Content uploaded by Alula M Teklu
Author content
All content in this area was uploaded by Alula M Teklu on Jan 03, 2017
Content may be subject to copyright.
197
Alula Meressa Tekelu, Kesetebirhan Delele Yirdaw. Ethiop Med J, 2016, Vol. 54, No. 4
ORIGINAL ARTICLE
ART EXPERIENCED PATIENTS FOR TACKLING ATTRITION
FROM HIV CARE: A MULTI-SITE COHORT STUDY
Alula Meressa Tekelu, MD, MPH
1*
, Kesetebirhan Delele Yirdaw, MD, MPH
1
ABSTRACT
Introduction: Retention of patients on anti-retroviral treatment in Ethiopia is a challenge. Use of anti-retroviral
treatment experienced patients to prepare and re-engage them when they miss follow-ups is recommended, but
evidence on its effectiveness is limited. This study evaluated its effectiveness.
Methods: A retrospective cohort study in 10 randomly selected health facilities was conducted to compare out-
comes before and after initiation of the adherence supporters program in HIV care and treatment from September
2001 to August 2013. Data analysis involved Kaplan-Meier survival and Log-rank test analysis on STATA statisti-
cal software Version 12 to compare survival experiences.
Results: Of 18,835 records that were available, 938 (4.36%) records with missing values were excluded and data
from the remaining 17,897 was analyzed. The incidence of first instance lost to follow-up was 22.2 per 100 person-
years (95% confidence interval 21.7-22.7). The risk of missing follow-ups after initiation of the program was high
(Hazard Ratio –1.22, P < 0.001). The incidence of restarting after missed follow-ups was 23 per 100 PY (95% CI
22.2-24.0). The likelihood of restarting after missed follow-ups was four times higher during the period adherence
supporters were present (P<0.001). Patients who stayed longer in care before missing follow ups were more likely
to restart (5.7 times the chance of restarting treatment for those whose first lost to follow-up occurred at≥12
months compared to <3 months, P< 0.001).Time to restarting treatment was shorter after the initiation of the ad-
herence supporters program (median 37 vs. 115 days). The risk of recurrence of being lost to follow-up in the
presence of adherence supporters was significantly higher than when there were no adherence supporters; 38.8
(95% CI 36.3-41.6) per 100 PY vs. 26.1 (95% CI 19.8-34.4) per 100 PY, respectively.
Conclusion: Adherence supporters were effective in improving re-engagement of patients in treatment and care
after they were lost to follow-up. Yet, prevention of lost to follow-up cases has remained a challenge to the pro-
gram.
Key Words: Adherence, adherence supporters, lost to follow-up, HIV, antiretroviral treatment
1
MERQ Consultancy PLC
* Corresponding author: ateklu72@gmail.com
INTRODUCTION
The HIV epidemic has affected millions and has left
close to a million people currently living with the
virus in Ethiopia (1,2). The country’s response to
HIV has shown remarkable results. The number of
new infections has declined and mortality has gone
down. Globally, the reduction in mortality goes as
high as 65% in certain settings (3,4).
The success of antiretroviral therapy (ART) can be
attributed to the increased access to the service by
those who need it. There has been a massive increase
in the number of people living with HIV/AIDS
(PLHIV) who are taking ART (4). With 79% cover-
age of eligible patients (CD4<350 or WHO Stage III
or IV), Ethiopia has put 492,649 patients on ART by
the end of 2013. Of the 492,649 patients, only
344,344 are currently on ART, which shows a 30%
difference. The difference was 23% four years ear-
lier. (5,6) With increasing number of patients on
ART, retention has continued to be an increasing
challenge (7).
Ensuring adherence to treatment and care will be a
major challenge for the health system if it continues
to work exclusively. Involving PLHIV meaningfully
in HIV services planning and delivery was found to
be essential, and many PLHIV have been involved in
the provision of support to fellow patients. A case in
198
point is the involvement of ART experienced PLHIV
in the provision of support to other patients in HIV
care. The national program identifies these cadres as
adherence supporters and they have been providing
services for the previous six years before this evalua-
tion.
