AIDS Behav (2006) 10:227–245DOI 10.1007/s10461-006-9078-6

ORIGINAL ARTICLE

Self-Report Measures of Antiretroviral Therapy Adherence:A Review with Recommendations for HIV Researchand Clinical ManagementJane M. Simoni · Ann E. Kurth · Cynthia R. Pearson ·David W. Pantalone · Joseph O. Merrill ·Pamela A. Frick

Published online: 3 June 2006C© Springer Science+Business Media, Inc. 2006

Abstract A review of 77 studies employing self-report mea-sures of antiretroviral adherence published 1/1996 through8/2004 revealed great variety in adherence assessment itemcontent, format, and response options. Recall periods rangedfrom 2 to 365 days (mode = 7 days). The most common cut-off for optimal adherence was 100% (21/48 studies, or 44%).In 27 of 34 recall periods (79%), self-reported adherence wasassociated with adherence as assessed with other indirectmeasures. Data from 57 of 67 recall periods (84%) indicatedself-reported adherence was significantly associated withHIV-1 RNA viral load; in 16 of 26 (62%), it was associatedwith CD4 count. Clearly, the field would benefit from itemstandardization and a priori definitions and operationaliza-tions of adherence. We conclude that even brief self-reportmeasures of antiretroviral adherence can be robust, and rec-

J. M. Simoni (�) · D. W. PantaloneDepartment of Psychology, University of Washington, Seattle,Washington 98195-1525 Box 351525e-mail: jsimoni@u.washington.edu

A. E. KurthSchool of Nursing/CFAR, University of Washington, Seattle,Washington

C. R. PearsonSchool of Public Health & Community Medicine, University ofWashington, Seattle,Washington

J. O. MerrillDepartment of Medicine, University of Washington, Seattle,Washington

P. A. FrickDepartment of Pharmacy, University of Washington, Seattle,Washington

ommend items and strategies for HIV research and clinicalmanagement.

Keywords HIV/AIDS . Antiretroviral . Medicationadherence . Self-report . Viral load

Introduction

An abundance of convergent empirical evidence has con-firmed that strict adherence to medication regimens is keyto the successful treatment of HIV infection with antiretro-viral therapy or ART (Bangsberg et al., 2000; Hogg et al.,2002; Paterson et al., 2000). However, there is decidedly lessagreement on the best strategy for assessing ART adherence.An ideal assessment instrument would be reliable, valid, andlogistically practical, with low participant and staff burden.

The search for an adherence assessment “gold standard”is not unique to the field of HIV (Geletko et al., 1996; Martinet al., 2001; Rudd, 1979; Rudd, Ahmed, Zachary, Barton,& Bonduelle, 1990; Straka, Fish, Benson, & Suh, 1997;Waterhouse, Calzone, Mele, & Brenner, 1993). Acrossmultiple clinical conditions, researchers have examined arange of methodologies for capturing medication adherence.These have been categorized as either direct or indirectmethods (Liu et al., 2001; Miller & Hays, 2000; Paterson,Potoski, & Capitano, 2002; Turner, 2002; Wutoh et al.,2003). Direct methods such as biological assays of activedrug, metabolite or other markers in blood, urine, or otherbodily fluids confirm active drug ingestion. Indirect meth-ods, which do not measure the presence of the drug in theindividual, include self-report, clinician assessment, medicalchart review, clinic attendance, behavioral observation suchas directly observed therapy, pill count (PC), pharmacy

Springer

228 AIDS Behav (2006) 10:227–245

refill (PR) records, electronic drug monitoring (EDM), andtherapeutic impact such as HIV-1 RNA viral load (VL), CD4lymphocyte count, Centers for Disease Control-definedstage of disease progression, and mortality. These assess-ment methods have advantages and disadvantages (Gao,Nau, Rosenbluth, Scott, & Woodward, 2000), with the trade-off generally assumed to be financial and logistical costversus psychometric and epidemiologic accuracy (Gordis,1979).

The present study focused on the most widely used indi-rect method of assessing ART adherence: self-report mea-sures. The practicality of self-report makes this approach alikely candidate for continued widespread use in clinical andresearch settings, including in resource-poor countries justgaining access to ART.

Patient self-report measures in the form of personal in-terviews or written questionnaires have many advantages,including low cost, minimal participant burden, ease andspeed of administration, flexibility in terms of mode ofadministration and timing of assessment, and the poten-tial to yield specific information about the timing of dosesand adherence to food requirements (Wagner & Miller,2004). Additionally, the specificity of self-report measuresis high, i.e., patients’ acknowledgment of nonadherence isgenerally credible (Bangsberg et al., 2001). Moreover, arecent meta-analysis found that despite significant studyheterogeneity, the pooled association between self-reportedART adherence and VL was statistically significant, ad-justed OR = 2.31, 95% CI = 1.99–2.68 (Nieuwkerk & Oort,2005).

On the other hand, self-report is susceptible to recallbias and inaccurate memory and potentially to social de-sirability bias; indeed, self-report does tend to produce esti-mates of adherence that are 10–20% higher than those fromEDM (Arnsten et al., 2001; Wagner & Miller, 2004). Be-cause of these limitations, some researchers have suggestedthat EDM or other less subjective methods may be prefer-able to self-report for adherence assessment in interventiontrials (Miller & Hays, 2000). Others have noted practicallimitations of EDM (Bova et al., 2005) and that adher-ence may be underestimated by EDM and overestimatedby self-report and pill count, thus warranting the use ofseveral adherence measures (Liu et al., 2001). This strat-egy, though, may be impractical for ongoing clinical use.Despite the perceived limitations, many clinicians and re-searchers alike continue to rely extensively on self-reportadherence measures, probably because they continue tobe the least costly and burdensome way to assess ARTadherence.

For the present report, we conducted a review of theliterature with the goals of identifying (a) the varietyof self-report measures used in ART adherence research,(b) the pattern of associations between self-report and other

adherence assessment strategies such as pill count and EDM,and (c) the relation between self-report and clinical indi-cators such as VL and CD4 lymphocyte count. Our aimwas to determine best practices with respect to selectingself-report measures for both research purposes and clinicalmonitoring.

Selection of studies for review

We conducted an extensive search of PsycINFO, AIDS Line,and MEDLINE for articles published in refereed journalsfrom January 1996 through August 2004 that contained somecombination of the terms (a) HIV or human immunodefi-ciency virus or AIDS or acquired immunodeficiency syn-drome and (b) adherence or compliance. Additionally, wescanned bibliographies of relevant articles and consultedwith experts in the field for other references. From the re-sulting list of over 600 articles, we selected the English-language publications describing studies of individuals atleast 18 years of age that utilized a self-report measure ofART adherence and reported its association with at least oneother adherence assessment method (such as pill count orpharmacy refill records) or with an indicator of clinical im-pact (such as VL or CD4 count). We excluded the few earlystudies examining adherence to ART monotherapy, result-ing in 77 published articles that met the a priori selectioncriteria.

Review strategy

From each article we extracted information on the study set-ting, location, and sample size; details regarding the self-report measure (including its source, number, and word-ing of items, and how adherence was operationalized foranalysis); the recall period; and the measure’s associa-tions with other adherence measures and clinical indicators.These are presented as a reference source in Table 1. Al-though not noted in the Table, we also recorded eligibil-ity criteria, sample characteristics, and study purpose anddesign.

After summarizing key descriptive information about thestudies, we focused on describing the self-report adherencemeasures in detail and use χ2 tests to assess the associationbetween self-report and other adherence measures. Our ex-amination of the reported associations between self-reportedadherence and clinical outcomes such as VL include a forestplot graph to visually summarize reported association effectsizes (Fig. 1). In a sub-analysis, we examined the effect ofrecall period length on the association between self-reportedadherence and VL using χ2 tests of proportions and logisticregression.

Springer

AIDS Behav (2006) 10:227–245 229

Tabl

e1

Stud

ies

repo

rtin

gth

eas

soci

atio

nof

self

-rep

orte

dan

tiret

rovi

rala

dher

ence

with

adhe

renc

eas

mea

sure

dby

othe

rin

dire

ctm

easu

res

orw

ithcl

inic

alin

dica

tors

Sour

ceSt

udy

Self

-rep

ortm

easu

reR

ecal

lpe

riod

Ass

ocia

tion

with

othe

rin

dire

ctad

here

nce

mea

sure

sA

ssoc

iatio

nw

ithcl

inic

alin

dica

tors

Setti

ng;l

ocat

ion;

sam

ple

size

Sour

ce/it

ems/

(non

adhe

renc

eop

erat

iona

lizat

ion)

(CO

:co

ntin

uous

,CT

:cat

egor

ical

,D

I:di

chot

omou

s)

elec

tron

icda

tam

onito

ring

;ph

arm

acy

refil

ls(P

R);

pill

coun

t(PC

);ot

her

HIV

-1R

NA

vira

lloa

d(V

L);

CD

4;ot

her

(adh

eren

tvs.

nona

dher

ent)

Alc

oba

etal

.(20

03)

2H

IVcl

inic

s;Sp

ain;

N=

106

N/R

;N/R

;DI:

<90

%of

pres

crib

eddo

ses

for

atle

ast

one

drug

4da

ysV

Lde

tect

able

(NS)

;pla

sma

indi

navi

rle

vels

(NS)

Alo

isie

tal.

(200

2)57

IDho

spita

luni

ts;I

taly

;N

=36

6N

/R;3

item

s;D

I:“Y

es”

toal

lth

ree

item

svs

.<3

6m

onth

sV

Lun

dete

ctab

le∗∗

∗68

%vs

.40%

@12

mo

Alti

ce,M

osta

shar

i,an

dFr

iedl

and

(200

1)4

pris

onH

IVcl

inic

s;C

T,U

S;N

=16

4Ic

kovi

cs’9

7;N

/R;D

I:80

%of

pills

take

n/pr

escr

ibed

7da

ysPR

r=

0.82

(sig

nific

ance

N/R

)C

D4

coun

t(N

S)

Am

mas

sari

etal

.(20

04)

11cl

inic

alce

nter

s;It

aly;

N=

135

Mur

ri’0

0;1

forc

ed-c

hoic

eite

mon

timin

gof

last

mis

sed

dose

;DI:

mis

sed

≥1

dose

over

last

7da

ys

7da

ysC

D4

high

erm

ean

(SD

)∗63

7(3

41)

vs.5

09(3

62)

Ant

inor

ieta

l.(2

004)

Stud

yco

hort

;Ita

ly;N

=23

8M

urri

’00;

N/R

;DI:

mis

sed

≥1

dose

over

last

7da

ys7

days

VL

rebo

und

>50

0co

pies

/mL

(NS)

Arn

sten

etal

.(20

01)

Hos

pita

lstu

dyco

hort

;B

ronx

,NY

,US;

N=

67N

/R;N

/R;C

O:%

ofpr

escr

ibed

dose

sta

ken

1da

y;7

days

ED

M;1

-day

r=

0.49

∗∗∗ ;

7-da

yr=

0.46

∗∗∗

VL

<50

0;1-

day

r=

0.43

∗∗∗ ;

7da

yr=

0.52

∗∗∗

Ban

gsbe

rget

al.(

2000

)C

omm

unity

coho

rt;S

anFr

anci

sco,

US;

N=

34N

/R;3

item

s;C

O:M

ean

valu

eof

3m

easu

res

of%

pres

crib

eddo

ses

3da

ysV

Lr=

−0.

60∗∗

∗

Ban

gsbe

rget

al.(

2001

)C

omm

unity

coho

rt;S

anFr

anci

sco,

US;

N=

45N

/R;D

ay-b

y-da

yre

view

ofdo

ses;

CO

:%pr

escr

ibed

dose

s,D

I:>

80%

3da

ysPC

r=

0.85

∗∗∗ ;

PCκ

=0.

65∗∗

∗ ;pr

ovid

eres

timat

e(N

S)B

angs

berg

etal

.(20

02)

Priv

ate

clin

ican

dco

unty

hosp

ital;

San

Fran

cisc

o,U

S;N

=11

0

AA

CT

G(C

hesn

ey’0

0);

Day

-by-

day

revi

ewof

dose

son

com

pute

r;D

I:90

and

80%

3da

ysPr

ovid

eres

timat

e∗∗∗

(tes

tst

atis

ticN

R)

VL

dete

ctab

le≥

500;

<80

%O

R3.

0(9

5%C

I1.

1–8.

1)

Bar

roso

etal

.(20

03)

HIV

refe

renc

ece

nter

;Rio

deJa

neir

o,B

razi

l;N

=64

N/R

;1ite

m;D

I:ta

king

aspr

escr

ibed

>80

%da

ys30

days

VL<

400;

OR

7.2

(95%

CI

1.6–

31.9

)in

sem

en;O

RA

8.2

(95%

CI

1.2–

56.7

)in

plas

ma

Bri

gido

etal

.(20

01)

Publ

icA

IDS

clin

ic;S

aoPa

ulo,

Bra

zil;

N=

168

N/R

;5ite

ms;

CT

:Reg

:all

dose

sta

ken,

qReg

:mis

sup

to4

dose

sor

1fu

llda

y/m

o,Ir

eg:a

llot

her

irre

gula

r

30da

ysV

Lm

edia

nlo

g 10∗ ;

Reg

2.0

(1.6

–5.6

);qR

eg2.

