Abstract Book Vol. 2 [International Conference on AIDS (14th: 2002: Barcelona, Spain)]

44 Abstracts WePeB5851-WePeB5854

XIV International AIDS Conference

WePeB5851 Determinants of antiretroviral treatment adherence
among patients with HIV and AIDS in Botswana
S.D. Weiser1, W.R. Wolfe1, P. Kebaabetswe2, J. Makhema3, D. Dickenson3,
K.F. Mompati4, M.K. Sherif3, S. Tlou5, H. Moffat6, C. Onen6, K. Hensle1,
I. Thior3, R.G. Marlink1. dHarvard AIDS Institute, 931 Stanyan Street, San
Francisco, CA, 94117, United States; 2Botswana Ministry of Health, Gaborone,
Botswana; 3Botswana-Harvard Partnership, Gaborone, Botswana;
4Botswana-Harvard Partnership, Francistown, Botswana; 5University of
Botswana, Gaborone, Botswana; 6Princess Marina Hospital, Gaborone,
Botswana
Background: Botswana has one of the highest rates of HIV infection in the world,
estimated at 36% of the population aged 15-49. To improve antiretroviral (ARV)
treatment delivery, we conducted a collaborative study of the social, cultural, and
structural determinants of treatment adherence. This research is being used by
the Botswana National AIDS Program to develop strategies for implementing universal access to treatment.
Methods: We used both qualitative and quantitative research methodology, including questionnaires and interviews with 112 HIV patients on ARVs, 60 local
healthcare providers, and 16 traditional healers. To measure adherence (defined
as taking 95% of prescribed doses), we used both patient self-report and clinician
assessment. To identify factors affecting adherence, we elicited patient knowledge, attitudes and practices; use of traditional medicines; structural barriers to
treatment; and social stigma.
Results: 54% percent of patients were adherent by self-report while 53% were
adherent by physician assessment. Observed agreement between patients and
providers was only 68%. Principal barriers to adherence included financial constraints (44%), stigma (15%), travel/migration (10%), side-effects (9%), and lack
of food (7%). While 98% of patients demonstrated accurate knowledge about the
mode of transmission and prevention of HIV, 70% of providers believed that lack
of knowledge played a key role in treatment non-adherence. Eliminating cost as a
barrier by logistical regression, projected adherence increased from 54% to 74%.
Conclusions: ARV adherence rates in this study are comparable to those seen
in most developed countries. As elsewhere, health providers in Botswana were
often unable to identify which patients adhere to their ARV regimens. The cost of
antiretroviral therapy was the most significant barrier to adherence. This situation
is encouraging in Botswana, where access to antiretrovirals in the public sector is
expanding.
Presenting author: Sheri Weiser, 931 Stanyan Street, San Francisco, CA,
94117, United States, Tel.: [phone], E-mail: [email protected]
I WePeB5852 Adherence to antiretroviral therapy: comparative
analysis of protease inhibitors-based and
non-nucleoside reverse transcriptase
inhibitors-based regimens
A. Antinori1, M.P Trottal, A. Cozzi Lepri2, M. Zaccarelli1, L. Minardi3,
P. Narciso1, L. Ravasio4, R. Murri5, F. Baldelli6, P. Piano7, S. Lo Caputo8, P. De
Longis', RP. Noto1, S. Nappa9, M. Dalessandrolo, A.W. Wul1, M. Moroni4,
G. Ippolito', A. d'Arminio Monforte4, A. Ammassari5. 'NIID L. Spallanzani, INMI
L. Spallanzani, via Portuense 292, 00149, Roma, Italy; 2Royal Free Hospital,
London, United Kingdom; 3Clinic Infectious Diseases, Brescia, Italy; 4Clinic
Infectious Diseases, Milano, Italy; 5Catholic University, Roma, Italy; 6Clinic
Infectious Diseases, Perugia, Italy; 7Clinic Infectious Diseases, Cagliari, Italy;
8Department Infectious Diseases, Firenze, Italy; 9Clinic Infectious Diseases,
Napoli, Italy; 1Clinic Infectious Diseases, Chieti, Italy; 1Johns Hopkins
University Baltimore, United States
Background: Simplification by NNRTI was suggested for improving adherence
(ADH) to antiretrovirals. Nevertheless the relationship between ADH and drug
classes has not yet been studied thoroughly.
Study design: Inter-cohort analysis on two Italian observational studies on ADH
to HAART (AdlCoNA, a multicenter study within the ICoNA cohort and AdeSPALL
study). Non-ADH was defined as having missed at least one HAART dose in the
last week or reporting interruption in drug supply by self-administered questionnaire.
Results: 596 pts included (72% men, mean age 37 yrs, 36.3% IDU, 21% AIDS).
At enrolment, median HIV-RNA and CD4 count was 1.90 (range 1.30-6.06)
logl0c/ml and 490 (6-1883) cells/pl. Median time on HAART was 21 months,
11 spent on the last scheme. The current HAART regimen included single-PI in
380 and NNRTI in 216 pts. Non-ADH was reported by 274 pts. Proportions of
non-ADH pts were 50.5% in NNRTI- and 38% in P1-group. In a multivariate analysis to receive an NNRTI-regimen (OR 0.56; 95%CI 0.37-0.84), age <35 yrs (1.53;
1.05-2.22) and IDU (2.99; 1.48-6.03) were associated with non-ADH. Among 254
pts in their first HAART, only EFV-based treatment decreased the risk of nonADH (0.20; 0.06-0.64); age <35 yrs, self-reported alcohol abuse, to be too busy
and vomit were associated with poorer ADH. As non-ADH determinants, younger
age, IDU, too many pills or pills too big, to have experienced previous side ef
fects, confusion, diarrhoea, fatigue, muscle pain, vomit, abnormal fat distribution
and sexual dysfunction were correlated to non-ADH in PI group, wheares nausea, IDU, unemployment and to be remembered of HIV disease affected ADH in
NNRTI group.
Conclusions: NNRTI-regimens are related to better ADH than those containing
PI. In the first-HAART EFV-regimen was the only variable positively associated
with ADH. PI-treated pts are more likely to be non-adherent because of treatmentrelated variables concerning complexity of scheme and medication side effects.