Given the shortage of human resources for health
(HRH) and with intention of ensuring greater and
meaningful involvement of PLHIV in HIV care, the
Federal HIV/AIDS Prevention and Control Office
(FHAPCO) of Ethiopia approved the notion of in-
volving ART experienced patients in the provision of
adherence counseling, tracking and restarting pa-
tients who missed follow up (LTFU) and helping
patients who have adherence/retention related chal-
lenges, as well as facilitating linkage of newly diag-
nosed HIV positive patients to HIV care (8).
The national adherence supporters program was
launched in 2007 in the country. It was introduced in
all regions within a few months’ time in 2007. An
implementation guideline with the details of the ac-
tivities that are expected from the adherence support-
ers was issued by the Federal Ministry of Health
(FMOH) (8). Supplementary documentation tools
were developed including registration of patients
who were LTFU and used to identify those who were
LTFU and forms to track those who were LTFU.
Though there have been some efforts to ascertain
effectiveness of the program, this has not been fully
and systematically assessed. This study has been
conducted to fill this gap by collecting and analyzing
data from ten randomly selected hospitals in four
regions of Ethiopia.
METHODS
Study Setting: This is a retrospective, observational,
longitudinal study that used existing records of pa-
tients in ten randomly selected hospitals located in
four regions of Ethiopia, namely Addis Ababa, Ben-
ishangul Gumuz, Gambella, and Southern Nations
Nationalities and Peoples (SNNP) Region. These
regions were selected because of their use of patient
monitoring electronic data management tool which
was similar across these regions.
The adherence supporters program was initiated in
April 2007 and is still being implemented. Adher-
ence supporters were selected using criteria which
were approved by FHAPCO and all were trained
using a standard material from the same source (8).
Those who managed to complete the training were
labeled as “adherence supporters” and were deployed
to facilities. The program was started in all hospitals
providing antiretroviral therapy in 2007. The number
of adherence supporters per facility varied depending
on patient load; the assignment used one adherence
supporter for 800 patients on HIV care as a standard.
Study Population: The study population included
patients taking ART in public and non-governmental
(NGO) hospitals in the selected regions. Of the 36
hospitals in these regions, ten were randomly se-
lected using a lottery method. All patients enrolled in
HIV care and treatment program started on ART
from September 2001 to August 2013 were included
in the study. Patients who were transferred from
other facilities and enrolled in the selected hospitals
for continuation of ART were excluded.
Variables and Source of Data: There were two out-
come variables. The first was ‘lost to follow-up’
which is defined as not present for more than one
month since the last appointment date for ART medi-
cation refill. The second outcome was ‘restarting
treatment’; this is whenever patients who missed
follow-ups restart ART after having been labeled as
“lost to follow-up”. The presence of adherence sup-
porters was the primary exposure variable. Other
explanatory variables to potentially affect the out-
come included age, gender, baseline World Health
Organization (WHO) staging, baseline CD4 cell
count, and time to first loss of follow-up were ex-
tracted from existing records.
Data Analysis: Epi Info statistical software Version
3.4 and MS Access 2007 were used to clean the data.
Descriptive analysis and survival analysis were car-
ried out using STATA statistical software Version
12. Log-rank test was used to compare survival ex-
periences. Patients were uncensored at the first epi-
sode of loss to follow-up. Those who stopped treat-
ment after medical consultation and those who were
transferred to another facility were censored. Patients
who were enrolled and started on ART before 2007
did so before the adherence supporters were de-
ployed and hence formed a historical control group.
These patients were censored on July 31, 2007 if
they were actively on follow-up until that time.
For patients who started treatment after October 1
2007, adherence supporters were available to provide
additional counseling as well as patient tracking;
hence they formed the exposure group. These groups
of patients were censored on the last day of follow-
up at the end of 2013 if they remained in follow-up
199
in their respective facilities. Because of some lack
of uniformity in the deployment of adherence sup-
porters which took sometimes before they were
fully functional, patients enrolled in that adjustment
period between August and September 2007 were
excluded from the analysis. Potential within-site
correlation of patient characteristics was controlled
by stratification using site unique identity (ID) num-
bers as cluster identifier. Records having at least
one missing variable were excluded from analysis.
Ethics Consideration: Ethics approval for the study
was obtained from the National Research Ethics
Review Committee (NRERC) of Ethiopia. Patients’
consent forms were not required since existing de-
identified and de-linked data was used before acqui-
sition and during analysis.