0(1

.6–5

.5);

Ireg

3.6

(1.6

–6.2

);C

D4

med

ian

gain

∗ ;(t

ests

tatis

ticN

/R);

AID

Sde

velo

pmen

tor

deat

h∗(t

ests

tatis

ticN

/R)

Car

rier

ieta

l.(2

001)

Rou

tine

clin

ical

site

s;Fr

ance

;N=

436

AA

CT

G;5

item

sfo

rea

chdr

ug;C

T:1

00%

,80–

99%

,<

80%

4da

ysV

Lun

dete

ctab

leat

4∗∗;1

2∗∗,a

nd20

mon

ths∗∗

(tes

tsta

tistic

N/R

)

Springer

230 AIDS Behav (2006) 10:227–245

Tabl

e1

Con

tinue

d

Sour

ceSt

udy

Self

-rep

ortm

easu

reR

ecal

lpe

riod

Ass

ocia

tion

with

othe

rin

dire

ctad

here

nce

mea

sure

sA

ssoc

iatio

nw

ithcl

inic

alin

dica

tors

Car

rier

ieta

l.,20

03R

outin

ecl

inic

alsi

tes;

Fran

ce;N

=36

0A

AC

TG

;5ite

ms

for

each

drug

;CT

:100

%,8

0–99

%,

<80

%

4da

ysV

Lsu

ppre

ssio

nat

3ye

ars:

Hig

hly

adhe

rent

OR

3.4

(95%

CI

1.4–

7.9)

;Mod

adhe

rent

NS;

CD

4in

crea

se>

200

by3

year

s:hi

ghly

adhe

rent

OR

2.4

(95%

CI

1.0–

5.5)

;Mod

adhe

rent

NS

Cat

z,K

elly

,Bog

art,

Ben

otsc

h,an

dM

cAul

iffe

(200

0)

Out

patie

ntID

clin

ic;

Milw

auke

e,U

S;N

=72

N/R

;2ite

ms;

CT

:mis

sed

dose

sda

ily,w

eekl

y,m

onth

ly,o

rne

ver

3m

onth

sV

L<

400∗

(tes

tsta

tistic

N/R

)

Ced

erfj

all,

Lan

gius

-Ekl

of,

Lid

man

,and

Wre

dlin

g(2

002)

Out

patie

ntH

IVcl

inic

;St

ockh

olm

,Sw

eden

;N=

99N

/R;1

,7,3

0da

ys=

%m

isse

din

1m

onth

;DI:

95%

1m

onth

VL

<50

71%

vs.4

5%∗ ;

CD

4<

200

8%vs

.32

%∗∗

∗

Cin

gola

niet

al.(

2002

)Te

rtia

ryca

reID

depa

rtm

ent;

Ital

y;N

=12

7M

urri

’00;

1ite

m,t

imin

gof

last

mis

sed

dose

;DI:

mis

sed

befo

re2–

4w

eeks

(adh

eren

t)vs

.yes

terd

ay,

last

wee

k

N/A

VL

<50

0at

3m

o∗(t

ests

tatis

ticN

R);

nona

dher

entO

R0.

37(9

5%C

I0.

1–0.

95)∗ ;

CD

4ch

ange

;3m

o+

50vs

.−

12∗∗

;6m

o+

62vs

.−

13∗∗

Coh

n,K

amm

ann,

Will

iam

s,C

urri

er,a

ndC

hesn

ey(2

002)

AA

CT

Gsi

tes;

29si

tes

inU

S;N

=64

3A

AC

TG

;2ite

ms;

DI:

100%

48hr

VL>

500

nona

dher

ence

over

56w

eeks

;OR

2.3

(95%

CI

N/R

);70

%vs

.50%

∗∗∗

Dor

zet

al.(

2003

)2

IDde

part

men

ts;P

adua

and

Ver

ona,

Ital

y;N

=10

9N

/R;1

item

;CO

:#pi

lls/#

pres

crib

ed,D

I:80

%7

days

VL

mea

n10

,854

vs.3

4,14

9∗ ;C

D4

mea

n68

99vs

.379

∗∗∗

Duo

nget

al.(

2001

)A

IDS

outp

atie

ntcl

inic

;D

ijon,

Fran

ce;N

=14

9PM

AQ

(Pat

erso

n’9

9);4

item

s;D

I:10

0%(n

onad

here

ntsc

ore<

4)

Com

bine

d:4

days

and

4w

eeks

VL

redu

ctio

n;O

RA

2.9

(95%

CI

1.2–

7.1)

;D

idno

tmis

san

yPI

last

4da

ysr

=.1

8∗

Dur

anet

al.(

2001

)St

udy

coho

rt;F

ranc

e;N

=27

7A

AC

TG

;5ite

ms;

DI:

100%

Com

bine

d:4

days

and

wee

kend

VL

unde

tect

able

4m

onth

saf

ter

AR

Tin

itiat

ion

59.4

%vs

.41.

6%∗∗

Dur

anet

al.(

2001

)St

udy

coho

rt;F

ranc

e;N

=57

N/R

;1ite

m;C

T:1

00%

,99

–80%

,<80

%1

wee

kV

Lm

edia

nlo

g 10∗∗

100%

:2.3

(2.3

–3.5

);80

%–9

9%:2

.3(2

.3–3

.6);

<80

%:3

.8(2

.6–5

.04)

;VL

unde

tect

able

∗∗10

0%:

73.1

%;8

0%–9

9%:6

9.2%

;<80

%:2

2.2%

;C

D4

(NS,

p=

0.06

);dr

ugle

vel∗∗

∗

Dur

anet

al.(

2003

)47

hosp

itals

;Fra

nce;

N=

642

N/R

;5ite

ms;

CT

:100

%,

99%

–80%

,<80

%4

days

VL

dete

ctab

le;1

00%

:OR

1.0;

99%

-80%

:O

R1.

5(9

5%C

I1.

0–2.

3);<

80%

:OR

2.3

(95%

CI

1.3–

4.1)

Eld

red,

Wu,

Cha

isso

n,an

dM

oore

(199

8)H

ospi

talH

IVcl

inic

;B

altim

ore,

MD

,USA

;N

=24

4

N/R

;1ite

mfo

rea

chtim

efr

ame;

DI:

80%

7da

ys;1

4da

ysM

edic

alre

cord

kapp

a71

%;7

day:

60%

vs.5

6%(N

S);1

4da

y:74

%vs

.67%

∗∗

Springer

AIDS Behav (2006) 10:227–245 231

Fong

etal

.(20

03)

HIV

clin

ic;H

ong

Kon

g;N

=16

1N

/R;N

umbe

rm

isse

ddo

ses;

DI:

100%

Sinc

ela

stvi

sit

VL<

500;

OR

A4.

2(9

5%C

I1.

8–12

.3)

Gao

etal

.(20

00)

3cl

inic

s;W

estV

irgi

nia,

US;

N=

72Sa

met

’92;

N/R

,ass

esse

ddo

ses;

CO

:pre

scri

bed

–m

isse

d/pr

escr

ibed

2da

ysD

isea

sese

veri

ty∗

Gar

cia

deO

lalla

etal

.(20

02)

HIV

hosp

italu

nit;

Bar

celo

na,S

pain

;N=

1219

N/R

;N/R

;DI:

90%

1m

onth

Mor

talit

y:N

onad

here

nt;R

elat

ive

haza

rd1.

5(9

5%C

I1.

2–1.

99)

Gif

ford

etal

.(20

00)

Com

mun

itypr

actic

es;S

anD

iego

CA

,US;

N=

133

CA

SQ(B

erry

’00)

4ite

ms

per

drug

take

n;C

T:1

00%

,80

–99%

,<80

%

7da

ysV

Llo

g 10

Eac

hin

crea

sein

adhe

renc

eca

tego

ryas

soci

ated

with

1.3

log 1

0

decr

ease

∗∗

Gio

rdan

o,G

uzm

an,C

lark

,C

harl

eboi

s,an

dB

angs

berg

(200

4)

Part

icip

ants

’us

ualp

lace

ofre

side

nce;

San

Fran

cisc

o,C

A,U

S;N

=84

AA

CT

Gan

dV

isua

lAna

log

Scal

e–

VA

S(W

alsh

’98)

;4ite

ms/

drug

(3da

y),V

AS-

1;C

O:M

ean

adhe

renc

eov

erth

ree

visi

ts

3da

ys(A

AC

TG

)3

or4

wee

ks(V

AS)

Una

nnou

nced

PCan

dV

AS:

r=

0.76

(95%

CI

0.65

–0.8

4);3

-day

r=

0.71

(95%

CI

0.59

–0.8

0);(

sig.

N/R

;NS

diff

betw

VA

San

d3

day)

VL

;VA

S:r=

−.4

9(9

5%C

I−

0.64

–0.3

1);

3-da

yr=

−.3

4(9

5%C

I−

0.51

–0.1

3);

(sig

.N/R

;NS

diff

betw

VA

San

d3

day)

God

in,G

agne

,and

Nac

cach

e(2

003)

4H

IVcl

inic

s;M

ontr

eal,

Que

bec

City

;N=

256

Res

earc

her-

crea

ted;

9ite

ms,

#pi

llsm

isse

d/#

pres

crib

ed;

DI:

95%

1,2,

7,30

days

VL

incr

ease

over

6m

onth

s;N

onad

here

nt1,

2,30

day

(NS)

;7da

ysO

R1.

9(9

5%C

I1.

0–3.

6)G

olin

etal

.(20

02)

3pu

blic

HIV

clin

ics;

N/R

;N

=11

7C

ompo

site

scor

ew

ithE

DM

,PC

,SR

inte

rvie

w(L

iu’0

1)1

item

CO

:#do

ses

take

n/#

pres

crib

ed

7da

ysE

DM

r=

0.38

(sig

.N/R

);PC

r=

0.62

(sig

.N/R

)

Gor

dillo

,del

Am

o,So

rian

o,an

dG

onza

lez-

Lah

oz(1

999)

HIV

refe

renc

ece

nter

;M

adri

d,Sp

ain;

N=

366

N/R

;N/R

;DI:

90%

Las

twee

kC

D4

aten

rollm

enta

ndgo

odad

here

nce;

>50

0O

RA

2.4

(95%

CI

1.3–

4.4)

;20

0–49

9O

RA

2.8

(95%

CI

1.4–

5.5)

Gou

jard

etal

.(20

03)

Hos

pita

lcen

ters

;Fra

nce;

N=

326

AA

CT

G,P

MA

Q,a

nd3

item

sre

:ins

truc

tions

(Met

calf

’98)

13ite

ms;

CO

:N

onad

here

nce

scor

e0–

26

N/R

VL

low

er(t

ests

tatis

tics

N/R

)∗∗∗

CD

4hi

gher

(tes

tsta

tistic

sN

/R)∗

Gua

rald

ieta

l.(2

003)

8te

rtia

ryce

nter

s;N

orth

ern,

Cen

tral

Ital

y;N

=17

5M

OS-

HIV

Hea

lthSu

rvey

N/R

;85%

(>1

dose

in7

days

)

7da

ysM

orph

olog

ical

tera

tions

;O

RA

2.36

(95%

CI

1.1–

5.0)

Hau

bric

het

al.(

1999

)5

univ

ersi

tyH

IVcl

inic

s;C

A,

US;

N=

164

@2

mon

ths,

119

@6

mon

ths

N/R

24ite

ms;

asse

ssed

%pr

escr

ibed

dose

sta

ken;

CT

:10

0%,9

9–95

%,<

95–8

0%,

<80

%

4w

eeks

Prov

ider

estim

ate

kapp

a=

0.02

(NS)

VL

log 1

0re

duct

ion

(SD

)@

2m

onth

s∗

100%

,99–

95%

,<95

–80%

,<80

%;0

.95

(2.2

);0.

79(2

.0);

0.57

(1.8

);0.

