Decentralization of HIV care and treatment services
in Central Province, Kenya: Adult patient characteristics and
outcomesPresenting author: William Reidy, PhD
Reidy W, Hawken M, Wang C, Koech E, Elul B, and Abrams EJ
for the Identifying Optimal Models of HIV Care in Africa: Kenya Consortium
Background: Kenya
• Population: 38.6 million
• Adult HIV prevalence: 6.2%
• Living with HIV: 1.6 million
• Estimated annual number of newly infected: 100,000
• Number died of AIDS-related causes in 2011: 49,126
Background: Decentralization of HIV care in Kenya
• HIV care/ART in Kenya was provided in a small number of secondary health facilities (HF): – District, sub-district, provincial, or
teaching/national referral hospitals• Beginning in 2004, started scaling up HIV
clinics at smaller, primary HF:– Health centers and dispensaries
• Performance of primary HF during scale-up is not well-established
Objective• To compare the performance of
primary and secondary HF in Central Province, Kenya during a period of scale-up: –Patient volume–Patient and facility characteristics–Quality of care–Patient retention
Population and data sources• 37 of 52 government health facilities in
Central Province supported by ICAP at Columbia University via PEPFAR funding – 15 secondary and 22 primary HF
• Included patients enrolled between 2006-10 (N= 26,690)
• Data sources:– HIV care/ART data from patient-level databases
maintained by facility staff– Annual facility survey conducted by ICAP
Key variables and outcomes (1)
• Patient volume– Number of patients enrolled in HIV
care, by year• Patient characteristics
– Gender, age, WHO stage, CD4 count at enrollment and ART initiation
• Facility characteristics– Rural/non-rural, nurse ART provision,
CD4 machine on-site
Key variables and outcomes (2)
• Quality of care– Assessment of ART eligibility
(CD4/WHO), prompt ART initiation• Patient retention
1. Death: Recorded as dead in facility database
2. Loss to follow-up: Not dead, not transferred out, and not attending clinic for >6 months for patients on ART, or >12 month for pre-ART patients
Analytic Methods• Descriptive statistics• Kaplan-Meier survival curves• Competing risks regression (pre-
ART) and Cox proportional hazards regression (ART) Multivariate regression models
included: site type (primary vs. secondary HF), WHO stage, CD4 count, age group, gender, year of patient enrollment in care or ART initiation
Results
Patient volume
Enrollment in HIV care and treatment at primary and secondary HF
2006 2007 2008 2009 20100
1000
2000
3000
4000
5000
6000
7000
0
5
10
15
20
25
Primary HF Secondary HF
Num
ber o
f pati
ents
# of
faci
lities
Enrollment in HIV care and treatment at primary and secondary HF
2006 2007 2008 2009 20100
1000
2000
3000
4000
5000
6000
7000
0
5
10
15
20
25
Primary HFs Secondary HF
Num
ber o
f pati
ents
# of
faci
lities
# Primary HF
# Secondary HF
Facility characteristics
Clinic location, nurse ART provision, and presence of CD4
machine on-site
Loca
ted in
rura
l are
a
Nurse pr
ovisi
on of
ART
CD4 mac
hine o
n site
0%20%40%60%80%
100%
Primary Health Facilities Secondary Health Facilities
% o
f fac
iliti
es
Patient characteristics
Characteristics at enrollment in HIV care
Primary HF Secondary HF(n=3,881) (n=22,809)
Female 72% 69%Age group
15-20 2% 2%20-30 20% 23%30-40 43% 42%40+ 35% 34%
CD4 count 40% missing 41% missing<100 25% 31%
100-200 22% 22%200-350 22% 20%
350+ 31% 27%WHO stage 11% missing 24% missing
I/II 69% 60%III/IV 31% 40%
Point of entry to HIV care
Primary HF
Secondary HF
(n=3,881)
(n=22,809)
Transferred in 20% 12%VCT 19% 29%PMTCT 12% 9%TB/HIV 5% 6%PITC 3% 6%Unknown/other 41% 34%
Characteristics of patients starting ART
Primary HF Secondary HF
(n=2,391) (n=13,486)CD4 value at ART initiation
19% missing
18% missing
<100 32% 38%100-200 30% 30%200-350 32% 26%
350+ 7% 5%WHO stage at ART initiation
13% missing
18% missing
I/II 52% 52%III/IV 48% 48%
Quality of care: ART eligibility assessment
and prompt initiation
At enrollment in HIV care
Within first 3 months of HIV
care
Within first 6 months of HIV
care
0%
20%
40%
60%
80%
Percent with ART eligibility assessed by CD4 or WHO
stage
Primary HF Secondary HF
% o
f pa
tien
ts
0%
40%
80%
Percent of patients ART-eli-gible at enrollment who
started ART
Primary HF Secondary HF
% o
f pa
tien
ts
Patient retention: Death and loss to follow-up
(LTF)
Death following enrollment in HIV care (pre-ART)
Adjusted SHR=1.2995% CI: (0.91-1.84)
Adjusted SHR=0.77 95% CI: (0.62-0.97)
LTF following enrollment in HIV care (pre-ART)
Death following ART initiation
Adjusted HR=0.94 95% CI: (0.67-1.32)
LTF following ART initiation
Adjusted HR=0.67 95% CI: (0.27-1.65)
Adjusted S/HR of non-retention in Primary vs. Secondary HF
All patients
Adjusted S/HR* 95% CI
Pre-ART Death 1.29 0.91-1.84
LTF 0.77 0.62-0.97
ART Death 0.94 0.67-1.32
LTF 0.67 0.27-1.65*Reference category: Secondary HF. Models control for WHO stage,
CD4 count, age group, gender, year of patient enrollment in care or ART initiation
Adjusted S/HR of non-retention in Primary vs. Secondary HF
Sensitivity analysis excluding transfer-in patients
All patientsExcluding transfer-in patients
Adjusted S/HR* 95% CI Adjuste
d S/HR* 95% CI
Pre-ART Death 1.29 0.91-1.84 1.32 0.92-
1.89LTF 0.77 0.62-
0.97 0.84 0.66-1.07
ART Death 0.94 0.67-1.32 0.94 0.65-
1.35LTF 0.67 0.27-
1.65 0.72 0.28-1.82
*Reference category: Secondary HF. Models control for WHO stage, CD4 count, age group, gender, year of patient enrollment in care or ART initiation
Summary• Patient enrollment at primary HF
increased dramatically during the period
• Patients enrolling in primary HF were somewhat healthier by WHO stage, CD4 count
• Quality of patient care and retention were comparable at primary and secondary HF – Among pre-ART patients, the rate of
LTF was lower at primary than at secondary facilities
• Primary HF have performed well within the context of decentralization in Central Province, Kenya
Acknowledgements• Kenya Ministry of Health• Government staff at the 37 facilities• ICAP staff in Kenya and in New York
– Dr. Muhsin Sheriff (Kenya), Mansi Agarwal (NY)
• US Centers for Disease Control and Prevention
• The President’s Emergency Plan for AIDS Relief
• This research was supported by PEPFAR through the CDC under the terms of Cooperative Agreement Number 5U62PS223540 and 5U2GPS001537