Transcript Presentation at the 4th International Adherence

Impact of Electronic Drug
Monitoring Feedback on Adherence
to Antiretroviral Therapy
April 6, 2009
Lora Sabin
Center for International Health
and Development
Boston University
1
China Adherence For Life (AFL)
study collaborators
Boston University SPH
• Lora Sabin, MA, PhD
• Christopher J. Gill, MS, MD
• Mary B. DeSilva, MS, ScD
• Davidson H. Hamer, MD
Tufts-New England Medical
Center
• Ira Wilson, MS MD
Ditan Hospital, Beijing
• Xu Keyi, MD
Dali Second People’s
Hospital, Dali
• Zhang Jianbo, MD
Horizon Research Group,
Beijing
• Yuan Yue, MA, PhD
• Fan Wen, MA
• Li Tao, MA
Funding provided by: USAID, WHO/Beijing, US CDC
Additional acknowledgments:
Don Thea, Jon Simon, Deirdre Pierotti, Mini Singh, Anna Knapp, James Chen, Wanju Wu, Guo Jianhua, Matt Bobo, Ahmar Hashmi, and Jordan Tuchman
Background
• China is rapidly scaling up ART, but treatment
programs are at an early stage:
• Little is known about levels of adherence,
particularly among IDUs and former IDUs
• Little is understood about how to improve
adherence
• Drug resistance is rising, and there are fears about
the cost and availability of 2nd and 3rd line regimens
• As in other countries, there is an urgent need
for interventions that are effective in improving
adherence among HIV-positive patients
3
The relationship between ART adherence
and HIV outcomes was deduced using
electronic drug monitors (EDM)
• EDM pill bottles have an
embedded microchip in the
cap
– Time/date stamps each
bottle opening
– Surrogate marker for
adherence
• Comparative studies show
that EDM are by far the
best measure of adherence
available.
STUDY QUESTION:
Can we improve adherence
to ART using Electronic
Drug Monitor (EDM)
feedback?
Overview of AFL
(Control)
Continued passive
observation
(Intervention)
N=80
Patients enrolled
N=68
Patients randomized
Active EDM
feedback
Phase I
Phase II
Phase III
6 months
6 months
6 months
Qualitative investigations on
Adherence observed
Randomized controlled trial to
what patients/doctors in Dali
prospectively via EDM,
determine effectiveness of EDM
view as key barriers to
relationship between barriers and
feedback strategy
adherence
actual adherence, clinical
outcomes measured
6
Study site, Dali, Yunnan Province
Dali
Yunnan province
Study population
• HIV epidemic driven by injectable drug use
• Lesser contribution from commercial sex
work
• Minimal spread into larger population
AFL Study objectives
1. Primary Objective
 To determine effect of EDM feedback on adherence
rates
2. Secondary Objectives
 To determine effect of EDM feedback on CD4-cell counts
and undetectable viral loads (UDVL)
The study was powered to detect a 15% difference in
adherence rates, as assessed by EDM
Randomization Procedure
• Block stratified randomization
• At end of Phase I, patients stratified by ‘high’ or
‘low’ adherence
• ≥95% = ‘high adherence’
• <95% = ‘low adherence’
• Based on average adherence during the 5 months prior
to randomization
• Equal numbers of patients allocated from within
each adherence stratum
• Ensured balanced allocation at start of intervention
What happened in intervention
group?
• EDM data reviewed at each monthly study visit
• Patients with <95% adherence by EDM in previous month
flagged for “additional adherence counseling”
• EDM report given to doctor and patient at each visit
• % doses taken
• % on time
• Histogram readout
• Additional counseling had no fixed script
• involved a conversation between doctor and patient in
which doctor asked about problems or challenges,
referring to EDM print-out
What happened in control group?