04(2

.0);

VL

log 1

0in

crea

se(S

D)

@6

mon

ths∗

100%

:−

1.1

(2.2

);<

80%

:0.2

(1.2

);C

D4

cells

@6

mon

ths∗∗

100%

,99–

95%

,<95

–80%

,<

80%

72(1

62);

+87

(154

);+

54(1

62);

−19

(74)

Springer

232 AIDS Behav (2006) 10:227–245

Tabl

e1

Con

tinue

d

Sour

ceSt

udy

Self

-rep

ortm

easu

reR

ecal

lpe

riod

Ass

ocia

tion

with

othe

rin

dire

ctad

here

nce

mea

sure

sA

ssoc

iatio

nw

ithcl

inic

alin

dica

tors

Ho

etal

.(20

02)

Clin

ic;H

ong

Kon

g;N

=16

1D

oung

’01

1ite

m,%

pres

crib

eddo

ses

take

nC

T:

100%

,99–

95%

,94–

90%

,<

90%

4–6

wee

ksV

Lde

tect

able

∗∗≤

99%

vs.1

00%

OR

4.2

(95%

CI

1.8–

12.3

)D

isea

sepr

ogre

ssio

n∗∗

Hor

neet

al.(

2004

)O

utpa

tient

clin

ic;B

righ

ton,

UK

;N=

109

VA

S1

item

,cor

rect

dose

timin

g;D

I:7

ptsc

ale

0–6:

cuto

ff≥

5

N/R

VL>

400;

18%

vs.2

6%(N

S);C

D4

(NS)

Hug

enet

al.(

2002

)U

nive

rsity

cent

re;N

ijmeg

enan

dA

rnhe

mN

ethe

rlan

ds;

N=

26

N/R

;VA

S;M

ultip

leite

ms;

CT

:3gr

oups

and

rang

e1–

10

N/R

ED

M%

take

non

time∗∗

∗

ρ=

.73;

%ta

ken∗∗

ρ=

.55

N/R

Icko

vics

etal

.(20

02)

21A

AC

TG

site

s;M

ultis

ites,

US;

N=

93A

AC

TG

1ite

mfo

rea

chdr

ugD

I:95

%4

days

VL>

50@

24w

eeks

<95

%ad

here

ntO

R2.

6(9

5%C

I1.

1–6.

1)C

D4

chan

ge(N

S)In

gers

oll(

2004

)U

nive

rsity

IDcl

inic

;V

irgi

nia;

N=

120

Med

icat

ion

Adh

eren

ceFo

rm(I

nger

soll

’99)

Mul

tiple

item

s;D

I:95

%PI

sta

ken;

DI:

Adh

eren

cesc

ore

1–3,

cuto

ff>

2

1w

eek

VL

unde

tect

able

77%

vs.2

3%∗

CD

4<

200

54%

vs.4

6%∗

Kim

mer

ling

etal

.(20

03)

Clin

ics,

com

mun

ity;L

osA

ngel

es,U

S;N

=58

N/R

#do

ses

take

nq.

d.of

last

3;C

O:s

core

3da

ysE

DM

:r=

0.47

∗∗∗

Subs

etre

port

ing

mis

sed

dose

s(N

S)K

leeb

erge

ret

al.(

2001

)R

esea

rch

coho

rt;B

altim

ore,

Chi

cago

,Pitt

sbur

gh,L

A,

US;

N=

393

Mod

ified

AA

CT

G;m

ultip

leite

ms;

DI:

100%

4da

ysV

Lun

dete

ctab

le<

50co

pies

;53.

3%vs

.37

.4%

∗∗C

D4

≥40

0(N

S);6

4.4%

vs.

58.3

%K

nobe

leta

l.(2

001)

Uni

vers

ityH

IVcl

inic

;B

arce

lona

,Spa

in;N

=67

9N

/R;N

/R;D

I:90

%1

mon

thV

L<

500;

Adh

eren

tOR

3.1

(95%

CI

2.2–

4.2)

∗∗∗ ;

nona

dher

entO

RA

0.4

(95%

CI

0.2–

0.7)

∗∗;C

D4

mea

nin

crea

se17

1vs

.10

7∗∗

Kno

bele

tal.

(200

1)69

hosp

itals

;Spa

in;

N=

2528

@3,

2127

@6,

and

1797

@12

mon

ths

SMA

Q(f

rom

Mor

isky

’86)

;6

item

s;D

I:95

%(m

isse

d>

2day

sin

3m

os;2

dose

s7

days

orye

sto

1/4

item

s)

Com

bine

d1 w

eeke

nd;

1w

eek;

3m

onth

s

ED

M:s

ensi

tivity

72%

;sp

ecifi

city

91%

;PPV

91%

;N

PV80

%

VL

<50

0;@

3m

onth

sO

R2.

2(9

5%C

I1.

8–2.

6)∗∗

∗ ;@

6m

onth

sO

R2.

6(9

5%C

I2.

2–3.

1)∗∗

∗ ;@

12m

onth

sO

R2.

5(9

5%C

I2.

0–3.

1)∗∗

∗ ;V

L>

500;

Non

adhe

rent

OR

A

1.7

(95%

CI

1.4–

2.1)

Kno

bele

tal.

(200

4)2

hosp

itals

;Bar

celo

na,

Spai

n;N

=85

SMA

Q(f

rom

Mor

isky

’86)

;6

item

s;D

I:90

%1

wee

kend

;1

wee

k;3

mon

ths

VL

>50

0@

first

year

:Non

adhe

rent

OR

5.2

(95%

CI

2.1–

13.3

);O

RA

4.4

(95%

CI

1.6–

12.3

)L

anie

ceet

al.(

2003

)3

heal

thcl

inic

s;D

akar

,Se

nega

l;N

=15

8N

/R;N

/R;#

take

n:#p

resc

ribe

d;D

I:90

%30

days

VL

mea

ndi

ffer

ence

nona

dher

ent;

Mon

th18

:1.7

log 1

0co

pies

∗ ;M

onth

24:1

.8lo

g 10

copi

es∗

Le

Moi

nget

al.(

2001

)47

clin

ical

cent

ers;

Pari

s/Fr

ance

;N=

750

N/R

;N/R

;DI:

100%

4da

ysV

L<

500:

84%

vs.7

3%∗∗

∗ ;O

R2.

0(9

5%C

I1.

3–3.

0)L

eM

oing

etal

.(20

02)

47cl

inic

alce

nter

s;Fr

ance

;N

=11

29N

/R;5

item

s;ca

tego

rica

l:100

%,8

0–99

%,

<80

%

4da

ysV

Lre

boun

d=

VL>

500;

27%

high

adhe

r.H

R=

0.4

(95%

CI

0.3–

0.6)

∗∗∗ ;

34%

mod

erat

ead

her.

HR

=0.

6(9

5%C

I0.

4–0.

8)∗∗

;53%

low

adhe

r.H

R=

1.0

Springer

AIDS Behav (2006) 10:227–245 233

Liu

etal

.(20

01)

Publ

icH

IVcl

inic

;N/R

;N

=10

8N

/R;2

item

s,co

mpo

site

adhe

renc

esc

ore;

CO

:mea

n1

wee

kE

DM

r=

0.38

∗∗∗ ;

PR:

r=

0.62

∗∗∗

VL

<40

0vs

.VL>

400

mea

nad

here

nce

8w

eek

(NS)

,24

wee

k∗0.

97(0

.85–

0.96

)vs

.0.

90(0

.85

−0.

96)

Lop

ez-S

uare

z,Fe

rnan

dez-

Gut

ierr

ezde

lA

lmo,

Pere

z-G

uzm

an,a

ndG

iron

-Gon

zale

z(1

998)

N/R

;Cad

iz,S

pain

;N=

65N

/R;N

/R;D

I:80

%N

/RV

Llo

g 10,2

drug

/3dr

ugre

gim

en;3

mon

ths:

2.9

vs.4

.4∗∗

∗ /3.

6vs

.4.9

∗ ;6

mon

ths:

3.1

vs.4

.5∗∗

∗ /3.

3vs

.4.8

∗ ;C

D4,

2dr

ug/3

drug

regi

men

;3m

onth

s:55

0vs

.356

∗∗∗ /

405

vs.

333∗ ;

6m

onth

s:56

7vs

.416

∗∗∗ /

540

vs.

400∗

Luc

as,C

heev

er,C

hais

son,

and

Moo

re(2

001)

John

sH

opki

nsA

IDS

Serv

ice;

Bal

timor

e,M

D,

US;

N=

533

N/R

;N/R

;DI:

mis

sed

>2

dose

sin

2w

eeks

2w

eeks

VL

log 1

0di

ffer

ence

0.4

(0.2

–0.7

);C

D4

cell

diff

eren

ce−

12(-

40–1

5)(s

ig.N

/R)

Mag

giol

oet

al.(

2002

)O

utpa

tient

clin

ic;B

erga

mo,

Ital

y;N

=59

7M

odifi

edA

AC

TG

(Che

sney

’00)

;N/R

;DI:

100%

90da

ysV

L<

50;7

5.6

vs.5

5.3%

∗∗∗

Man

nhei

mer

etal

.(20

02)

18C

PCR

ASi

tes;

US;

N=

1095

CPC

RA

(For

m64

6,’0

2);

N/R

;CT

:100

%,8

0–99

%,

<80

%

7da

ysV

Llo

g 10

decr

ease

∗∗∗ ;

100%

:2.8

,80–

99:

2.3,

<80

%:0

.7;C

D4

incr

ease

∗∗∗ ;

100%

:17

9,80

–99:

159,

<80

%:5

3M

artin

etal

.(20

01)

Hos

pita

lHIV

unit;

Mad

rid,

Spai

n;N

=24

2N

/R;4

item

s,#

ofpi

llsde

liver

ed/p

resc

ribe

d6

days

SRvs

.PR

;80%

adhe

renc

ecu

toff

:sen

s=25

%,

spec

=86

%,P

PV=

49%

,po

sitiv

elik

elih

ood

ratio

(LR

)=1.

8;90

%ad

here

nce:

sens

=19

%,s

pec=

84%

,PP

V=

58%

,pos

itive

LR

=1.

2M

artin

-Fer

nand

ezet

al.

(200

1)H

IVun

it;M

adri

d,Sp

ain;

N=

283

Tul

dra

’99;

2ite

ms:

Cap

able

(1–5

,cut

off<

4);E

ffor

t(1

00pt

scal

e,cu

toff

<=

36);

Phar

mac

yre

fill=

gold

stan

dard

;DI:

95%

N/R

Are

aun

der

curv

eof

mea

sure

;Cap

able

0.61

(0.5

4–0.

67);

Eff

ort0

.64

(0.5

7–0.

70).

Con

cord

ance

betw

een

nega

tive

resp

onse

on2

SRto

PRka

ppa

0.25

(0.1

3–0.

36)

Mat

hew

set

al.(

2002

)U

nive

rsity

HIV

Clin

ic;S

anD

iego

,CA

,US;

N=

175

Mod

ified

AA

CT

G(C

hesn

ey’0

0);5

item

s;D

I:Sc

ore

0–33

cuto

ff5

30da

ysE

DM

ρ=

−0.

40(s

ig.N

/R)

VL

log 1

0di

ffer

ence

∗ ;1

mon

th:.

04,3

mon

th:1

.1,6

mon

th:1

.3;V

Lun

dete

ctab

le∗ ;

CD

4∗ ;pl

asm

ale

vel

ρ=

−0.

48(s

ig.N

/R)

Mel

bour

neet

al.(

1999

)Ph

ysic

ian

offic

es;

Prov

iden

ce,R

I,U

S;N

=44

N/R

;N/R

;CO

:mea

n%

1m

onth

SR(S

D)

vs.E

DM

(SD

);1

mon

th:9

8%(3

.6)

vs.9

0%(1

4)∗ ;

2m

onth

:96%

(5)

vs.

90%

(12.

6)∗

Moa

ttiet

al.(

2000

)H

ospi

tals

;Mar

seill

es,

Avi

gon,

Nic

e,Pa

ris,

Fran

ce;

N=

164

N/R

;N/R

;DI:

80%

7da

ysV

Lm

edia

nlo

g 10

(ran

ge)∗∗

;2.7

(2.3

–5.6

)vs

.3.

9(2

.3–5

.8);

VL

unde

tect

able

orde

crea

se>

1lo

g 10

57%

vs.4

0.3%

∗ ;C

D4

med

ian

incr

ease

(NS)

;dis

ease

prog

ress

ion

(NS)

Springer

234 AIDS Behav (2006) 10:227–245

Tabl

e1

Con

tinue

d

Sour

ceSt

udy

Self

-rep

ortm

easu

reR

ecal

lpe

riod

Ass

ocia

tion

with

othe

rin

dire

ctad

here

nce

mea

sure

sA

ssoc

iatio

nw

ithcl

inic

alin

dica

tors

Mur

riet

al.(

2001

)U

nive

rsity

HIV

clin

ic;

Rom

e,It

aly;

N=

140

Res

earc

her-

crea

ted;

16ite

ms;

DI:

forg

ot1

dose

vs.

>1

dose

in3

days

1da

y;3

days

VL

dete

ctab

le;n

onad

here

nt3

days

;OR

2.2

(95%

CI

1.0–

4.7)

;Pla

sma

leve

lPI

1da

y:O

R15

.9(9

5%C

I4.

9–50

.7),

3da

y:O

R4.

4(9

5%C

I1.

