poor linkage to care despite significant improvement in access to early cart – data from test and...
TRANSCRIPT
Poor linkage to care despite significant improvement in access to early cART – data from Test and Keep in Care (TAK) project.
Leah Shepherd, Magdalena Ankiersztejn-Bartczak, Aneta Cybula, Hanna Czeszko-Paprocka, Ewa Firląg-Burkacka, Andrzej Horban, Amanda Mocroft and Justyna D. Kowalska
Test and Keep in Care
Hospital for Infectious Diseases, HIV Out-Patients Clinic, Warsaw, Poland, Medical University of Warsaw, Department for Adults' Infectious Diseases, Warsaw, Poland
So…where we are?
Unemployment Avarage salary Migration (saldo)
Background - Europe Up to 50% of HIV-infected remain undiagnosed
Half of newly diagnosed are late presenters
Increased HIV testing and improving care pathway are strongly encouraged
Only persons linked to care can fully benefit from being diagnosed
It is unknown how many diagnosed patients remain outside care
Antinori A. Antivir Ther 2010, Mocroft A. PLoS 2013, Wynberg E JIAS 2014
Test and Keep in Care (TAK) project
Prospective follow-up of HIV persons diagnosed in CBVTC in central Poland
Data collection: pre-clinical from CBVTC questionnaire (standardized by National
AIDS Centre) clinical from HIV clinics
Main end-point: linkage to care
Study aims: estimate the prevalence of HIV(+) lost or late to care investigate related factors and target effective interventions
Ankiersztejn-Bartczak M. HIV Med. 2015
Methods
data from CBVTC (2010-2013/14) and HIV clinics
were linked using unique Western-Blot number
linked to care = tested at CBVTC and registered in
HIV clinics
latest follow up date was 4/06/2014
Counselling
ELISA + result not collected by CBVTC client
Lost to care
WB+ result not collected by CBVTC client
Lost to care
Decoding not agree Lost to care
Clinic not regisdtered Lost to care
Linked to care
From CBVTC to HIV clinicH
IV c
lin
ic
A
no
nym
ou
st
test
ing
at
CB
VT
C
Methods - Statistical analyses
Cox-proportional hazard models were used to identify factors associated with:
being linked to care (LTC) – baseline date of HIV test
starting cART – baseline date of first visit to HIV clinic
Both pre-clinical (16) and clinical factors (5) were included into analyses and tested as potential predictors of linkage to care
Results – linked to care
232 persons tested HIV+ in CBVTC
144 (62.1%) were linked to care
(95%CI:60%-70%)
81.2% registered within 3 months from testing
Medial follow-up per person 3 months
(95%CI:1-23)
239 person years of follow-up
Baseline characteristics for those who were tested
Median (IQR) Total n (%) Not linked to care Linked to care P value
Total number 232 ( 100) 88 ( 100) 144 ( 100) -
Age at test <=30* 116 (50.0) 50 (56.8) 66 (45.8) 0.105
Male sex 220 (94.8) 83 (94.3) 137 (95.1) 0.784
Polish nationality 223 (96.1) 83 (94.3) 140 (97.2) 0.276
High education 180 (77.6) 63 (71.6) 117 (81.3) 0.089
MSM 175 (75.4) 55 (62.5) 120 (83.3) <.001
Test at CBVCT 128 (55.2) 44 (50.0) 84 (58.3) 0.216
No test last year 65 (28.0) 25 (28.4) 40 (27.8) 0.917Year of test 2010 32 (13.8) 14 (15.9) 18 (12.5) 0.575 2011 76 (32.8) 31 (35.2) 45 (31.3) 2012 70 (30.2) 22 (25.0) 48 (33.3) 2013 54 (23.3) 21 (23.9) 33 (22.9)
* median age 30.1 (IQR: 25.2-35.9) years
Baseline characteristics for those who were tested
Median (IQR) Total n (%) Not linked to care Linked to care P value
STI 35 (15.1) 11 (12.5) 24 (16.7) 0.391
Partner tested 84 (36.2) 32 (36.4) 52 (36.1) 0.969
Partner HIV+ 53 (22.8) 13 (14.8) 40 (27.8) 0.024
Stable relationship within last year
27 (11.6) 15 (17.0) 12 ( 8.3) 0.049
Condom use with stable partners
100 (43.1) 30 (34.1) 70 (48.6) 0.031
Condom use with casual partners
132 (56.9) 47 (53.4) 85 (59.0) 0.402
No of stable partners
1-5, unknown 67 (28.9) 29 (33.0) 38 (26.4) 0.549 6-20 93 (40.1) 34 (38.6) 59 (41.0) >21 72 (31.0) 25 (28.4) 47 (32.6)
No of casual partners
1-5, unknown 176 (75.9) 69 (78.4) 107 (74.3) 0.495 6-20 40 (17.2) 12 (13.6) 28 (19.4) >21 16 ( 6.9) 7 ( 8.0) 9 ( 6.3)