RESULTS
Baseline Characteristics: There were 18,835 re-
cords (excluding patients who were transferred in)
out of which some had variables with missing val-
ues. After excluding all records with at least one
missing variable, there were 17,897 records (95%).
Overall, patients were observed for a minimum of
one day, and a maximum of 6.1 years. The median
time of follow-up was 1.25 years.
The median age (interquartile range (IQ)) was 34
(28, 40), and the minimum, and maximum ages 3
months and 86 years respectively. Pediatric patients
below 15 years of age were 2% while the majority
(90%) was between ages 15 and 49 years. Females
accounted for 54%. Patients with WHO stage III or
IV constituted 67%, while those with baseline CD4
cell count of less than 200 cells/mm
3
were 78%. In
the period where adherence supporters were present,
there were 10,265 patients accounting for 57%
(Table 1).
Prevalence and Incidence of Loss to Follow-up:
Thirty seven percent of patients (n=6,595) were
LTFU at least once. Patients were followed for a
total period of 29,696 person-years. This makes the
incidence of LTFU 22 per 100 person-years of fol-
low-up (95% CI:21.7-22.7).
Predictors of LTFU: Figure 1 shows the survival
experience of patients with respect to presence or
Table 1. Baseline characteristics of patients on ART (N=17,897), 2001-2013, Ethiopia
Variable Category Total
(Column %)
Adherence supporters
Present Not present
Number (row %) Number (row %)
Age <15 269 (2) 248 (92) 21 (8)
15-49 16138 (90) 9241 (57) 6897 (43)
>49 1490 (8) 776 (52) 714 (48)
Sex Female 9722 (54) 5970 (61) 3752 (39)
Male 8175 (46) 4295 (53) 3880 (47)
Baseline WHO Stage I or II 5967 (33) 4431 (74) 1536 (26)
III or IV 11930 (67) 5834 (49) 6096 (51)
Baseline CD4 Cell Count/mm3 <200 13907 (78) 7038 (51) 6869 (49)
200-349 3496 (19) 2819 (81) 677 (19)
>349 494 (3) 408 (83) 86 (17)
Ever lost to follow-up? No 11302 (63) 6054 (54) 5249 (46)
Yes 6595 (37) 4211 (64) 2383 (36)
200
Figure 1: Kaplan Meier Survival for being lost to follow-up among patient on ART,
2001-2013, Ethiopia
It can be seen that, in general, LTFU was greater for
both cases (with or without adherence supporters) in
the first year of follow-up. Patients continued to be
LTFU after that period, but the rate was much
slower, a drop in survival of less than 10% per year
as compared to 30% for the first year. The overlap
between the two curves in the first year of follow-up
indicates that the probability of remaining in care
was comparable for the two groups during the early
years of ART program initiation when adherence
supporters were non-existent. After that period, pa-
tients enrolled in the presence of adherence support-
ers were less likely to remain in care. Table 2 below
summarizes the effect of all variables on the hazard
of being LTFU.
During the period that adherence supporters were
present, the relative hazard of becoming LTFU in-
creased by 22%. Being male and being in WHO
stage III or IV were also associated with increased
relative hazard of being lost to follow-up (p value
<0.05). Age was another predictor but was signifi-
cant only at the 0.1 level. Adult patients were at an
increased relative hazard of being LTFU as com-
pared to pediatric patients. No association was found
between LTFU and CD4 cell count category.
Prevalence and incidence of restarting treatment
after first LTFU: Among patients LTFU (n=6,595),
2,221 (34%) patients restarted treatment (Table 3).
LTFU patients were followed for a total time of
9,617 person-years until last observation. The inci-
dence of restarting treatment was 23 per 100 person-
years (95% CI: 22.2-24.0). Overall, final status was
known for 3,568 (54%) of the patients LTFU. Of
these, 1,347 (38%) patients had died. During the time
adherence supporters were present, the proportion of
patients LTFU with unknown status at the time of
censoring was lower compared to the period they
were not present, 37% vs 62% (p value <0.001).