7–11

.9)

Nie

uwke

rket

al.(

2001

)14

hosp

itals

;The

Net

herl

ands

,Bel

gium

;N

=16

0

Res

earc

her-

crea

ted

3ite

ms;

DI:

100%

7da

ysPC

mea

sure

for

saqu

iniv

ir∗ ;

PCm

easu

refo

rri

tona

vir

(NS)

VL>

400

@48

wee

ks;n

onad

here

nt40

%,

adhe

rent

15%

∗

Nie

uwke

rket

al.(

2001

)22

hosp

itals

;The

Net

herl

ands

;N=

224

Res

earc

her-

crea

ted

4ite

ms;

DI:

100%

7da

ysV

L>

500;

nona

dher

entO

R2.

1(9

5%C

I0.

9–4.

9);n

onad

here

ntO

RA

4.0

(95%

CI

1.4–

11.6

);dr

ugle

velm

edia

nco

ncen

trat

ion

(ran

ge)

1.1

(0.6

–1.4

)vs

.0.8

(0.5

1.1)

∗∗∗

Oyu

giet

al.(

2004

)R

esea

rch-

affil

iate

dcl

inic

san

dho

spita

ls;K

ampa

la,

Uga

nda;

N=

34

AA

CT

G(C

hesn

ey’0

0)an

dV

isua

lAna

logu

eSc

ale

(VA

S);N

/R;C

O:M

ean

3da

ys(A

AC

TG

);30

day

(VA

S)

ED

M3

day

r=

0.87

∗∗∗ ;

VA

S0.

77∗∗

∗ ;PC

3da

yr=

0.89

∗∗∗ ;

VA

S0.

86∗∗

∗ ;3

day

and

VA

Sr=

0.82

∗∗∗

VL

<40

0@

12w

eeks

;3-d

ayr=

−0.

42∗∗

;30-

day

VA

Sr=

−.0

36∗

Pale

pu,H

orto

n,T

ibbe

tts,

Mel

i,an

dSa

met

(200

4)M

edic

alan

dm

etha

done

clin

ics,

resp

itefa

cilit

y;B

osto

n,M

A,U

S;N

=19

4

N/R

;N/R

;DI:

95%

;CO

:M

ean

30da

ysV

Llo

g 10

mea

n(S

D);

1.8

(1.8

)vs

.2.7

(1.9

)∗∗∗ ;

CD

4m

ean

(SD

);41

4(2

54)

vs.

375

(216

)(N

S)Pi

nhei

roet

al.(

2002

)Pu

blic

clin

ic;P

elot

as,B

razi

l;N

=19

5R

esea

rche

r-cr

eate

d;N

/R;D

I:95

%2

days

VL

<50

0;67

.5%

vs.3

1.5%

∗∗∗ ;

CD

Cdi

seas

est

age

(NS)

Prad

ier

etal

.(20

01)

Res

earc

hco

hort

:12

outp

atie

ntho

spita

lsin

Mar

seill

es,A

vigo

n,N

ice,

and

Pari

s,Fr

ance

;N=

119

N/R

;1ite

mfo

rea

chm

edic

atio

n;D

I:10

0%3

grou

ps:(

1)no

VL

chan

geor

<0.

5de

crea

se,(

2)>

0.5

decr

ease

buts

tilld

etec

tabl

e,(3

)un

dete

ctab

le

7da

ysV

Llo

g 10;G

3vs

.G2

=O

RA

5.8

(95%

CI

1.5–

22.1

);G

3vs

.G1

=O

RA

5.6

(95%

CI

1.3–

24.7

)

Rab

oud

etal

.(20

02)

N/R

;Ita

ly,T

heN

ethe

rlan

ds,

Can

ada,

and

Aus

tral

ia;

N=

311

INC

AS,

AV

AN

TI

2an

d3

stud

ies;

N/R

;DI:

Adh

eren

cefr

om3

diff

eren

tst

udie

sw

hich

each

dich

otom

ized

diff

eren

tly92

.3,7

5,75

%

28da

ysPC

(N/R

)V

irol

ogic

failu

re;R

R3.

0(9

5%C

I1.

4–6.

1);

Test

stat

istic

sN

Rfo

rth

efo

llow

ing:

Vir

olog

icsu

ppre

ssio

n∗∗;T

ripl

edr

ug∗∗

,do

uble

drug

(NS)

;VL

unde

tect

able

∗

Schu

man

etal

.(20

01)

Res

earc

hco

hort

;Bal

timor

e,C

hica

go,D

etro

it,N

ewY

ork,

LA

,Was

h.D

C,U

S;N

=37

1

N/R

;1ite

m;D

I:75

%2

wee

ksV

Lun

dete

ctab

le;O

R3.

9(9

5%C

I1.

8–8.

5);

CD

4≥

200;

OR

2.1

(95%

CI

1.0–

4.3)

∗

Silv

eira

etal

.(20

02)

HIV

/AID

Sse

rvic

e;Pe

lota

s,B

razi

l;N

=24

4N

/R;1

item

(#ta

blet

sta

ken)

;C

T:

≥95

%,9

4–80

%,

79–6

0%,<

60%

48hr

VL

<80

acro

ssgr

oups

OR

≥95

%:O

R5.

5(9

5%C

I2.

6–11

.9);

60–7

9%:O

R4.

2(9

5%C

I1.

3–13

.3);

80%

–94%

:OR

5.6

(95%

CI

2.2–

14.1

);<

60%

:1.0

Springer

AIDS Behav (2006) 10:227–245 235

Spir

eet

al.(

2002

)R

esea

rch

coho

rt,4

7ho

spita

ls,F

ranc

e;N

=44

5N

/R;3

item

s;D

I:10

0%4

days

VL

log 1

0m

edia

nde

crea

se@

4m

onth

s;1.

7vs

.1.3

∗∗∗ ;

VL

≤50

077

%vs

.60%

∗∗∗

Tro

ttaet

al.(

2003

)R

esea

rch

coho

rts;

Rom

ean

dot

her

site

s,It

aly;

N=

596

Mur

ri’0

0;16

item

s;D

I:86

%(m

isse

d≥

1do

sela

st7

days

)

7da

ysV

L≤

500;

nona

dher

entO

R0.

7(9

5%C

I0.

5–0.

9);C

D4

<20

0/m

m;O

R0.

6(9

5%C

I0.

3–1.

0,p

=.0

6)V

inck

ean

dB

olto

n(2

002)

N/R

;Bel

gium

;N=

86PI

attit

ude

scal

e(W

eiss

N/R

)3

item

sO

rdin

al:s

cale

(1–5

,5:

exce

llent

)

4w

eeks

Clin

icia

n:r=

−0.

25(N

S);

Sign

ifica

ntot

her:

r=

−0.

42∗∗

VL

;R=

0.30

(sig

.N/R

)

Wag

ner

etal

.(20

01)

3V

AM

edic

alC

ente

rs;

Cle

vela

nd,O

H,H

oust

on,

TX

,Man

hatta

n,N

Y,U

S;N

=79

3

N/R

;4ite

ms;

CT

:0(p

oor)

,1,

2(p

erfe

ct)

4da

ysSR

and

prov

ider

agre

emen

t;ka

ppa

−.0

3,(−

.09,

.03)

(NS)

VL

<40

0;55

%vs

.36%

vs.2

2%∗∗

∗ ;O

RA

0.9

(95%

CI

0.8–

1.3)

;VL

med

ian

141

vs.

393

vs.1

679∗∗

∗ ;O

RA

0.04

(95%

CI

−0.

2–0.

1)W

agne

r,20

02C

BO

s,cl

inic

s;L

osA

ngel

es,

USA

;N=

180

Mod

ified

AA

CT

G(C

hesn

ey’0

0)N

/R;C

O:M

ean

3da

ys4-

wee

kE

DM

:r=

0.34

∗∗

Wag

ner

etal

.,20

03M

enta

lhea

lthco

mm

unity

;L

A,C

A,U

S;N

=47

N/R

;N/R

;CO

:Mea

ns3

days

;2w

eeks

ED

M3

days

∗∗∗

r=

0.61

;2w

eeks

∗∗∗

r=

0.63

VL

log 1

0r=

−0.

39∗

(rec

allp

erio

dN

/R);

VL

log 1

0m

ean

(SD

);3

day:

2.3

log 1

0(1

.0)

vs.3

.5lo

g 10

(1.2

)∗∗;C

D4

3da

ys(N

S);1

4da

ys(N

S)W

alsh

etal

.(20

01)

Publ

icly

fund

edcl

inic

;N/R

;N

=17

8R

esea

rche

r-cr

eate

d;N

/R;

CO

:Med

ian

30da

yan

dV

AS

(30

days

)

PRρ

=0.

19∗∗

;Nur

sera

ting

ρ=

0.51

∗∗;M

Dra

ting

ρ=

0.33

∗∗¶

Wal

shet

al.(

2002

)Pu

blic

HIV

clin

ic;L

ondo

n,E

ngla

nd;N

=78

AA

CT

G(C

hesn

ey’0

0),

Hec

ht’9

8,Fl

etch

er’7

9;6

item

s:3

–3da

y;1–

2w

eek,

1la

stm

isse

d;V

AS

30da

y(0

–100

%);

CO

:Mea

n

3da

y;2

wee

ks;3

0da

y(V

AS)

ED

M:U

niva

riat

elin

ear

regr

essi

on;3

day

r=

0.32

;0.

68(9

5%C

I0.

23–1

.13)

∗∗;

2w

eek

r=

0.62

;1.2

1(9

5%C

I0.

86–1

.56)

∗∗∗ ;

VA

Sr=

63;1

.09

(95%

CI

0.78

–1.3

9)∗∗

∗

VL<

50;3

day

(NS)

;14

day

ρ=

−0.

30∗∗

;V

AS

ρ=

−0.

28∗∗

Wei

ser

etal

.(20

03)

3pr

ivat

ecl

inic

s;G

abor

one,

Fran

cist

own,

Bot

swan

a;N

=93

–109

Mod

ified

AA

CT

G(C

hesn

ey’0

0);N

/R;D

I:95

%1

year

SRan

dpr

ovid

erag

reem

ent;

Kap

pa=

.35,

x2=

11.1

3∗∗∗

Wut

ohet

al.(

2001

)2

larg

eH

IVcl

inic

s;W

ashi

ngto

nD

C,U

S;N

=10

0

N/R

;N/R

;CO

:Mea

n7

days

VL

mea

n;ρ

=.−

312∗∗

Not

es.N

/R:N

otre

port

ed,N

S:N

on-s

igni

fican

t,ID

:Inf

ectio

usdi

seas

e,SD

:Sta

ndar

dde

viat

ion,

PI:P

rote

ase

inhi

bito

r.aO

dds

ratio

s,ha

zard

ratio

s,an

dre

lativ

eri

sks

are

unad

just

edun

less

deno

ted

bysu

bscr

ipt“

A”;

95%

confi

denc

ein

terv

als

deno

tesi

gnifi

canc

eun

less

only

ap-

valu

eis

give

n.bC

orre

latio

nst

atis

tics

are

Pear

son’

sr

orSp

earm

an’s

ρ.

c Sign

ifica

nce

leve

lwas

calc

ulat

edfr

omda

tapr

ovid

edin

the

artic

leus

ing

a1

sam

ple

test

ofpr

opor

tion.

dSi

gnifi

canc

ele

velw

asca

lcul

ated

from

data

prov

ided

inth

ear

ticle

usin

ga

1sa

mpl

et-

test

.∗∗∗

Springer

236 AIDS Behav (2006) 10:227–245

Fig. 1 Association is between (a) adherence and VL suppression or(b) nonadherence and VL increase or rebound. Excludes 4 studies thatshowed statistically significant associations due to overly-wide confi-dence intervals (Barroso et al., 2003) or because the association was

reported differently (e.g., nonadherence as protective from VL suppres-sion) and could not be re-calculated from published data (Cingolaniet al., 2002; LeMoing et al., 2002; Trotta et al., 2003)

Findings from the review

Study description

Study date, location, and setting

The number of publications peaked in the years 2001–2002(1997 n = 1; 1998 n = 1; 1999 n = 3; 2000 n = 6; 2001n = 22; 2002 n = 22; 2003 n = 14; and through August 2004n = 8). The vast majority of studies were conducted in theUnited States (US, n = 26) and Europe (n = 38), mainlyFrance (n = 12), Spain (n = 9), or Italy (n = 9). There weretwo from Asia, both from Hong Kong (Fong et al., 2003;Ho, Fong, and Wong, 2002), four from South America,all from Brazil (Barroso et al., 2003; Brigido et al., 2001;Pinheiro, de-Carvalho-Leite, Drachler, and Silveira, 2002)and only three recent reports from Africa, in Uganda (Oyugiet al., 2004); Botswana (Weiser et al., 2003); and Senegal(Laniece et al., 2003). Most studies (n = 61) oc-curred in hospital-based outpatient clinics, either of-fering HIV primary care or specializing in infectiousdiseases.

Eligibility criteria and sample characteristics

Eligibility criteria varied greatly across studies. Some stud-ies enrolled any adult patients on ART, while others hadextensive inclusion and exclusion criteria that created highlyspecific samples. Most studies referred to at least one of thefollowing as part of their eligibility criteria: Disease statusor clinical status as measured by CD4 count and VL; co-existing problems such as substance use; type of regimen(most required inclusion of a protease inhibitor); treatmentexperience (many studies required participants to be ART-naıve or on ART for no more than a specified amount oftime); and pregnancy status (some studies excluded preg-nant women).

