Important note: Information in this article was accurate in November 2002. The state of the art may have changed since the publication date.
Investigators and clinicians continue to be interested in predicting, measuring, and improving adherence to highly active antiretroviral therapy (HAART). However, new information was sparse at the International AIDS conference in Barcelona in July, despite the presentation of over 60 abstracts on the subject. Many definitions of adherence were used, without clear agreement, as evidenced by definitions based upon number of doses missed within a specified time period, appointments missed, self-report, caregiver reports, pill counts, electronic monitoring devices, and pharmacy refill records.
Several groups looked at illicit drug and alcohol use as a predictor of adherence with conflicting results. Sharpe and colleagues interviewed 1655 African American women who attended health department clinics in Georgia between July 1997 and December 2000 [Abstract WePeB5817]. They found that women who used crack or other illicit substances were less likely to take medications as prescribed (OR 0.37 and 0.47, respectively) than non-users. Ekstrand and associates interviewed 200 people (90% men, 54% Caucasian, 28% African American, 12% Latino) who were on antiretroviral therapy, had detectable viral loads, and admitted to some recent use of alcohol or illicit drugs [Abstract WePeB5841]. In a multivariate analysis, those who were non-adherent to prescribed regimens had more reasons for missing doses, were depressed, had less structure to their daily lives, felt more helpless/overwhelmed, and felt less able to manage side effects. In contrast to Sharpe's study, adherence was not associated with alcohol or illicit drug use. Hinkin and colleagues also evaluated the impact of alcohol, illicit drug use, and neuro-psychiatric factors on non-adherence [Abstract WePeB5828]. Medication Event Monitoring System (MEMS) caps were used for 1 month in 142 patients. Patients >50 years of age were 4.6 times more likely to be good adherers (>90% adherence). Those who had neuropsychological impairment or neuropsychiatric dysfunction were more likely to be non-adherent (2.5 times and 4 times more likely, respectively). Current drug use, but not alcohol abuse, was strongly associated with non-adherence.
Durvasula and others looked at gender differences between 130 men and 32 women [Abstract WePeB5823]. Mean adherence scores by MEMS caps did not differ by gender (79.6% for men versus 71.6% for women); however, the reasons for non-adherence were different. Barriers to adherence for men were mainly cognitive, such as higher levels of efficacy and intent to follow medication recommendations, while women tended to have more practical reasons for non-adherence (care-giving duties and fewer economic resources).
Research on HAART adherence in adolescents has been limited. Hosek and colleagues interviewed 25 males and 17 females, ages 16-25 years [Abstract WePeB5826]. No gender differences were found, but depression/anxiety and age of first marijuana use were found to be predictors of non-adherence. The most important predictors of adherence included their belief in whether the drugs would work, side effects, number of pills, and number of doses. The most frequent reasons for missing doses were forgetting, sleeping through the dosing time, and not having medications at the time of the prescribed dose.
We continue to explore the relationship between clinician and patient and whether we, as healthcare providers, can successfully predict who will be adherent. Mannheimer and colleagues compared provider-estimated adherence to patient self-report [Abstract WePeB5845]. Agreement between the two was low (Kappa=0.11). They did find that clinicians' estimates of non-adherence were primarily influenced by laboratory markers (viral load, CD4 cell count) and that patients associated non-adherence with depression and a history of psychiatric illness. Spire and colleagues also investigated the clinician-patient relationship and found a similar level of agreement (Kappa=0.25) [Abstract WePeB5827]. When adherence was estimated to be low by either self-report or by the physician, blood levels of protease inhibitors were undetectable in 32% and 31% of patients, respectively. Similarly, for those who were deemed to be highly adherent by self-report or by physician, viral loads were undetectable in 75% of patients in both groups.
With so many different regimens now available, particularly for naïve or minimally experienced patients, there is a movement toward once daily regimens that are well tolerated and effective.