Time to restarting treatment was shorter after initia-
tion of the adherence supporters program, median 37
vs 115 days (p value <0.001). The risk of recurrence
of being lost to follow-up while with availability of
adherence supporters was significantly higher than
when there were no adherence supporters, 38.8 (95%
CI 36.3-41.6) per 100 PY vs 26.1 (95% CI 19.8-34.4)
per 100 PY.
Predictors of restarting treatment among those
LTFU: Table 4 summarizes factors affecting the
restarting of treatment among those who were LTFU.
201
Table 2. Predictors for being lost to follow-up of patients on ART after the first six months
(n= 12,123) 2001-2013, Ethiopia
* Stratified by site
Table 3. Tracing outcome of patients lost to follow-up from ART (first loss only), 2001-2013, Ethiopia
Table 4. Predictors for restarting treatment for patients lost to follow-up from ART,
2001-2013, Ethiopia (n=6,595)
Variable Category Hazard ratio,
Crude*
P value Hazard
ratio*, adj
P value
Age <15 1 0.0735 1 0.059
15-49 1.38 1.38
>49 1.39 1.41
Sex Female 1 0.0012 1 0.0001
Male 1.13 1.16
Baseline WHO Stage I or II 1 <0.0001 1 <0.0001
III or IV 1.18 1.20
Baseline CD4 Cell Count/mm
3
<200 1 0.5532
200-349 1.05
>349 1.00
Adherence supporters present? No 1 <0.0001 1 <0.0001
Yes 1.22 1.23
Total
First LTFU
Number
Dead (%)
Number
Restarted (%)
Number
Still lost at the end of
observation (%)
No 2383 687 (29) 215 (9) 1481 (62)
Yes 4212 660 (16) 2006 (47) 1546 (37)
Total 6595 1347 (20) 2221 (34) 3027 (46)
Variable Category
Hazard ratio,
crude P value
Hazard
ratio, adj P value
Baseline WHO Stage I or II 1 <0.0001 1 0.0015
III or IV 0.61 0.87
Baseline CD4 Cell Count/mm
3
<200 1 <0.0001 1 <0.0001
≥200 1.75 1.35
Time to 1
st
LTFU <3 months 1 <0.0001 1 <0.0001
3-5 months 1.81 1.44
6-8 months 2.23 1.67
9-11 months 3.04 2.26
>=12 months 4.44 2.88
Adherence supporters Present 1 <0.0001 1
Not present 5.17 2.84 <0.0001
202
Those who died were excluded from the analysis. It
can be seen that, all variables had an association with
resumption of treatment by patients. The relative
likelihood of restating treatment among those LTFU
during the period adherence supporters were present
was three times as high as that of patients seen when
adherence supporters were not present (Figure 2).
During the time adherence supporters were present,
restarting treatment was much better. Most restarted
treatment in the first year after being lost to follow-
up. This levels of in later years.
In addition, being in the pediatric age group (< 15
years), being female, having a WHO stage I or II
disease and having a higher baseline CD4 cell count
category were all associated with a higher relative
chance of restarting treatment (p value <0.05). Also,
restarting treatment was more likely if treatment was
discontinued much later during follow-up than soon
after ART initiation (Figure 3).
Those who missed follow-ups early after starting
antiretroviral therapy were less likely to restart as
compared to those who became lost later during fol-
low-up. Those confirmed to have died were excluded
from the analysis.
Figure 2. Cumulative hazard for restarting treatment among patients lost to Follow-up
from ART by presence of adherence supporters, 2001-2013. Ethiopia.
203
Figure 3: Cumulative hazard for restarting treatment among patients lost to follow-up
by time to first LTFU, 2001-2013.
DISCUSSION
Over one-third of the patients started who were on
ART were LTFU at least once during the follow-up.
The peer support program was effective in improving
re-engagement, but it did not reduce the magnitude
of those absent for the first time. It also improved the
proportion of patients whose final status was known.
A number of studies have reported on the level of
treatment discontinuation ranging between 20% and
30% (9-12), which is less than what is identified in
this study of 37%. The reason for the difference is
that the present study assessed the proportion of
those who discontinued treatment for at least one
month any time during follow-up while previous
studies reported on final follow-up status at the time
of study.
A study in Tanzania that used a definition for LTFU
similar to our study found that LTFU was 8% among
12,000 followed-up patients (13). The proportion of
LTFU in the Tanzanian study was much smaller be-
cause patients were followed up for only six months.