Study sample size ranged from 26 (Hugen et al., 2002) to2528 (Knobel et al., 2002); only five studies had fewer than50 participants. The majority of participants in almost ev-ery study was male (range = 29 to 100% male). Specifically,in the 71 studies reporting sex of participants, 62 includedsamples that had at least 60% males; two studies had no fe-male participants, and two studies had no male participants.Most studies did not include sufficient numbers of women

Springer

AIDS Behav (2006) 10:227–245 237

to conduct analyses by sex. Where reported, these generallyindicated that there were no sex differences in adherencelevels and no interactions by sex among the adherence mea-sures and other factors. Most participants in the US studieswere members of racial/ethnic minority groups; in Europeansamples, race/ethnicity was rarely reported. Some studiesprovided data on baseline disease stage, VL, or CD4 count.

Study design and purpose

Eighteen studies employed cross-sectional survey designs,often including chart-extracted reports of VL and CD4counts. The earlier studies generally aimed to identify pre-dictors of nonadherence and often were embedded withinclinical trials; later studies often involved sub-analyses ofintervention trials. Six studies set out specifically to eval-uate adherence measures (i.e., Martin-Fernandez, Escobar-Rodriguez, Campo-Angora, & Rubio-Garcia, 2001; Martinet al., 2001; Murri et al., 2001; Vincke & Bolton, 2002;Wagner et al., 2001; Walsh, Mandalia, & Gazzard, 2002).

Self-report adherence measures

The most common self-report measure consisted of a singleitem querying the number of prescribed doses the partici-pant had missed in a specified time period (n = 22). Therewas great heterogeneity among other assessment measures,which included items assessing missed doses on the week-ends and adherence to dietary restrictions. Apart from theAdult AIDS Clinical Trials Group (AACTG) adherence mea-surement form and its variations, which were used in 15studies, a visual analog scale (six studies), and the Simpli-fied Medication Adherence Questionnaire (two studies), noother single instrument was used in more than one study.

Twenty-five studies did not provide important detailsabout the adherence assessment strategy they employed.Those that did described measures ranging from one item tothe lengthy AACTG measure that addresses each medicationover each of the last 3 days in terms of number of doses takenper day, number of pills taken per dose, and adherence to anyspecial dietary instructions (Chesney et al., 2000). Measuresvaried with respect to recall period (from 2 to 365 days); itemresponse format (i.e., closed-ended, open-ended, Likert-type,visual analogue); and whether introductory statements nor-malizing nonadherence were included. Psychometric prop-erties such as internal consistency of multi-item scales werereported in only three studies.

Most self-report interview modalities appeared toinvolve paper instruments, although this information wasnot always explicitly provided. Two studies employedcomputer-assisted self-interviews (Bangsberg, Bronstone, &Hofmann, 2002; Pinheiro et al., 2002); two were conductedover the telephone (Silveira, Draschler Mde, Leite, Pinheiro,

& da Silveira, 2002; Wagner, Kanouse, Koegel, & Sullivan,2003); and none involved the internet. Few studies reportedwhether providers, study staff, or the patients themselvesadministered the interviews.

The construct of adherence was operationalized for thedata analyses in a variety of ways–sometimes multiple waysin the same study. A continuous measure of percentage ofdoses taken was calculated often as

Prescribed doses − missed doses

Prescribed doses× 100.

Other researchers created a summary score based on somecombination of multiple items. Frequently, adherence datawere converted to dichotomous indicators of adherent ver-sus nonadherent patients, with thresholds, often apparentlyassigned post hoc, of 80% (n = 6/48 or 13% of recall periodsassessed), 90% (n = 7/48, 15%), 95% (n = 11/48, 23%), or100% (n = 21/48, 44%) or less of prescribed doses taken.

Association of self-report and other measuresof adherence

As seen in Table 2, 27 of the studies reported data on theassociation between self-reported adherence and adherenceas assessed with another indirect measure of adherence, in-cluding EDM (n = 11); pharmacy refill records (n = 9); clin-ician assessments (n = 7); pill counts (n = 3, of which twowere unannounced); chart review (patient report of adher-ence to provider; n = 1); and morphologic alterations (n = 1).In 27 of the 34, or 79%, of the recall periods examined inthese studies, associations were significant or resulted inmoderately strong kappa values. Sample sizes were insuf-ficient to compare the level of association by assessmenttechnique.

Association of self-reported adherence andclinical indicators

Most of the studies (60 of 77 or 78%) assessed VL, al-though the types of tests and their detection thresholds (e.g.,Roche Amplicor, 50 copies/µL) were not uniformly de-scribed. Many were taken from a review of medical recordsinstead of based on blood samples drawn on the same dayadherence was assessed. Analyses of the relation betweenself-reported adherence and VL most often involved bivariatetests of association such as Pearson product moment corre-lations. These rarely controlled for confounders or assessedpotential effect modifiers such as previous experience withART. When they did, the association between self-reportedadherence and VL usually remained statistically significant(e.g., Alcoba et al., 2003; Nieuwkerk, Gisolf, Sprangers, &Danner, 2001).

Springer

238 AIDS Behav (2006) 10:227–245

Tabl

e2

Ass

ocia

tion

ofse

lf-r

epor

ted

antir

etro

vira

ladh

eren

cew

ithad

here

nce

asm

easu

red

byot

her

indi

rect

mea

sure

sor

with

clin

ical

indi

cato

rs,b

yre

call

peri

od

Self

-rep

orte

dad

here

nce

reca

llpe

riod

(day

s)

12

34

714

2830

30(V

AS)

90,1

80,o

r36

5

Com

bine

dtim

epe

riod

sor

“las

tm

isse

d”To

tal

Clin

ical

impa

ctH

IV-1

RN

Avi

ral

load

(VL

)1/

32–4

3/43,

5–7

4/64,

8–12

9/10

13–2

213

/162,

3,23

–35

3/311

,12

,36

3/337

–39

10/1

13,10

,12

,40

–47

2/210

,12

3/348

,49

506/

651–5

657

/67

Plas

ma

Rx

leve

l1/

141/

140/

1222/

224,

351/

1405/

6C

D4

coun

t0/

1111/

414,

16,

177/

1123

,24

,26

,28

,29

,32

,34

,57

–59

2/336

,60

1/154

4/541

,42

,44

,46

,47

1/151

16/2

6D

isea

sepr

ogre

s-si

on/m

orta

lity

1/161

0/16

0/129

1/139

2/241

,62

4/6

Oth

erin

dire

ctm

easu

reof

adhe

renc

eE

lect

roni

cda

tam

onito

ring

1/12

4/411

,12

,63

,64

2/22,

272/

211,

122/

2652/

210,

1213

/13

Pill

coun

t1/

1661/

1431/

2302/

210,

435/

6Ph

arm

acy

refil

l1/

127,

671/

1682/

2Pr

ovid

eres

timat

e0/

29,66

0/121

0/169

2/268

1/170

3/7

Oth

er1/

237,

711/

1711/

237,

693/

5Si

gnifi

canc

eno

tre

port

edV

L43

PC38

VL

46,

69PL

46E

DM

72V

L43

Not

es.

Frac

tions

indi

cate

the

prop

ortio

nof

asso

ciat

ions

that

wer

est

atis

tical

lysi

gnifi

cant

(i.e

.,p

<0.

05or

95%

confi

denc

ein

terv

als

excl

udin

g1.

0).

Una

djus

ted

resu

ltsar

ere

port

edw

here

avai

labl

e.Su

pers

crip

ted

num

bers

refe

rto

cita

tions

that

are

mar

ked

with

anas

teri

skin

the

Ref

eren

ces

sect

ion.

Not

eth

atso

me

stud

ies

prov

ided

data

onm

ore

than

one

reca

llpe

riod

.N

otre

pres

ente

din

this

tabl

e:G

ouja

rdet

al.,

2003

;H

orne

etal

.,20

04;

Hug

enet

al.,

2002

;L

opez

-Sua

rez

etal

.,19

98;

Mar

tin-F

erna

ndez

etal

.20

01,

asse

lf-r

epor

tre

call

peri

odco

uld

not

bede

term

ined

.R

esul

tsfo

rm

ean

VL

,de

tect

able

VL

,an

ddi

ffer

ent

VL

outc

ome

cate

gori

es(e

.g.,

>50

0ce

lls,

200–

499)

wer

een

tere

din

divi

dual

lyif

sepa

rate

anal

yses

wer

eco

nduc

ted.

‘Oth

er’=

med

ical

reco

rd,m

orph

olog

ical

tera

tions

,or

sign

ifica

ntot

her.

Springer

AIDS Behav (2006) 10:227–245 239

In 57 of 67 (85%) of recall periods assessed (note thatsome studies reported data on more than one recall period),self-reported adherence was significantly related to VL(see Table 2). The magnitude of the significant correlationsranged from 0.30 to 0.60. Across different recall periods,odds ratios and hazard ratios of the association betweenself-reported adherence and VL were on the order of2.0, with 95% confidence bounds generally excluding 1.0(see Fig. 1). Findings from analyses of the proportionof patients with good adherence (with viral suppressionas the outcome) and of the proportion of patients withpoor adherence (with higher VL as the outcome) werecomparable.

As seen in Table 2, fewer studies found a positive cor-relation between self-reported adherence and CD4 count(16/26 or 62%) of recall periods. Five studies (Brigido et al.,2001; Gao et al., 2000; Ho et al., 2002; Moatti et al., 2000;Pinheiro et al., 2002) reported associations of self-reportedadherence with disease progression as defined by develop-ment of a new opportunistic infection or disease staging;three were significant. Two studies assessed mortality asthe outcome; in both, the association with self-report wassignificant (Brigido et al., 2001; Garcia de Olalla et al.,2002).

Association of length of recall period and VL

As seen in Table 2, there was some suggestion of aneffect of the length of the self-report adherence assess-ment recall period on the relation with VL: Adherencewas associated with VL in 88% of recall periods thatwere greater than 3 days and in 64% of those that were3 days or less, χ2 (N = 63) = 4.16, p = 0.04. However,an unadjusted bivariate logistic regression included 1.0(crude odds ratio 0.25, 95% confidence interval 0.06–1.0,p = 0.05).

Conclusions and implications

A review of the literature on self-report measures of ARTadherence identified 77 published articles meeting eligibilitycriteria. Most were published in 2000–2001 and were basedon data from hospital-based clinic samples of predominantlymen from the US and Europe. The most common assess-ment strategy involved asking patients about the numberof missed doses over a specified recall period; otherwise,there was great variability in the content of the items, the re-sponse format, and the recall period. The lack of widespreaduse of standardized measures made it difficult to evalu-ate any particular measure or to compare measures acrossstudies.

Nonetheless, self-reported adherence was significantlyrelated to adherence as assessed by other indirect mea-sures such as EDM and pill count in 79% of studiescomparing measurement approaches. Although we werenot able to statistically examine these issues in this re-view, it would be helpful to know which techniques aremost closely associated with VL and whether any socio-demographic indicators moderate these relationships. Self-report measures may not be feasible with some individu-als (such as the cognitively impaired); therefore, data onwhich other methods are appropriate options would beuseful.

We observed a robust pattern of association between self-reported adherence and VL: In 84% of recall periods, self-reported adherence was associated with VL based on oddsratios or simple measures of correlation. The association wasstatistically significant across a variety of self-report mea-sures, administration modalities, and recall periods. Thesefindings are consistent with the conclusions of a recent meta-analysis of adherence studies (Nieuwkerk & Oort, 2005).These results may provide some reassurance to practitionersand researchers employing self-reported adherence strate-gies.

There was some suggestion that longer recall periods maybe more likely than shorter ones to yield estimates of ad-herence that are significantly correlated with VL, althoughthis was not statistically conclusive in our review or in thepreviously published meta-analysis (P. Nieuwkerk, personalcommunication April 21, 2005). The association betweenself-report and CD4 was less consistent, a finding that is notentirely unexpected, as viral load and CD4 count generallycorrelate but discordant results are common. Furthermore,CD4 response can be somewhat delayed following initialART initiation. For this reason, many experts believe that VLis the best measure of therapeutic response to ART, thoughCD4 remains the best clinical prognostic indicator (Bartlett& Gallant, 2004).

These findings are limited by several factors. Becausemost of the studies were conducted in the West, resultsmay not be generalizable to resource-poor settings. Thelack of data on refusal rates and the preponderance of non-probability samples of patients who were largely in care,participants in cohort studies, or volunteers receiving mon-etary incentives further limit the generalizability of thesefindings to other HIV populations. Relatedly, we were notable to determine whether self-report measures have differ-ential validity for groups varying in socio-demographic ordisease factors, because these variables, if assessed and re-ported, were not usually included in the analyses and smallsample sizes limited the ability to conduct subgroup anal-yses. The possibility of publication bias—that studies withnon-significant associations between adherence and VL are

Springer

240 AIDS Behav (2006) 10:227–245

less likely to be published—also cannot be definitively ruledout.