Presenting author: Andrea Antinori, INMI L. Spallanzani, via Portuense 292,
149, Roma, Italy, Tel.: [phone], Fax: [phone], E-mail: antinori
@inmi.it
WePeB5853 A novel method for visualising individual
electronic monitoring adherence data
K. Fenniel, G. Knafl', C. Bova2, K. Dieckhaus3, A. Williams1. 1 Yale University
School of Nursing, yale university school of nursing, po box 9740, new haven,
connecticut, 06536-0740, United States; 2University of Massachusetts School of
Nursing, Worcester, MA, United States; 3University of Connecticut School of
Medicine, Farmington, CT United States
Background: Electronic monitoring device (EMD) data are rich and complex,
and may not be analysed to their full potential. Summary measures are essential
aspects of an individual's overall adherence, but adherence can vary over time.
Thus it is important to visualise the pattern of change over time. The purpose of
this study is to describe a novel method for visualising individual adherence event
rates and for modeling expected event rates. Methods and
Results: This is an analysis of EMD data from a clinical trial examining a
home-based nursing intervention. Data were adjusted to correct for missing data
and non-control of medications. Individual level data were graphed as cumulative event counts over time. A linear cumulative count curve suggests consistent medication-taking. Concave and convex curves indicate decreasing and increasing event rates respectively. Fluctuating curves are indicative of intermittent
medication-taking. Values are not readily apparent from these curves, so it is difficult to determine a slope consistent with prescribed dosing vs an incorrect one.
Therefore, data were grouped into equal-sized intervals and event rates graphed.
These intervals were compared to prescribed event rates in order to assess rates.
Finally, count data for intervals were modeled using Poisson regression to describe the pattern of expected event counts over time, then converted to event
rates for comparison to prescribed rates. Heuristics were used to adaptively select sets of power transforms of time with which to model event count data.
Conclusions: These methods are useful for obtaining insight into adherence
through estimation of expected EMD event rates, comparison to prescribed rates,
identification of variability, and subjective classification of pattern types. These
methods can be helpful in assessing whether subjects are at risk for maximizing
viral resistance to antiretrovirals.
Presenting author: Kristopher Fennie, yale university school of nursing, po box
9740, new haven, connecticut, 06536-0740, United States, Tel.: [phone],
Fax: [phone], E-mail: [email protected]
WePeB5854I Adherence as a gender issue: psychosocial
factors influencing antiretroviral adherence in
HIV-seropositive women
F Starace1, N. Abrescial, G. Angarano2, L. Cafaro1, G. Guaraldi3, S. Lo
Caputo4, R. Maserati5, A. Massa', A. Tramarin6, A. Chirianni1, L. Sherr7.
SCotugno Hospital, Consultation Psychiatry Unit, Cotugno Hospital, Via
Quagliariello, 54, 80131 Napoli, Italy; 2University of Bari, Bari, Italy; 3Policlinico
di Modena, Modena, Italy; 4S. Maria Annunziata Hospital, Firenze, Italy;
5 Policlinico S. Matteo, Pavia, Italy; 6S. Bartolo Hospital, Vicenza, Italy; 7Royal
Free & University College, London, United Kingdom
Background: In recent years, adherence to antiretroviral (ARV) therapy in HIVseropositive persons has been explored. However, major gaps still exist in our
knowledge of the psychosocial and behavioural aspects of non-adherence (NA)
in women.
Methods: A nation-wide multicentre Italian study has been carried out to assess
psychosocial variables influencing ARV medication adherence in HIV-seropositive
women utilising the Italian version of the ACTG adherence questionnaire. NA was
defined as either taking less than 95% of the prescribed doses in the last four days
or not following the specific schedule or special instructions most of the time, in
the same period.
Results: 176 women (mean age: 34.6 ~ 6.1) have been recruited after providing their informed consent. 43.7% of the sample had a diagnosis of AIDS.
The average number of pills taken daily was 9.8 ~ 4.5. 29% of women were
NA to the prescribed ARV regime in the four days prior to interview. NA was
significantly associated with a number of psychological parameters rather than
logistic, regimen, employment or demographic factors. Prominent depressive
symptomatology and higher levels of perceived stress (OR=2.1; 95%CI=1.1-4.2;
and OR=2.5; 95%C1=1.2-5.0, respectively) were associated with non-adherence.
These women showed lower self-efficacy (OR=5.7; 95%CI=2.7-12.0), uncertainty
about positive effects of ARV therapy (OR=4.3; 95%CI=2.1-8.7), and disbelief
non-adherence related resistance (OR=2.5; 95%C1=1.3-4.8). Alcohol misuse was
significantly associated to NA (OR=5.6; 95%CI=1.7-18.5).
Conclusions: Psychosocial factors significantly influence adherence to ARV
therapy in HIV-seropositive women. Special attention should be paid to the early
recognition and proper treatment of alcohol, depressive and anxiety symptoms

in women. Gender differences must be explored and interventions should be tailored to women's needs.
Presenting author: Fabrizio Starace, Consultation Psychiatry Unit, Cotugno
Hospital, Via Quagliariello, 54, 80131 Napoli, Italy, Tel.: [phone], Fax:
[phone], E-mail: [email protected]