The study also revealed that peer supporters were
able to re-engage 38% of those LTFU which is simi-
lar to the finding in the present study. A good num-
ber of patients could not be restarted because some
had died and others were not traceable due to wrong
addresses or reside far away the city where treatment
is provided. In some cases patients change treatment
centers without prior notice to health care providers
or even started traditional treatment (9, 13).
The first year after initiation of ART was when most
patients were LTFU. This is the time when most
treatment related side effects occur, in addition to the
occurrence of opportunistic infections and immune
reconstitution inflammatory syndrome (IRIS) and a
good proportion of the patients die. These factors add
to existing barriers for adherence such as problems in
access, lack of family or social support and stigma
(14). Treatment supporters were not able to prevent
LTFU in our series.
In fact, the occurrence of being LTFU was worse
when adherence supporters were present. This may
partly be explained by the improvement in documen-
tation of patients’ final status as the adherence sup-
204
porters used documentation tools and actively looked
out for patients who were LTFU. This may also be
due to problems in the continuum of care with re-
spect to access for quality ART services, and use of
tools to guide counseling. As with other chronic ill-
nesses, patients may also discontinue treatment
whenever they feel better. HIV/AIDS is also progres-
sively becoming less stigmatizing and as patients
learn that it is manageable they become reluctant to
adhere to treatment plan (15).
The increase in the absolute number of patients
LTFU could very well be a reflection of facts stated
above. These facts may also make it more difficult to
prevent LTFU than in the earlier period when ART
was initially being rolled out. Moreover the work
burden on care providers is ever increasing because
of the rising number of PLHIV placed in care and
treatment program. Children, female patients, pa-
tients with less advanced disease, and those being
LTFU after three months of follow-up were more
likely to restart treatment. This is in line with the
earlier observation indicating that people are more
vulnerable to illness and side effects in the early
months after treatment initiation and are more likely
to discontinue treatment.
This study has several implications for clinical prac-
tice and policy. With the current level of implemen-
tation despite the gains made, a significant propor-
tion of clients remain LTFU. That implies additional
early tracing options need to be sought especially for
those residing far away from treatment facilities.
This could be in the form of engaging health exten-
sion workers or other community agents in tracing
activities, which calls for a more robust referral net-
working system involving health posts, the lowest
health units. Continuing to reduce patient load in big
treatment centers, mostly hospitals, by offloading
stable patients to health centers not only decrease the
burden of work on care providers, but also reduce the
time required by patients to travel and better access
care and treatment (9,16).
The present study also indicates that it is important to
focus counseling and tracing efforts in the first few
months after starting treatment when most of patients
will become LTFU. Particular attention should be
directed towards the most vulnerable which may
include the very sick, those lacking social support,
and those coming from remote areas. The adherence
supporter to patient ratio may also need revision as
the current ratio of 1:800 is low taking into consid-
eration the burden of counseling needs. Healthier
clients are starting treatment according to the revised
ART treatment guidelines (CD4≤500) which may
make it harder to convince individuals about the
benefits of sustained treatment and treatment adher-
ence, adding to existing challenges (17). A custom-
ized adherence protocol focusing on patients who
start ART at higher CD4 levels is required. Strength-
ening the involvement of adherence supporters in
multi-disciplinary teams supplemented by structured
guide (cue cards) is also essential.
The fact that this was a multicenter study involving a
large cohort of clients in four regions of Ethiopia
spanning over a decade of follow-up, makes the find-
ings very informative. Yet, because of lack of pri-
mary data, this is an over simplified description that
does not show much of the effort put to address
LTFU, including the effort by adherence supporters
and the multi-disciplinary ART team in adherence
preparation, and ongoing counseling for those with
poor treatment adherence. The adherence counselor-
patient ratio was also very low at around 1:800.
Therefore, data on which clients actually received
support from the counselors would have provided
further opportunity to identify their roles in prevent-
ing patient loss and re-engaging those lost to follow-
up.
A separate study specifically designed to look at the
detailed activities of adherence supporter program is
highly needed. Lastly, the interpretations of findings
must be made carefully since all limitations that ap-
ply to the use of historical controls apply in this
study. Also, the fact that the study spans over a long
period means that progressive changes in quality of
care may influence study outcomes.