Lack of information about the interviewer’s relationshipto the participant and mode of interview administration (Di-Matteo, 2004; Rudd et al., 1990), as well as the lack ofany systematic manipulation of these two variables in thestudies we reviewed, limits the extent to which we cancomment on their relevance to our findings. It is worthexploring whether audio computer-assisted self-interviews(ACASI) can contribute to the quality and validity of ARTself-reporting, as has been seen with respect to sex and othersensitive behaviors (Schroder, Carey, & Vanable, 2003). Anexample of the visual analog scale as presented in a hand-held computer can be viewed at http://faculty.washington.edu/wcurioso/emulator/emulator.htm.

Finally, the timing of the adherence assessment may af-fect the strength of its association with clinical outcome.We would not expect perfect agreement between assessmentof self-reported adherence over a brief, recent recall periodand current VL, given all the other potential effect modifierssuch as co-morbidity and earlier periods of nonadherencethat may have resulted in resistance (Bangsberg et al., 2003).Most studies examined the association of adherence and VLcross-sectionally, but adherence over time (serial measure-ments within patients) may better predict VL prospectively.Longitudinal HIV studies increasingly include tests for geno-typic or phenotypic resistance, parameters that may be usefulin future ART adherence evaluations.

Obtaining accurate data on the association between as-sessed ART adherence and relevant outcomes requiresmethodologically precise studies. Future research in this areashould report baseline characteristics that may confound ormodify (Raboud, Harris, Rae, & Montaner, 2002) the associ-ation between self-reported adherence and health outcomes,including CD4 count nadir, baseline VL, class and durationof previous ART experience, and possibly, evidence of spe-cific ART viral resistance. This precision will enable moreaccurate estimations of the quality of assessment methods,although given the complex and dynamic nature of HIVdisease, no single adherence assessment measure can be ex-pected to correlate perfectly with clinical indicators or clin-ical outcomes.

Recommendations for best practices in HIV researchand clinical management

Our findings suggest that both researchers and cliniciansmay proceed with the use of self-report measures of ARTadherence with some confidence in their validity at least interms of their associations with other indirect measures ofadherence and VL, a reliable surrogate marker of clinicalimpact. Some experts have advocated the use of multipleadherence measures (Caplan, Harrison, Wellons, & Frech,

1980; Ickovics, 1997; Konkle-Parker, 2000; Samet, Sullivan,Traphagen, & Ickovics, 2001). Our findings suggest this maynot be routinely required in clinical arenas, where VL andother biological markers are often readily available and fundsfor additional assessments are limited. However, there areat least two situations in which further assessment may bewarranted.

First, in intervention trials, the use of less subjective meth-ods such as EDM or unannounced pill counts may be worth-while because of the potential reporting bias with self-reportstrategies in the intervention conditions. Second, althoughpatient reports of nonadherence can generally be believed,clinicians may be at a loss to interpret individual patientreports of perfect (100%) adherence. Pharmacy refill data,where accessible, may be useful in validating self-reported“perfect” adherence. In one study, adherence as measured bytime-to-pharmacy refill was able to distinguish VL impactamong self-reportedly perfect adherers (Grossberg, Zhang,& Gross, 2004). Other strategies to mitigate the ceiling ef-fect of reportedly perfect adherence include calculating theproportion of times across multiple interviews that 100% ad-herence was reported and supplementing the standard 3-daymissed dose item with another item assessing the timing ofthe last missed a dose or whether any doses were missedin the last 30 days (Mannheimer, Friedland, Matts, Child,& Chesney, 2002). These approaches may assist cliniciansin identifying patients claiming to be adherent who, in fact,need ART adherence support.

When employing self-report strategies, researchers andclinicians alike should capitalize on the flexibility of self-report methodologies and inquire beyond the assessmentof missed doses, gathering information on other aspectsof adherence such as knowledge of medication namesand prescribed dosing regimens, attention to special di-etary instructions, and patterns of nonadherence on week-ends, mid-day, or when daily schedules change. Barri-ers to adherence and facilitators are also important fac-tors that are inaccessible with other adherence assessmentmethodologies.

Adherence experts have developed guidelines for assess-ment that are geared toward minimizing social desirability.These include using self-administered measures with open-ended and forced choice items; broaching the topic with apreamble acknowledging the low prevalence and difficultyof perfect adherence; wording items in such a way that non-adherence is presented as expected and accepted; queryingreasons for nonadherence; focusing on recent behavior; spec-ifying a time frame; aiding recall when possible using medi-cation lists and diagrams of pills; anchoring reports to salientevents; embedding threatening with non-threatening items;using authority to justify and normalize the behavior; andending with a reliability check of the accuracy of responses(Miller & Hays, 2000).

Springer

AIDS Behav (2006) 10:227–245 241

Notes: aBased on Golin et al. (2002); bBased on Wash, Mandalia, and Gazzard (2002). An exact percentage can be calculated by measuring thedistance from 0 to mark in cm or inches; cBased on Knobel et al. (2002).

Researchers designing statistical analyses and cliniciansseeking guidance for advising patients could benefit fromrecommendations regarding an appropriate threshold of ad-herence necessary for favorable clinical outcomes. In thestudies we reviewed, thresholds appeared to be often deter-mined post hoc, increasing the probability of Type I error. Insome instances, a threshold was predetermined but analyseswere conducted with a continuous measure of adherence.Generally speaking, parametric tests of continuous variableswill have more power than nonparametric analyses of di-chotomous variables but will not define a clinically relevantcutoff. Given that continuous measures of self-report arehighly skewed and non-normal, it may be most valid to di-chotomize at 100% for statistical analyses. However, as aclinical goal, this level may be unreasonable for patients inthe long term. Optimal virologic success declines rapidlyin patients taking fewer than 95% of their prescribed doses(Paterson et al., 2000). Nonetheless, one study using phar-macy refill data among 923 HIV-positive patients showedthat there was no difference in the risk of disease progressionbetween those with moderate (70–90%) and high (>90%)levels of adherence compared to those with low (<70%) ad-herence (Kitahata et al., 2004). It is worth exploring whetherpatients can reliably make fine distinctions about their ad-herence behavior, such as judging it as either less than 80%or less than 85% (Bangsberg, Moss, & Deeks, 2004).

Which recall period is best to use is an open question.Patients do report more accurately over briefer time periods,with accuracy dropping off as rapidly as beyond 24 hr(Turner & Hecht, 2001; Wagner & Miller, 2004; Walsh,Horne, Dalton, Burgess, & Gazzard, 2001). It is worth

considering, however, whether somewhat longer recallperiods may yield more useful data as the increasing use ofonce-daily ART dosing may now result in too few dosingtimes in a very brief (i.e., 1–3 day) recall period to providesufficient variability in adherence (Paterson et al., 2000). Avery short interval may not allow for differentiation betweenpatients whose good adherence is consistent and those whoreport good adherence over a recent brief time period butwho are generally less adherent. A particular advantage of a7-day recall period is that it will always include a weekend,during which adherence is often problematic.

Recommended self-report measures are presented inFig. 2. These items are drawn both from the literature andfrom clinical experience and incorporate use of normalizinglanguage, 7-day recall, and exploration of barriers to ad-herence (Morisky, Green, & Levine, 1986). Many differentself-report measures appear to have an association with VL.Researchers and clinicians may choose single or multipleitems based on their needs, weighing the need to assess inac-curate dosing or dietary adherence with the desire to reducerespondent burden. Longitudinal use of the increasingly uti-lized visual analog scale may be enhanced by measuring theexact distance from zero to the patient’s mark. We suggestuse of the term “dose” over “pills” as patients generally donot take partial doses (G. Wagner, personal communicationMarch 2, 2005) and it is easier to calculate the number ofmissed doses than the exact number of pills missed acrossmissed doses. Exploring the reasons why patients “forget”to take their medications may uncover important issues thatcan be addressed with subsequent potential problem-solving(Bartlett, 2002). More consistent use of items such as these

Springer

242 AIDS Behav (2006) 10:227–245

would allow comparison of self-report measure psychomet-ric and clinical performance across populations.

The ability to make more definitive recommendations re-garding precise measurement strategies will be enhancedwith further research that explicitly addresses some ofthe issues we have raised. In the meantime, results fromthis extensive literature review offer some direction forHIV researchers and clinicians in their critically im-portant work attempting to address and enhance ARTadherence.

Acknowledgments We are grateful to researchers who shared theirmaterials with us and provided feedback regarding this review, es-pecially Pythia Nieuwkerk and Glenn Wagner. This work was sup-ported by University of Washington Center for AIDS Research So-ciobehavioral and Prevention Research Core (P30 AI 27757) fundingto Dr. Kurth, 2 R01 MH58986 to Dr. Simoni, and F31 MH71179 toMr. Pantalone.

References

(Numbered asterisks indicate studies cited in Fig. 1)

*22 Alcoba, M., Cuevas, M. J., Perez-Simon, M. R., Mostaza, J.L., Ortega, L., Ortiz de Urbina, J., et al. (2003). Assessment ofadherence to triple antiretroviral treatment including indinavir:Role of the determination of plasma levels of indinavir. Journalof Acquired Immune Deficiency Syndromes, 33(2), 253–258.

*50 Aloisi, M. S., Arici, C., Balzano, R., Noto, P., Piscopo, R.,Filice, G., et al. (2002). Behavioral correlates of adherence toantiretroviral therapy. Journal of Acquired Immune DeficiencySyndromes, 31(Suppl 3), S145–S148.

*57 Altice, F. L., Mostashari, F., and Friedland, G. H. (2001). Trustand the acceptance of and adherence to antiretroviral therapy.Journal of Acquired Immune Deficiency Syndromes, 28(1), 47–58.

*58 Ammassari, A., Antinori, A., Aloisi, M. S., Trotta, M. P.,Murri, R., Bartoli, L., et al. (2004). Depressive symptoms,neurocognitive impairment, and adherence to highly active an-tiretroviral therapy among HIV-infected persons. Psychoso-matics, 45(5), 394–402.

*34 Antinori, A., Cozzi-Lepri, A., Ammassari, A., Trotta, M. P.,Nauwelaers, D., Hoetelmans, R., et al. (2004). Relative prog-nostic value of self-reported adherence and plasma NNRTI/PIconcentrations to predict virological rebound in patients ini-tially responding to HAART. Antiviral Therapy, 9(2), 291–296.

*2 Arnsten, J. H., Demas, P. A., Farzadegan, H., Grant, R. W.,Gourevitch, M. N., Chang, C. J., et al. (2001). Antiretroviraltherapy adherence and viral suppression in HIV-infected drugusers: Comparison of self-report and electronic monitoring.Clinical Infectious Diseases, 33(8), 1417–1423.

*8 Bangsberg, D. R., Hecht, F. M., Charlebois, E. D., Zolopa,A. R., Holodniy, M., Sheiner, L., et al. (2000). Adherence toprotease inhibitors, HIV-1 viral load, and development of drugresistance in an indigent population. AIDS, 14(4), 357–366.

*9 Bangsberg, D. R., Bronstone, A., and Hofmann, R. (2002).A computer-based assessment detects regimen misunderstand-ings and nonadherence for patients on HIV antiretroviral ther-apy. AIDS Care, 14(1), 3–15.Bangsberg, D. R., Charlebois, E. D., Grant, R. M., Holodniy,M., Deeks, S. G., Perry, S., et al. (2003). High levels of ad-

herence do not prevent accumulation of HIV drug resistancmutations. AIDS, 17(13), 1925–1932.

*66 Bangsberg, D. R., Hecht, F. M., Clague, H., Charlebois, E. D.,Ciccarone, D., Chesney, M., et al. (2001). Provider assessmentof adherence to HIV antiretroviral therapy. Journal of AcquiredImmune Deficiency Syndromes, 26(5), 435–442.Bangsberg, D. R., Moss, A. R., and Deeks, S. G. (2004). Para-doxes of adherence and drug resistance to HIV antiretroviraltherapy. Journal of Antimicrobial Chemotherapy, 53(5), 696–699.Bangsberg, D. R., Charlebois, E. D., Grant, R. M., Holodniy,M., Deeks, S. G., Perry, S., et al. (2003). High levels of ad-herence do not prevent accumulation of HIV drug resistantmutations. AIDS, 17, 1925–1932.

*40 Barroso, P. F., Schechter, M., Gupta, P., Bressan, C., Bomfim,A., and Harrison, L. H. (2003). Adherence to antiretroviraltherapy and persistence of HIV RNA in semen. Journal ofAcquired Immune Deficiency Syndromes, 32(4), 435–440.Bartlett, J. A. (2002). Addressing the challenges of adherence.Journal of Acquired Immune Deficiency Syndromes, 29(Suppl1), S2–S10.Bartlett, J. G., and Gallant, J. E. (2004). Medical managementof HIV infection (2004 ed.). Baltimore, MD: Johns HopkinsMedicine Health Publishing Business Group.Bova, C. A., Fennie, K. P., Knafl, G. J., Dieckhaus, K. D.,Watrous, E., and Williams, A. B. (2005). Use of electronicmonitoring devices to measure antiretroviral adherence: Prac-tical considerations. AIDS and Behavior, 9(1), 103–110.