• Self-report data reviewed at each monthly visit
• EDM data not provided to doctor/patient
• Patients with <95% adherence by self report in
previous month flagged for “additional adherence
counseling”
• Like intervention arm, additional counseling involved a
conversation in which doctor asked about problems or
challenges faced, referring to patient’s self-report
Definition of Primary Outcome
Metric
Composite EDM measure
includes proportion taken and timing of
doses:
# doses taken +/- 1 hour of scheduled time
# prescribed doses
Clinical measures
•
•
CD4-cell count
Undetectable Viral load (UDVL)
(Using RT PCR: <400 copies/ml =
“undetectable”
RESULTS
Patient Characteristics at randomization (Mo. 6)
Intervention
Characteristic
Gender
Male
Female
Number (%) Mean (SD)
25 (74)
9 (26)
Age (Mean, SD)
Control
Number (%) Mean (SD)
25 (74)
9 (26)
36.1 (8.3)
35.1 (8.0)
Education*
Elementary
Junior high
Senior high/technical school
7 (21)
17 (50)
10 (29)
13 (38)
20 (59)
1 (3)
Marital status
Single
Married
15 (44)
19 (56)
16 (47)
18 (53)
Ethnic background
Han Chinese
Bai
Other
18 (53)
14 (41)
2 (6)
15 (44)
17 (50)
2 (6)
Household size
Employment status
Currently employed
Currently unemployed
4.2 (1.5)
10 (31)
22 (69)
* Statistically significant at the p<0.01 level
4.5 (1.4)
12 (37)
20 (63)
16
Patient Characteristics at randomization (Mo. 6)
Characteristic
Heroin use in previous 3 mos
Yes
No
Intervention
Control
Number (%) Mean (SD)
Number (%) Mean (SD)
4 (12)
30 (88)
Depression (Beck's, continuous)
Depression (Beck's, binary)
Yes
No
9.8 (3.6)
6 (18)
27 (82)
CD4, Month 6 (continuous)
UDVL, Month 6 (yes)
5 (15)
29 (85)
10.2 (4.0)
10 (30)
23 (70)
297 (145)
357 (196)
30 (88.2)
28 (87.5)
16 (47)
18 (53)
17 (50)
17 (50)
Mean adherence, Months 0-5**
High (>= 95%)
Low (<95%)
** basis for block randomization procedure
17
Point Adherence at Months 6 and 12
*p<0.05
** p<0.01
At Month 6, no significant differences between intervention
and control groups
At Month 12, large increase in adherence in intervention arm;
no significant increase in control arm.
Mean adherence over time, periods 1 and 2
** p<0.01
At Month 6, no significant differences between intervention and
control groups (in Months 1-6 adherence)
Large increase in adherence in Months 7-12 in intervention arm;
no significant increase in control arm.
Achievement of mean adherence
≥95% throughout Months 7-12
Intervention
n/N (%)
Control
n/N (%)
23/31 (74)
11/33 (33)
RR = 2.23
(95% CI 1.3-3.8)
***p=0.001
Composite Adherence by group and time
100%
95%
90%
Adherence
85%
80%
75%
70%
65%
60%
Low adherers, intervention group
Low adherers, control group
55%
High adherers, intervention group
High adherers, control group
50%
1
2
3
4
5
6
7
Month
8
9
10
11
12
Clinical outcomes: Changes in CD4-cell
counts between months 6 and 12
Intervention
No. (%)
Control
No. (%)
22/31
(71%)
15/31
(48%)
Proportion with CD4
increase, months 6-12
RR 1.5 (1.0-2.2)
p=0.072
Mean change in CD4
(x1000 cells/ml)
+ 90
p=0.020
Note: regarding UDVL: little change from Month 6
-9
Patient-level EDM view:
A near perfect patient profile
Patient-level EDM view:
A patient with poor adherence
Patient-level EDM view:
A patient with improved adherence
Intervention
phase
6 months
Pre-intervention
phase
Main Findings
• EDM feedback improved ART adherence
• Adherence rise was prompt and sustained
• Intervention arm: adherence improved
• Control arm: adherence stayed steady with a falling trend
• Effect seen in both Month 6 v. Month 12 point comparisons and in
pre-intervention v intervention phase comparisons
• Patients more likely to achieve ≥95% adherence
• EDM feedback improved clinical outcomes
• CD4-cell counts rose significantly
• Trend towards higher proportion of rising CD4s among
intervention arm
• EDM feedback is a promising intervention – it warrants
further evaluation in other populations
Thank you for
your attention
Any
questions?