Roca and colleagues compared adherence of patients taking ddI (chewable buffered tabs) twice daily to adherence in the same patients after switching to the enteric coated (EC) formulation given once daily. [Abstract WePeB5815]. Patients were considered adherent if they kept their appointments, reported taking >80% of doses, and their HIV RNA levels were at least 1.5 log10 c/mL below baseline. Data from visits taking place during twice daily dosing (-6, -3, and 0 months) were pooled and compared to pooled once daily visits (+3, +6, and +9 months). Pooled data showed significant improvement in adherence after switch to once daily therapy compared to adherence on twice daily therapy. HIV RNA was also lower with once daily therapy. No difference was observed in CD4 cell counts.
Fumaz and a group from Spain evaluated quality of life (QOL) and adherence when HAART was switched to a once daily regimen of nevirapine/ddI/tenofovir [Abstract WePeB5834]. In this prospective study of 58 patients with baseline viral suppression, 30 were changed to the once daily regimen and 28 remained on their previous combinations. QOL measured by MOS-HIV and adherence measured by self-report were evaluated at baseline and week 12. By week 12, QOL was better in the once daily versus the twice daily group; however, adherence did not differ significantly between groups.
Knobel and colleagues looked at the impact of simplification of therapy by switching to AZT/3TC/ABC (Trizivir) in patients who were previously deemed severely non-adherent by self-report (<50% of doses taken in the previous 3 months or complete withdrawal from therapy) and who had viral loads in excess of 5000 c/mL [Abstract WePeB5829]. In an intention-to-treat analysis, 44% of patients achieved a viral load of <500 c/mL (66.5% by on treatment analysis). The mean viral load change was -1.6 log10 c/mL, mean CD4 count change was +124 cells/mm3, mortality was 6%, and 40% were lost to follow-up. At the end of the year of follow-up, 50% of patients reported >90% adherence to the simplified regimen of one pill twice a day.
Goelz and colleagues performed a retro-spective analysis, comparing durability of 3 categories of regimens for initial therapy: 2 NRTIs + 1-2 PIs, 2 NRTIs + 1 NNRTI, and 3 NRTIs (99% of which were AZT/3TC/ABC) [Abstract WePeB5838]. After 52 weeks, 71% of those on 3 NRTIs, 50% on NNRTI-based regimens, and 21% of those on PI-based regimens were still taking their initial regimen. Non-adherence caused a treatment change in 12% of those on PIs, 6% on NNRTIs, and 2% on NRTIs.
Antinori and colleagues reported the findings of an inter-cohort analysis of 2 Italian observational adherence studies comparing NNRTI regimens to PI-based regimens [Abstract WePeB5852]. Non-adherence was defined as having missed >1 dose in the last week or having an interruption of drug supply by self-report. In a multivariate analysis, those receiving an NNRTI regimen had decreased odds of non-adherence (OR 0.56), and those who were <35 years of age or injection drug users had increased odds of non-adherence (OR 1.53 and 2.99, respectively). For treatment-naïve patients, only efavirenz-based regimens decreased the risk of non-adherence (OR 0.20).
Two groups looked at the durability of regimens as a function of adherence. Dezii and colleagues evaluated mean adherence rates to NRTIs over 3 years, based on California Medicaid pharmacy claims for 2369 patients initiating therapy between July 1994 and June 2000 [Abstract WePeB5842]. Their data suggested that those who were poor adherers (<60%) at the start of therapy did not change their degree of adherence over time. However, for those who initially had better adherence (>60%), it waned over the 3 year period: a mean adherence of 70% in year one declined to 59% by year 3, and those who started with 91% adherence experienced a decline to 74% by year 3. Spire and associates reported on the dynamics of adherence, also over a 3-year period, for 360 patients who took part in the French APROCO cohort and who started a PI-containing regimen [Abstract ThPpB2137]. Initial adherence was measured at month 4 and over the course of month 12 through month 36. Those who were considered highly adherent (100% of doses taken) decreased over time (59.2% at month 4 versus 54.2% at month 36). Being highly adherent versus moderately adherent versus non-adherent at month 4 was significantly associated with virologic success (64.8%, 42.2%, and 35.3%, respectively), defined as an undetectable viral load at months 28 and 36 (the undetectable level was not defined) and immunologic success (60.1%, 55.2%, 38.7%, respectively), defined as a >200 cell increase in CD4 count at month 36.