*41 Brigido, L. F., Rodrigues, R., Casseb, J., Oliveira, D.,Rossetti, M., Menezes, P., et al. (2001). Impact of adherence toantiretroviral therapy in HIV-1-infected patients at a universitypublic service in Brazil. AIDS Patient Care and STDS, 15(11),587–593.Caplan, R. D., Harrison, R. V., Wellons, R. V., and Frech, J. R.P. (1980). Social support and patient adherence: Experimentaland survey findings. Ann Arbor, MI: Survey Research Center,University of Michigan.

*13 Carrieri, P., Cailleton, V., Le Moing, V., Spire, B., Dellamon-ica, P., Bouvet, E., et al. (2001). The dynamic of adherence tohighly active antiretroviral therapy: Results from the FrenchNational APROCO cohort. Journal of Acquired Immune Defi-ciency Syndromes, 28(3), 232–239.

*14 Carrieri, M. P., Raffi, F., Lewden, C., Sobel, A., Michelet, C.,Cailleton, V., et al. (2003). Impact of early versus late ad-herence to highly active antiretroviral therapy on immuno-virological response: A 3-year follow-up study. Antiviral Ther-apy, 8(6), 585–594.

*48 Catz, S. L., Kelly, J. A., Bogart, L. M., Benotsch, E. G.,and McAuliffe, T. L. (2000). Patterns, correlates, and barri-ers to medication adherence among persons prescribed newtreatments for HIV disease. Health Psychology, 19(2), 124–133.

*42 Cederfjall, C., Langius-Eklof, A., Lidman, K., and Wredling,R. (2002). Self-reported adherence to antiretroviral treatmentand degree of sense of coherence in a group of HIV-infectedpatients. AIDS Patient Care and STDS, 16(12), 609–616.

*51 Cingolani, A., Antinori, A., Rizzo, M. G., Murri, R., Ammas-sari, A., Baldini, F., et al. (2002). Usefulness of monitoring HIVdrug resistance and adherence in individuals failing highly ac-tive antiretroviral therapy: A randomized study (ARGENTA).AIDS, 16(3), 369–379.Chesney, M. A., Ickovics, J. R., Chambers, D. B., Gifford, A. L.,Neidig, J., Zwickl, B., et al. (2000). Self-reported adherenceto antiretroviral medications among participants in HIV clin-ical trials: The AACTG adherence instruments. Patient CareCommittee and Adherence Working Group of the Outcomes

Springer

AIDS Behav (2006) 10:227–245 243

Committee of the Adult AIDS Clinical Trials Group (AACTG).AIDS Care, 12(3), 255–266.

*5 Cohn, S. E., Kammann, E., Williams, P., Currier, J. S., andChesney, M. A. (2002). Association of adherence to Mycobac-terium avium complex prophylaxis and antiretroviral therapywith clinical outcomes in acquired immunodeficiency syn-drome. Clinical Infectious Diseases, 34(8), 1129–1136.DiMatteo, M. R. (2004). Variations in patients’ adherence tomedical recommendations: A quantitative review of 50 yearsof research. Medical Care, 42(3), 200–209.

*23 Dorz, S., Lazzarini, L., Cattelan, A., Meneghetti, F., Novara,C., Concia, E., et al. (2003). Evaluation of adherence to an-tiretroviral therapy in Italian HIV patients. AIDS Patient Careand STDs, 17(1),33.

*52 Duong, M., Piroth, L., Peytavin, G., Forte, F., Kohli, E.,Grappin, M., et al. (2001). Value of patient self-report andplasma human immunodeficiency virus protease inhibitor levelas markers of adherence to antiretroviral therapy: Relationshipto virologic response. Clinical Infectious Diseases, 33(3), 386–392.

*15 Duran, S., Peytavin, G., Carrieri, P., Raffi, F., Ecobichon, J. L.,Pereira, E., et al. (2003). The detection of non-adherence byself-administered questionnaires can be optimized by proteaseinhibitor plasma concentration determination. AIDS, 17(7),1096–1099.

*53 Duran, S., Saves, M., Spire, B., Cailleton, V., Sobel, A.,Carrieri, P., et al. (2001). Failure to maintain long-term ad-herence to highly active antiretroviral therapy: The role oflipodystrophy. AIDS, 15(18), 2441–2444.

*24 Duran, S., Solas, C., Spire, B., Carrieri, M. P., Fuzibet, J. G.,Costagliola, D., et al. (2001). ‘Do HIV-infected injecting drugusers over-report adherence to highly active antiretroviral ther-apy?’ A comparison between patients’ self-reports and serumprotease inhibitor concentrations in the French Manif 2000cohort study. AIDS, 15(8), 1075–1077.

*70 Eldred, L., Wu, A., Chaisson, R., and Moore, R. (1998). Ad-herence to antiretroviral and pneumocystis prophylaxis in HIVdisease. Journal of Acquired Immune Deficiency Syndromesand Human Retrovirology, 18(2), 117–125.

*54 Fong, O. W., Ho, C. F., Fung, L. Y., Lee, F. K., Tse, W. H.,Yuen, C. Y., et al. (2003). Determinants of adherence to highlyactive antiretroviral therapy (HAART) in Chinese HIV/AIDSpatients. HIV Medicine, 4(2), 133–138.

*61 Gao, X., Nau, D., Rosenbluth, S., Scott, V., and Woodward, C.(2000). The relationship of disease severity, health beliefs andmedication adherence among HIV patients. AIDS Care, 12(4),387–398.

*62 Garcia de Olalla, P., Knobel, H., Carmona, A., Guelar, A.,Lopez-Colomes, J. L., and Cayla, J. A. (2002). Impact of ad-herence and highly active antiretroviral therapy on survival inHIV-infected patients. Journal of Acquired Immune DeficiencySyndromes, 30, 105–110.Geletko, S. M., Segarra, M., Mayer, K. H., Fiore, T. C.,Bettencourt, F. A., Flanigan, T. P., et al. (1996). Electronic com-pliance assessment of antifungal prophylaxis for human im-munodeficiency virus-infected women. Antimicrobial Agentsand Chemotherapy, 40(6), 1338–1341.

*25 Gifford, A., Bormann, J., Shively, M., Wright, B., Richman, D.,and Bozzette, S. (2000). Predictors of self-reported adherenceand plasma HIV concentrates in patients on multi drug an-tiretroviral regimens. Journal of Acquired Immune DeficiencySyndromes, 23(5), 386–395.

*43 Giordano, T. P., Guzman, D., Clark, R., Charlebois, E. D., andBangsberg, D. R. (2004). Measuring adherence to antiretroviraltherapy in a diverse population using a visual analogue scale.HIV Clinical Trials, 5(2), 74–79.

Godin, G., Gagne, C., and Naccache, H. (2003). Validationof a self-reported questionnaire assessing adherence to an-tiretroviral medication. AIDS Patient Care and STDs, 17(7),325.

*72 Golin, C. E., Liu, H., Hays, R. D., Miller, L. G., Beck, C. K.,Ickovics, J., et al. (2002). A prospective study of predictors ofadherence to combination antiretroviral medication. Journal ofGeneral Internal Medicine, 17(10), 756–765.

*59 Gordillo, V., del Amo, J., Soriano, V., and Gonzalez-Lahoz,J. (1999). Sociodemographic and psychological variables in-fluencing adherence to antiretroviral therapy. AIDS, 13(13),1763–1769.Gordis, L. (1979). Conceptual and methodologic problems inmeasuring patient compliance. In D. W. T. R. B. Haynes andD. L. Sackett (Eds.), Compliance in health care (pp. 23–48).Baltimore, MD: Johns Hopkins University.

*23 Goujard, C., Bernard, N., Sohier, N., Peyramond, D., Lancon,F., Chwalow, J., et al. (2003). Impact of a patient educationprogram on adherence to HIV medication: A randomized clin-ical trial. Journal of Acquired Immune Deficiency Syndromes,34(2), 191–194.Grossberg, R., Zhang, Y., and Gross, R. (2004). A time-to-prescription-refill measure of antiretroviral adherence pre-dicted changes in viral load in HIV. Journal of Clinical Epi-demiology, 57(10), 1107–1110.

*37 Haubrich, R. H., Little, S. J., Currier, J. S., Forthal, D. N.,Kemper, C. A., Beall, G. N., et al. (1999). The value ofpatient-reported adherence to antiretroviral therapy in predict-ing virologic and immunologic response. AIDS, 13, 1099–1107.

*39 Ho, C. F., Fong, O. W., and Wong, K. H. (2002). Patient self-report as a marker of adherence to antiretroviral therapy. Clin-ical Infectious Diseases, 34(11), 1534–1535.Hogg, R., Heath, K., Bangsberg, D. R., Yip, B., Press, N.,O’shaughnessy, M. V., et al. (2002). Intermittent use of triplecombination therapy is predictive of mortality at baseline andafter one year of follow-up. AIDS, 16, 1051–1058.

*74 Horne, R., Buick, D., Fisher, M., Leake, H., Cooper, V., andWeinman, J. (2004). Doubts about necessity and concerns aboutadverse effects: Identifying the types of beliefs that are associ-ated with non-adherence to HAART. International Journal ofSTD AIDS, 15(1), 38–44.

*75 Hugen, P. W., Langebeek, N., Burger, D. M., Zomer, B., vanLeusen, R., Schuurman, R., et al. (2002). Assessment of adher-ence to HIV protease inhibitors: Comparison and combinationof various methods, including MEMS (electronic monitoring),patient and nurse report, and therapeutic drug monitoring. Jour-nal of Acquired Immune Deficiency Syndromes, 30(3), 324–334.Ickovics, J. (1997). Measures of adhference. Paper presented atthe Adherence to New HIV Therapies: A Research Conference,Washington, DC..

*16 Ickovics, J. R., Cameron, A., Zackin, R., Bassett, R., Chesney,M., Johnson, V. A., et al. (2002). Consequences and determi-nants of adherence to antiretroviral medication: Results fromAdult AIDS Clinical Trials Group protocol 370. Antiviral Ther-apy, 7(3), 185–193.

*26 Ingersoll, K. (2004). The impact of psychiatric symptoms, druguse, and medication regimen on non-adherence to HIV treat-ment. AIDS Care, 16(2), 199–211.

*63 Kimmerling, M., Wagner, G., and Ghosh-Dastidar, B. (2003).Factors associated with accurate self-reported adherence toHIV antiretrovirals. International Journal of STD AIDS, 14(4),281–284.

*17 Kleeberger, C. A., Phair, J. P., Strathdee, S. A., Detels, R.,Kingsley, L., and Jacobson, L. P. (2001). Determinants of

Springer

244 AIDS Behav (2006) 10:227–245

heterogeneous adherence to HIV-antiretroviral therapies in theMulticenter AIDS Cohort Study. Journal of Acquired ImmuneDeficiency Syndromes, 26(1), 82–92.Kitahata, M. M., Reed, S. D., Dillingham, P. W., Van Rompaey,S. E., Young, A. A., Harrington, R. D., et al. (2004). Pharmacy-based assessment of adherence to HAART predicts virologicand immunologic treatment response and clinical progressionto AIDS and death. International Journal of STD AIDS, 15(12),803–810.

*55 Knobel, H., Alonso, J., Casado, J., Collazos, J., Gonzalez, J.,Ruiz, I., et al. (2002). Validation of a simplified medicationadherence questionnaire in a large cohort of HIV-infected pa-tients: The GEEMA Study. AIDS, 16, 605–613.

*44 Knobel, H., Guelar, A., Carmona, A., Espona, M., Gonzalez,A., Lopez-Colomes, J. L., et al. (2001). Virologic outcome andpredictors of virologic failure of highly active antiretroviraltherapy containing protease inhibitors. AIDS Patient Care andSTDs, 15, 193–199.

*56 Knobel, H., Vallecillo, G., Guelar, A., Pedrol, E., Soler, A.,Carmona, A., et al. (2004). Simplified therapy with zidovu-dine, lamivudine, and abacavir for very nonadherent, treatment-failing patients. HIV Clinical Trials, 5(2), 65–73.Konkle-Parker, D. J. (2000). What is the best way to measuremedication adherence? HIV Clinical Trials, 12(3), 11–12.Laniece, I., Ciss, M., Desclaux, A., Diop, K., Mbodj, F.,Ndiaye, B., et al. (2003). Adherence to HAART and its prin-cipal determinants in a cohort of Senegalese adults. AIDS,17(Suppl 3), S103–S108.Le Moing, V., Chene, G., Carrieri, M. P., Alioum, A., Brun-Vezinet, F., Piroth, L., et al. (2002). Predictors of virologi-cal rebound in HIV-1-infected patients initiating a proteaseinhibitor-containing regimen. AIDS, 16(1), 21–29.Le Moing, V., Chene, G., Carrieri, M. P., Besnier, J. M.,Masquelier, B., Salamon, R., et al. (2001). Clinical, biologic,and behavioral predictors of early immunologic and virologicresponse in HIV-infected patients initiating protease inhibitors.Journal of Acquired Immune Deficiency Syndromes, 27(4),372–376.

