacing the time variable by an unspecified spline function in a fixed effect model allowed nonlinear graphic examination of nonadherence.
One hundred 49 patients were eligible, 16 refused, and 10 dropped out (7 never or only occasionally used the EM device, 2 lost the device, and 1 manipulated it during the entire measurement period). The final sample for analysis therefore consisted of 123 patients (69% male, mean age: 42.8 ± 8.9 years). Twenty-seven patients (22%) self-reported a positive intervention effect after the introduction of EM in their daily life, and 25 (20%) and 71 (58%) reported a negative or neutral effect, respectively.
The plotted nonlinear regression lines (Fig. 1) showed that the level of nonadherence in the complete sample increased over time after the start of EM (P < 0.0001) and stabilized after 40 days, suggesting an EM intervention effect on adherence of at least 40 days. Patients self-reporting an intervention effect initially had a lower nonadherence level than patients self-reporting a neutral or negative effect (P < 0.0001). Furthermore, the rise in nonadherence over time was more pronounced in patients with a self-reported intervention effect compared with patients without a self-reported intervention effect (P < 0.0001).
This study demonstrated that EM does induce an intervention effect that wanes over 40 days. Each EM study should take the presence of this intervention effect into consideration in developing studies and determining the period prevalence of nonadherence. Furthermore, it is worthwhile to consider using EM as an intervention strategy to improve adherence in HIV patients.
Ann E. Deschamps, MSN
Eric Van Wijngaerden, MD,PhD
Kris Denhaerynck, RN, PhD
Sabina De Geest
*Department of Internal Medicine University Hospitals of Leuven, Leuven Belgium
Institute of Nursing Science University of Basel, Basel Switzerland
Rega Institute, Katholieke Universiteit Leuven, Leuven Belgium
Centre for Health Services and Nursing Research, School of Public Health Faculty of Medicine, University of Leuven Leuven Belgium
1. Van Vaerenbergh K, De Geest S, Derdelinckx I, et al. A combination of poor adherence and a low baseline susceptibility score is highly predictive for HAART failure. Antivir Chem Chemother. 2002;13:231-240.
2. Paterson DL, Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med. 2000;133:21-30.
3. Arnsten JH, Demas PA, Farzadegan H, et al. Antiretroviral therapy adherence and viral suppression in HIV-infected drug users: comparison of self-report and electronic monitoring. Clin Infect Dis. 2001;33:1417-1423.
4. Liu H, Golin CE, Miller LG, et al. A comparison study of multiple measures of adherence to HIV protease inhibitors. Ann Intern Med. 2001;134:968-977.
5. Hugen PW, Langebeek N, Burger DM, et al. Assessment of adherence to HIV protease inhibitors: comparison and combination of various methods, including MEMS (electronic monitoring), patient and nurse report, and therapeutic drug monitoring. J Acquir Immune Defic Syndr. 2002;30:324-334.
6. Deschamps AE, De Graeve V, Van Wijngaerden E, et al. Prevalence and correlates of non-adherence to antiretroviral therapy in a population of HIV-patients using medication event monitoring system. AIDS Patient Care STDS. 2004;18:644-657.
7. Bova CA, Fennie KP, Knafl GJ, et al. Use of electronic monitoring devices to measure antiretroviral adherence: practical considerations. AIDS Behav. 2005;9:103-110.
8. De Geest S, Schäfer-Keller P, Denhaerynck K, et al. Supporting Medication Adherence in Renal Transplantation (SMART): a pilot study RCT to improve adherence with immunosuppressive regimen. Clin Transplant. 2006; 20(3): 3590-368.
© 2006 Lippincott Williams & Wilkins, Inc.