Szabo and colleagues reported findings from a 5-stage, multi-disciplinary intensive adherence program (IAP) [Abstract WePeB5810]. Stage I consisted of a readiness assessment performed by a social worker, intended to identify potential barriers to adherence. In Stage II the IAP team formulated a plan, and the pharmacist reviewed an intensive medication plan with the patient. Stage III involved initiation of HAART. In Stage IV, patients received phone calls and adherence visits with the pharmacist during the first 2 weeks of therapy, along with a visit with their care provider at week 4. Stage V involved the chronic maintenance phase of follow-up, consisting of phone calls, pharmacist and care provider visits, and questionnaires. Between December 1999 and September 2001, 23 patients (14 antiretroviral-naïve) were screened for IAP. Multiple barriers were identified in 57% of patients. At 6 months, 53.8% of patients achieved a viral load of <50 c/mL. After a mean of 10.2 months on therapy, there was a mean increase of 76 CD4 cells and viral load decreased from a mean of 244,000 c/mL to 8800 c/mL. Only 38.5% of patients had viral loads <50 c/mL (61.5% were <400 c/mL).
Carmona reported a subgroup analysis of a prospective, multicenter study on improved adherence in severely non-adherent patients (defined as <70% adherence by self-report) who received ongoing individualized counseling with a treatment adherence counselor (TAC) [Abstract WeOrF1365]. TAC sessions were performed at baseline and at 3, 6, and 12 months after starting therapy by either a nurse or pharmacist, in addition to adherence counseling by a physician (standard of care). At month 12, adherence was improved to >90% in 76% of those who received TAC as opposed to 56% of those who received standard of care. For those who received TAC, 45% had a viral load <500 c/mL vs 33% of those who received standard of care.
Caprio and colleagues described the impact of a multi-disciplinary adherence team (ACT) on improving adherence among indigent women in Houston [Abstract WePeB5857]. ACT consisted of a nurse, pharmacist, HIV-infected peer counselor, and a social worker/drug use specialist. Antiretroviral-experienced patients (57%) had a mean baseline adherence rate of 67%, which improved to 90% and 92% at weeks 4 and 8, respectively. Antiretroviral-naïve patients had 98% and 100% adherence at weeks 4 and 8, respectively. After 8 weeks, 73% of patients had >1.5 log10 decline in viral load.
Bentz and colleagues evaluated the effect of counseling by a trained adherence nurse [Abstract WePeB5867]. This prospective, randomized controlled trial compared 123 people in the intervention group to 121 in a control group. Patients in the intervention group received 3 counseling sessions with a nurse. Adherence was significantly higher in the intervention group (75% vs 61%) at 6 months. In addition, there was a significant difference in the mean decrease in viral load in the intervention group versus the control group.
Ashraf and associates described an educational pilot project, "Project Adherence," which included a team of educators (a nurse, pharmacist, and social worker), all experienced in the care of HIV-infected patients [Abstract WePeB5873]. The 134 patients enrolled in this program had previously failed treatment and had a history of poor compliance to clinic appointments. After 6 weeks of adherence education, 71% had an improvement in virologic suppression, 35% had viral loads <400 c/mL, and 10% had >1 log10 decline in viral load.
Myers and colleagues reported their experience with a short-term directly observed therapy (DOT) intervention for patients with refractory non-adherence [Abstract WePeB5868]. Patients were referred to the "Success Through Anti-Retroviral Therapy" (START) DOT program for a 4-6 week admission to a skilled nursing facility in which volunteers, social workers, and medical staff supervised therapy. HIV education and support groups were also utilized during the stay. After discharge, 85 patients were followed for a mean of 24 months and 75% maintained adherence to therapy by self-report and care providers. At 24 months, 48% maintained a viral load of <1000 c/mL and 19% sustained a viral load of <50 c/mL.