Kaplan-Meier plots
Unadjusted modelAdjusted model
Linked to care Unadjusted hazard ratios and 95% CI
VariableAge at test (per year older)
Education lower /unknown
Bi/heterosexual orienation
Partner not HIV+ *
Not in stable relationship last year*
No condom use with stable partners*
0.25 0.5 1 2 4
HR (95%CI)* or unknown
Only variables significant in univariate analyses (p <0.1) are shown above
better linkageworse linkage
Linked to careUnadjusted and adjusted hazard ratios and 95% CI
VariableAge at test (per year older)
Education lower /unknown
Bi/heterosexual orienation
Partner not HIV+ *
Not in stable relationship last year*
No condom use with stable partners*
0.25 0.5 1 2 4
HR (95%CI)
Unadjusted modelAdjusted model
Multivariate models adjusted for variables significant in univariate analyses (p <0.1) as shown above
* or unknownbetter linkageworse linkage
Results – starting cART
116 (80.5%) started cART during follow up
CD4 count was 393 (292-506) cells/ul *
HIV RNA was 4.5 (3.9-5.1) log copies/ml*
Median follow-up per person 3 (95% CI:1-16)
months
118 person years of follow-up
Pre-clinical characteristics were similar for patients
who started and not started cART
* Median (IQR) of baseline measurement at first clinic visit
Clinical characteristics for those who were linked and started cART
Median (IQR) Total n(%)Not started
cARTStarted cART P value
anti-HBc at first visit
Yes 16 (11.1) 5 (18.5) 11 ( 9.4) 0.40
No 103 (71.5) 18 (66.7) 85 (72.6)
Unknown/missing 25 (17.4) 4 (14.8) 21 (17.9)
anti- HCV at first visit
Yes 3 ( 2.1) 0 ( 0.0) 3 ( 2.6) 0.78
No 125 (86.8) 25 (92.6) 100 (85.5)
Unknown/missing 16 (11.1) 2 ( 7.4) 14 (12.0)
Syphilis at first visit
Yes 30 (20.8) 7 (25.9) 23 (19.7) 0.47
No/undetermined 100 (69.4) 19 (70.4) 81 (69.2)
Unknown/missing 14 ( 9.7) 1 ( 3.7) 13 (11.1)
HIV RNA at baseline
0 - 10,000 40 (27.8) 16 (59.3) 24 (20.5) <.001
<10,000 104 (72.2) 11 (40.7) 93 (79.5)
CD4 at baseline
1 - 350 49 (34.0) 2 ( 7.4) 47 (40.2) 0.005
351 - 500 58 (40.3) 12 (44.4) 46 (39.3)
> 500 37 (25.7) 13 (48.1) 24 (20.5)
Kaplan-Meier plots
Kaplan-Meier plots
50%
60%
90%
Unadjusted modelAdjusted model
Started cART Unadjusted hazard ratios and 95% CI
VariableAge at test (per year older)
Education lower /unknown
Year of linked to care 2012 vs. 2010/11
2013/14 vs. 2010/11
No STI/unknown
No condom use with casual partners*
Number of casual partners6-20 vs 1-5
>21 vs 1-5
HIV RNA (/10 fold higher)
CD4 (/2 fold higher)
0.125 0.25 0.5 1 2 4 8
HR (95%CI)
* or unknown
Only variables significant in univariate analyses (p <0.1) are shown
starting cARTdelaying cART
Started cARTUnadjusted and adjusted hazard ratios and 95% CI
Unadjusted modelAdjusted model
VariableAge at test (per year older)
Education lower /unknown
Year of linked to care 2012 vs. 2010/11
2013/14 vs. 2010/11
No STI/unknown
No condom use with casual partners*
Number of casual partners6-20 vs 1-5
>21 vs 1-5
HIV RNA (/10 fold higher)
CD4 (/2 fold higher)
0.125 0.25 0.5 1 2 4 8
HR (95%CI)
* or unknown
starting cARTdelaying cART
Limitations
patients classified as lost to care could migrated to or registered at non-regional clinicsrather trend towards centralisation of migration for
care
it is not possible to perform external quality assurance to exclude double testing in Polish settings (anonymous registration)patients re-test mostly due to undetermined WB
result
Conclusions
Benefits of HIV care, measured by access to early treatment, steadily improved in recent 4 years
1 in 3 persons aware of their HIV status remain outside professional healthcare
Bi/heterosexual persons with primary/unknown education are at higher risk of remaining outside care despite being diagnosed
Perspectives
All CBVTC in Poland record Western blot test numbers as a possible cross-check
To develope coding system allowing for outcome measures of care pathway
Next TAK step is introducing targeted intervention in 2015
Acknowledgments
To all people who continuously support us.Thank you