Adherence Abstracts

Adherence to Combination Therapy in AIDS Clinical Trials (1997)

Chesney, M., Ickovics J., for the Recruitment, Adherence and Retention Committee of the ACTG (1997). Presented at the Annual Meeting of the AIDS Clinical Trials Group, July 1997,Washington, D.C.

The Recruitment, Adherence and Retention Subcommittee of the AIDS Clinical Trials Group administered two questionnaires to 76 patients on combination therapy from 10 clinic AIDS Clinical Trials Units during May and June of 1997 (results were presented at the July, 1997 ACTG meeting by Drs. Margaret Chesney and Jeannette Ickovics). Eighty percent of the respondents were male, 30% were persons of color, the mean age was 40 years, 41 % were college graduates and the mean income was US$ 25,000. Of the 76 patients, 41% reported missing at least one dose “yesterday” (i.e., the day before completing the survey). Fourteen percent reported missing at least one dose the “day before yesterday.” When these two days were examined together, a total of 18% of the patients missed at least one dose in the last two days. When asked about the last two weeks, 36% reported missing at least one dose. These data probably underestimate the problem because most of these patients were relatively new to their regimens. Adherence research indicates that adherence is better early on in the course of treatment and declines with time.

The report of the ACTG survey also provided preliminary findings on some of the variables that are associated with or ‘predict’ nonadherence. These variables are important because they suggest ways that individuals who may have difficulties with adherence could be identified. The intent of studies finding such “predictors” is not to characterize persons who might not be prescribed medication but rather, to identify persons who may need additional assistance and to provide information that could be used to maximize the effectiveness of the assistance.

The ACTG survey identified two predictors of nonadherence. The first of these was the frequency of alcohol intake, with a higher frequency associated with skipped doses. The average number of drinks per month among those who did not report skipping medication was 9, whereas the average number of drinks per month among those who reported skipping medication was 17. The second variable significantly associated with non adherence was “working outside the home for pay’ ” Specifically, 59% of the adherent survey respondents worked outside the home, the prevalence of working outside the home was significantly higher, at 85%, among those who are nonadherent. This latter variable is consistent with the data indicating that among the reasons for missing medications is being away from home and busy with other daily activities.

A primary purpose of this survey was to test the feasibility of the two instruments: the baseline and the adherence follow-up questionnaires. The questionnaires took an average of 10 minutes each to complete and 89% and 93% thought the lengths of each (respectively) were fine. Ninety-six and 99% of the patients said that they thought others would be willing to complete the two instruments, respectively.

Feb 01, 1998

Adherence and Effectiveness of Protease Inhibitors in Clinical Practice

Abstract of Presentation from the 5th Conference on Retroviruses and Opportunistic Infections February 1-5, 1998, Chicago, ILHECHT FM1, COLFAX G2, SWANSON M1, CHESNEY MA1

1University of California San Francisco, CA and 2Department of Public Health, SF CA


Adherence to protease inhibitor containing regimens for HIV infection is thought to be a important factor in determining the effectiveness of treatment, but there is limited data linking adherence to virologic outcomes.

We measured adherence to protease inhibitor (PI) regimens in a public hospital clinic setting, and determined the association between adherence and undetectable HIV viremia.


In 1-97 and 2-97 we surveyed patients at half of all clinic sessions at the San Francisco General Hospital AIDS clinic. Adherence was measured using a self-administered questionnaire that was reviewed by a trained interviewer for completeness. The questionnaire asked how many doses of protease inhibitors had been missed in each of the past 3 days. Patients were also asked if they took less pills than their doctor told them to take at each dose. A composite adherence measure was produced by calculating the proportion of recommended medication actually taken by patients in the prior 3 days, accounting for both missed and reduced doses.

HIV-1 plasma RNA was measured by the bDNA test (Chiron, limit of detection 500 copies/ml), using measurements requested by physicians the day of the interview or the first measure performed after the interview.


Table 1: Patient Characteristics (n=135)

Characteristic Number Percent
Male 123 91.1 %
White 90 67.2 %
African American 20 14.9 %
Latino 15 11.2 %
Median age (years) 39.8 years Range 27.0-59.5 years
HIV Risk
MSM* 93 68.9 %
IDU* 12 8.9 %
MSM/IDU 11 8.1 %
Other 7 14.1 %
Baseline CD4**
0-100 49 36.6 %
101-200 29 21.6 %
201-500 39 29.1 %
> 500 2 1.5 %
Unknown 16 11.9 %
Baseline median VL (n=61) 16060 copies/ul
Protease Inhibitor
Saquinavir 17 13.1 %
Indinavir 80 61.5 %
Ritonavir 24 18.5 %
Nelfinavir 2 1.5 %
Saquinavir/Ritonavir 7 5.4 %
Duration of PI Tx 205 days Range 60-624 days
Adherence to PI Tx
100% adherent 98 72.6 %
80-99% adherent 10 7.4 %
< 80% adherent 27 20.0 %

* MSM=Men having sex with men and IDU=Injection drug use
** Before starting treatment with protease inhibitors.

Patients: 388 patients agreed to fill out the survey (response rate 72%). Of these, 183 had taken protease inhibitors. Of the 183, 135 had taken protease inhibitors for more than 2 months at the time of the interview, and provided a medical record number to match laboratory data with the questionnaire. Overall, 41% of patients had detectable viremia.

Figure 1: Proportion of Patients with Undetectable Viremia by Adherence

Multivariate predictors of undetectable viremia: In a multiple logistic regression model controlling for CD4 count prior to beginning PI treatment, type of protease inhibitor, and whether new or changed reverse transcriptase inhibitors were started with the PI, adherence was associated with non-detectable viremia, OR=4.7, 95% CI 1.1 ñ 20.6.


(1) The proportion of patients with undetectable viremia was nearly twice as high in patients who reported taking 100% of their recommended protease inhibitor medication in a 3 day period, compared with patients taking less than 80% of medication. Adherence to protease inhibitor treatment is an important predictor of reaching undetectable viremia in clinical practice.

(2) While self-reported adherence is likely underestimate missed doses, a simple self-report measure identifies clinically important non-adherence.

(3) Nearly half our pts had detectable viremia. This is higher than reported in several clinical trials of protease inhibitor regimens, and suggest that the effectiveness of protease inhibitor regimens in clinical practice may be lower than the efficacy of these treatments established in clinical trial settings.

Frederick M. Hecht, MD
San Francisco General Hospital
995 Potrero Ave, Ward 84
San Francisco, CA, 94110
Phone:             (415) 476-4082       x.431
Fax: (415) 476-6953
[email protected]

More Details:

Author(s): Margaret Chesney
Population: HIV+
Category: Domestic
Resource Type: Reports and Monographs
Published: 1997