Bova and colleagues questioned whether electronic monitoring devices (EMD) should be considered the reference standard for measurement of adherence [Abstract WePeB5821]. They described an exploratory study of EMD in 172 patients who were enrolled in a 15-month randomized controlled trial and who were treated with 3 or more drugs. Interviews were conducted throughout the study period and a structured questionnaire was administered at month 12. Thirty-six percent of patients admitted to not using the EMD consistently; 41% took out more than 1 dose at a time; 26% reported opening the bottle but not taking the medication; 29% said use of a pillbox was a major barrier to EMD use; and social problems accounted for 11% of all reported problems with EMD use.
In contrast, Deschamps and colleagues compared electronic event monitoring (EEM), which they referred to as the "gold standard," to patient self-report and physician assessment of adherence [Abstract ThPpB2136]. When asked about the effect EEM had on their pill-taking habits, 26% felt EEM had a positive effect, 14% reported a negative effect, and 60% felt it had no effect. Many patients (84%) complained about the practicality of using EEM. Comparing rates of adherence according to the 3 measures, non-adherence was observed in 40% by EEM, 5% by self-report, and 28% by physician assessment. When compared to EEM, self-report had an accuracy of 60%, and physicians were accurate 56% of the time.
Safren and colleagues randomized 44 patients who were <90% adherent by MEMS caps in the 2 weeks prior to randomization to compare continuation of MEMS alone versus the addition of reminder interventions via pager (MediMOM) over a 12-week period [Abstract WePeB5864]. Patients in the MediMOM group showed greater improvement in adherence by MEMS caps (55% at baseline versus 64% at week 12) than those who only continued with MEMS caps alone (57% versus 52%, respectively).
Research efforts continue to explore methods to improve adherence, through multi-disciplinary programs, readiness assessments, and simplification of regimens (either through decreased pill burden, less frequent dosing, or better tolerability). Our ability to predict adherence remains poor. Patients who abuse drugs or alcohol appear more likely to be non-adherent in many, but not in all studies, and other baseline characteristics appear to have little predictive value. Perhaps the best way to improve adherence is with the use of multi-disciplinary programs that provide continuous reminders from multiple members of the healthcare team. The interventions should be maintained throughout the course of therapy, not just during the initial period, as numerous studies demonstrate that adherence is improved as long as an intervention is in place, but that it wanes over time. Additional tools may be helpful to enhance adherence, such as pillboxes or electronic devices that remind patients when a dose is due. The use of electronic monitoring devices to measure or improve adherence remains controversial, as they may underestimate adherence, and they may interfere with other behavioral interventions, such as the use of pillboxes. The data suggest that more than one method of adherence measurement should be used when evaluating how well patients take their medications.
20021101
JH20021104
©1997,1998,1999,2000,2001, 2002. The Johns Hopkins University AIDS Service, Division of Infectious Diseases. Permission to use and reproduce portions of this newsletter is hereby granted provided that author and publication are fully credited and both copyright and permission notice appear with reprinted material. Inquiries may be directed to Sharon McAvinue, Managing Editor. Website: Johns Hopkins AIDS Service.
AEGiS is made possible through unrestricted grants from Boehringer Ingelheim, iMetrikus, Inc., the National Library of Medicine, and donations from users like you. Always watch for outdated information. This article first appeared in 2002. This material is designed to support, not replace, the relationship that exists between you and your doctor.
AEGiS presents published material, reprinted with permission and neither endorses nor opposes any material. All information contained on this website, including information relating to health conditions, products, and treatments, is for informational purposes only. It is often presented in summary or aggregate form. It is not meant to be a substitute for the advice provided by your own physician or other medical professionals. Always discuss treatment options with a doctor who specializes in treating HIV.
Copyright ©1980, 2002. AEGiS. All materials appearing on AEGiS are protected by copyright as a collective work or compilation under U.S. copyright and other laws and are the property of AEGiS, or the party credited as the provider of the content.