Conclusions: Adherence supporters were effective in
improving re-engagement, but preventing LTFU re-
mains a challenge.
ACKNOWLEDGMENT
We thank all health care providers who remain dedi-
cated to provide HIV/ART services. The electronic
data management tool used at the four facilities was
established by Johns Hopkins University Technical
Support for the Ethiopian HIV/AIDS ART Initiative
using funds from PEPFAR obtained through CDC-
Ethiopia. We would also like to thank staff working
on data at the various facilities and at the Regional
Health bureaus.
205
REFERENCES
1. AIDS info. Guidelines for the Use of Antiretroviral Agents in HIV-1-Infected Adults and Adolescents. Ac-
cessed at http://aidsinfo.nih.gov/guidelines2013.
2. Ethiopian Health and Nutrition Research Institute. HIV Related Estimates and Projections for Ethiopia. 2012.
3. Joint United Nations Program on HIV/AIDS. Global report: UNAIDS report on the Global AIDS Epidemic.
2013.
4. Joint United Nations Program on HIV/AIDS. Together we will end AIDS. 2012. Accessed at http://
w w w . u na i d s . o r g / e n / me d i a /u n a i d s / c o n t e n t a s s e t s / d o c u m e n t s / e p i d em i o l o g y / 2 0 1 2
JC2296_UNAIDS_TogetherReport_2012_en.pdf
5. Federal HIV/AIDS Prevention and Control Office. Country progress report on the HIV response, 2014. 2014.
6. Federal HIV/AIDS Prevention and Control Office & Federal Ministry of Health. Country Progress Report on
HIV/AIDS Response. Ethiopia. 2012.
7. Assefa Y, Kiflie A, Tesfaye D, Mariam DH, Kloos H, Edwin W, et al. Outcomes of antiretroviral treatment
program in Ethiopia: retention of patients in care is a major challenge and varies across health facilities. BMC
health services research. 2011;11(1):81.
8. Federal Ministry of Health &Federal HIV/AIDS Prevention & Control Office. Guidelines for Greater Involve-
ment of People Living With HIV/AIDS (GIPA) in Ethiopia. 2009.
9. Mulissa Z, Jerene D, Lindtjørn B. Patients present earlier and survival has improved, but pre-ART attrition is
high in a six-year HIV cohort data from Ethiopia. PloS one. 2010;5(10):e13268.
10. Berheto TM, Haile DB, Mohammed S. Predictors of loss to follow-up in patients living with HIV/AIDS after
initiation of antiretroviral therapy. North American Journal of Medical Sciences. 2014;6(9):453.
11. Yirdaw KD, Jerene D, Gashu Z, Edginton M, Kumar AM, Letamo Y, et al. Beneficial Effect of Isoniazid Pre-
ventive Therapy and Antiretroviral Therapy on the Incidence of Tuberculosis in People Living with HIV in
Ethiopia. PloS one. 2014;9(8):e104557.
12. Rachlis B, Ochieng D, Geng E, Rotich E, Ochieng V, Maritim B, et al. Implementation and Operational Re-
search: Evaluating Outcomes of Patients Lost to Follow-up in a Large Comprehensive Care Treatment Pro-
gram in Western Kenya. JAIDS. 2015;68(4):e46-e55.
13. Bupamba M, Mbatia, R, Strachan, M et al. Ambassadors for adherence: provision of highly effective defaulter
tracing and re-engagement by peer educators in Tanzania. XVIII International AIDS Conference, Vienna Aus-
tria, 18–23 July 2010 (Abstract MOAE0303). 2010.
14. World Health Organisation. Consolidated guidelines on the use of antiretroviral drugs for treating and prevent-
ing HIV infection: recommendations for a public health approach. Geneva, Switzerland. 2013.
15. World Health Organization. Adherence to long term therapy: evidence for action. Geneva, Switzerland. 2003.
16. Southern Nations Nationalities, & Peoples Regional State Health Bureau. Standard Operating Procedures for
Comprehensive HIV/AIDS Prevention, Care, Treatment and Support Services in SNNP Region. 2014.
17. Federal
Ministry of Health. National comprehensive HIV care and treatment training for health care providers,
participant manual, Ethiopia. 2014.