*27 Liu, H., Golin, C. E., Miller, L. G., Hays, R. D., Beck, C. K.,Sanandaji, S., et al. (2001). A comparison study of multiplemeasures of adherence to HIV protease inhibitors. Annals ofInternal Medicine, 134(10), 968–977.

*60 Lucas, G. M., Cheever, L. W., Chaisson, R. E., and Moore,R. D. (2001). Detrimental effects of continued illicit drug use onthe treatment of HIV-1 infection. Journal of Acquired ImmuneDeficiency Syndromes, 27(3), 251–259.

*76 Lopez-Suarez, A., Fernandez-Gutierrez del Almo, C., Perez-Guzman, E., and Giron-Gonzalez, J. A. (1998). Adherenceto the antiretroviral treatment in asymptomatic HIV-infectedpatients. AIDS, 12(6), 685–686.

*49 Maggiolo, F., Ripamonti, D., Arici, C., Gregis, G., Quinzan, G.,Camacho, G. A., et al. (2002). Simpler regimens may enhanceadherence to antiretrovirals in HIV-infected patients. HIV Clin-ical Trials, 3(5), 371–378.

*28 Mannheimer, S., Friedland, G., Matts, J., Child, C., andChesney, M. (2002). The consistency of adherence to antiretro-viral therapy predicts biologic outcomes for human immun-odeficiency virus-infected persons in clinical trials. ClinicalInfectious Diseases, 34(8), 1115–1121.Martin, J., Sabugal, G. M., Rubio, R., Sainz-Maza, M., Blanco,J. M., Alonso, J. L., et al. (2001). Outcomes of a health ed-ucation intervention in a sample of patients infected by HIV,most of them injection drug users: Possibilities and limitations.AIDS Care, 13(4), 467–473.Martin-Fernandez, J., Escobar-Rodriguez, I., Campo-Angora,M., and Rubio-Garcia, R. (2001). Evaluation of adherence

to highly active antiretroviral therapy. Archives of InternalMedicine, 161(22), 2739–2740.

*67 Martin, J., Sabugal, G. M., Rubio, R., Sainz-Maza, M., Blanco,J. M., Alonso, J. L., et al. (2001). Outcomes of a health ed-ucation intervention in a sample of patients infected by HIV,most of them injection drug users: Possibilities and limitations.AIDS Care, 13(4), 467–473.

*46 Mathews, W. C., Mar-Tang, M., Ballard, C., Colwell, B.,Abulhosn, K., Noonan, C., et al. (2002). Prevalence, predic-tors, and outcomes of early adherence after starting or chang-ing antiretroviral therapy. AIDS Patient Care and STDS, 16(4),157–172.

*65 Melbourne, K. M., Geletko, S. M., Brown, S. L., Willey-Lessne, C., Chase, S., and Fisher, A. (1999). Medication ad-herence in patients with HIV infection: A comparison of twomeasurement methods. The AIDS Reader, 9(5), 329–338.Miller, L. G., and Hays, R. D. (2000). Adherence to com-bination antiretroviral therapy: Synthesis of the literature andclinical implications. The AIDS Reader, 10(3), 177–185.Miller, L. G., Hays, R. D. (2000). Measuring adherence toantiretroviral medications in clinical trials. HIV Clinical Trials,1(1), 36–46.

*29 Moatti, J. P., Carrieri, M. P., Spire, B., Gastaut, J. A.,Cassuto, J. P., and Moreau, J. (2000). Adherence to HAART inFrench HIV-infected injecting drug users: The contribution ofbuprenorphine drug maintenance treatment. The Manif 2000Study Group. AIDS, 14(2), 151–155.

*29 Morisky, D. E., Green, L. W., and Levine, D. M. (1986). Con-current and predictive validity of a self-reported measure ofmedication adherence. Medical Care, 24(1), 67–74.

*4 Murri, R., Ammassari, A., De Luca, A., Cingolani, A., Marconi,P., Wu, A. W., et al. (2001). Self-reported nonadherence withantiretroviral drugs predicts persistent condition. HIV ClinicalTrials, 2(4), 323–329.

*30 Nieuwkerk, P., Gisolf, E., Sprangers, M., and Danner, S. (2001).Adherence over 48 weeks in an antiretroviral clinical trial: Vari-able within patients, affected by toxicities and independentlypredictive of virological response. Antiviral Therapy, 6(2), 97–103.Nieuwkerk, P. T., and Oort, F. J. (2005). Self-reported adher-ence to antiretroviral therapy for HIV-1 infection and virologictreatment response: A meta-analysis. Journal of Acquired Im-mune Deficiency Syndromes, 38(4), 445–448.

*35 Nieuwkerk, P. T., Sprangers, M. A., Burger, D. M., Hoetelmans,R. M., Hugen, P. W., Danner, S. A., et al. (2001). Limitedpatient adherence to highly active antiretroviral therapy forHIV-1 infection in an observational cohort study. Archives ofInternal Medicine, 161(16), 1962–1968.

*10 Oyugi, J. H., Byakika-Tusiime, J., Charlebois, E. D., Kityo,C., Mugerwa, R., Mugyenyi, P., et al. (2004). Multiple vali-dated measures of adherence indicate high levels of adherenceto generic HIV antiretroviral therapy in a resource-limited set-ting. Journal of Acquired Immune Deficiency Syndromes, 36(5),1100–1102.

*47 Palepu, A., Horton, N. J., Tibbetts, N., Meli, S., and Samet,J. H. (2004). Uptake and adherence to highly active antiretrovi-ral therapy among HIV-infected people with alcohol and othersubstance use problems: The impact of substance abuse treat-ment. Addiction, 99(3), 361–368.Paterson, D., Potoski, B., and Capitano, B. (2002). Measure-ment of adherence to antiretroviral medications. Journal ofAcquired Immune Deficiency Syndromes, 31, S103–S106.Paterson, D., Swindells, S., Mohr, J., Brester, M., Vergis, E.,Squier, C., et al. (2000). Adherence to protease inhibitor ther-apy and outcomes in patients with HIV infection. Annals ofInternal Medicine, 133(1), 21–30.

Springer

AIDS Behav (2006) 10:227–245 245

Pinheiro, C. A., de-Carvalho-Leite, J. C., Drachler, M. L., andSilveira, V. L. (2002). Factors associated with adherence toantiretroviral therapy in HIV/AIDS patients: A cross-sectionalstudy in Southern Brazil. Brazilian Journal of Medical andBiological Research, 35(10), 1173–1181.

*31 Pradier, C., Carrieri, P., Bentz, L., Spire, B., Dellamonica, P.,Moreau, J., et al. (2001). Impact of short-term adherence onvirological and immunological success of HAART: A casestudy among French HIV-infected IDUs. International Journalof STD AIDS, 12(5), 324–328.

*38 Raboud, J. M., Harris, M., Rae, S., and Montaner, J. S. (2002).Impact of adherence on duration of virological suppressionamong patients receiving combination antiretroviral therapy.HIV Medicine, 3(2), 118–124.Rudd, P. (1979). In search of the gold standard for compli-ance measurement. Archives of Internal Medicine, 139(6), 627–628.Rudd, P., Ahmed, S., Zachary, V., Barton, C., Bonduelle, D.(1990). Improved compliance measures: Applications in anambulatory hypertensive drug trial. Clinical Pharmacology andTherapeutics, 48(6), 676–685.Samet, J. H., Sullivan, L. M., Traphagen, E. T., and Ickovics, J.R. (2001). Measuring adherence among HIV-infected persons:Is MEMS consummate technology? AIDS and Behavior, 5(1),21–30.Schroder, K. E., Carey, M. P., and Vanable, P. A. (2003).Methodological challenges in research on sexual risk behavior:I. Item content, scaling, and data analytical options. Annals ofBehavioral Medicine, 26(2), 76–103.

*36 Schuman, P., Ohmit, E., Cohen, M., Sacks, H., et al. (2001).Prescription of and adherence to antiretroviral therapy amongwomen with AIDS. AIDS and Behavior, 5(4), 371.

*7 Silveira, M. P., Draschler Mde, L., Leite, J. C., Pinheiro, C.A., and da Silveira, V. L. (2002). Predictors of undetectableplasma viral load in HIV-positive adults receiving antiretroviraltherapy in southern Brazil. The Brazilian Journal of InfectiousDiseases, 6(4), 164–171.

*20 Spire, B., Duran, S., Souville, M., Leport, C., Raffi, F., andMoatti, J. P. (2002). Adherence to highly active antiretroviraltherapies (HAART) in HIV-infected patients: From a predictiveto a dynamic approach. Social Science and Medicine, 54(10),1481–1496.Straka, R. J., Fish, J. T., Benson, S. R., and Suh, J. T. (1997).Patient self-reporting of compliance does not correspond withelectronic monitoring: An evaluation using isosorbide dinitrateas a model drug. Pharmacotherapy, 17(1), 126–132.

*32 Trotta, M. P., Ammassari, A., Cozzi-Lepri, A., Zaccarelli, M.,Castelli, F., Narciso, P., et al. (2003). Adherence to highlyactive antiretroviral therapy is better in patients receiving non-nucleoside reverse transcriptase inhibitor-containing regimensthan in those receiving protease inhibitor-containing regimens.AIDS, 17(7), 1099–1102.

Turner, B. (2002). Adherence to antiretroviral therapy by hu-man immunodeficiency virus-infected patients. The Journal ofInfectious Disease, 185(Suppl 2), S143–S151.Turner, B. J., and Hecht, F. M. (2001). Improving on a coin tossto predict patient adherence to medications. Annals of InternalMedicine, 134(10), 1004–1006.

*69 Vincke, J., Bolton, R. (2002). Therapy adherence and highlyactive antiretroviral therapy: Comparison of three sources ofinformation. AIDS Patient Care and STDS, 16(10), 487–495.

*64 Wagner, G. J. (2002). Predictors of antiretroviral adherence asmeasured by self-report, electronic monitoring, and medicationdiaries. AIDS Patient Care and STDS, 16(12), 599–608.

*11 Wagner, G. J., Kanouse, D. E., Koegel, P., and Sullivan, G.(2003). Adherence to HIV antiretrovirals among persons withserious mental illness. AIDS Patient Care and STDs, 17(4),179.Wagner, G., and Miller, L. G. (2004). Is the influence of so-cial desirability on patients’ self-reported adherence overrated?Journal of Acquired Immune Deficiency Syndromes, 35(2),203–204.

*21 Wagner, J. H., Justice, A. C., Chesney, M., Sinclair, G.,Weissman, S., and Rodriguez-Barradas, M. (2001). Patient-and provider-reported adherence: Toward a clinically usefulapproach to measuring antiretroviral adherence. Journal ofClinical Epidemiology, 54(Suppl 1), S91–S98.

*68 Walsh, J. C., Horne, R., Dalton, M., Burgess, A. P., andGazzard, B. G. (2001). Reasons for non-adherence to antiretro-viral therapy: Patients’ perspectives provide evidence of mul-tiple causes. AIDS Care, 13(6), 709–720.

*12 Walsh, J. C., Mandalia, S., and Gazzard, B. G. (2002). Re-sponses to a 1 month self-report on adherence to antiretroviraltherapy are consistent with electronic data and virological treat-ment outcome. AIDS, 16(2), 269–277.Waterhouse, D. M., Calzone, K. A., Mele, C., Brenner, D. E.(1993). Adherence to oral tamoxifen: A comparison of patientself-report, pill counts, and microelectronic monitoring. Jour-nal of Clinical Oncology, 11(6), 1189–1197.

*70 Weiser, S., Wolfe, W., Bangsberg, D., Thior, I., Gilbert, P.,Makhema, J., et al. (2003). Barriers to antiretroviral adherencefor patients living with HIV infection and AIDS in Botswana.Journal of Acquired Immune Deficiency Syndromes, 34(3),281–288.

*33 Wutoh, A. K., Brown, C. M., Kumoji, E. K., Daftary, M. S.,Jones, T., Barnes, N. A., et al. (2001). Antiretroviral adher-ence and use of alternative therapies among older HIV-infectedadults. Journal of the National Medical Association, 93(7–8),243–250.Wutoh, A., Elekwachi, O., Clarke-Tasker, V., Daftary, M.,Powell, N., and Campusano, G. (2003). Assessment and pre-dictors of antiretroviral adherence in older HIV-infected pa-tients. Journal of Acquired Immune Deficiency Syndromes,33(2), S106–S114.

Springer

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.