Differences In HAART Uptake And HIV Disease Progression According

differences in haart uptake and hiv disease progression according to geographical origin and ethnicity julia del amo, inma jarrín, john gil

Differences in HAART uptake and HIV disease progression according to
geographical origin and ethnicity
Julia del Amo, Inma Jarrín, John Gill, Ronald Geskus, Laurence Meyer,
Giota Touloumi, Maria Prins, Kholoud Porter, Santiago Pérez-Hoyos
Summary
Inequity and social exclusion, along with cultural and language
barriers to HIV/AIDS care can impair the uptake and response to HAART
by migrant and ethnic minority populations in Western countries. The
few studies that have addressed this question have analysed data from
seroprevalent cohorts. HIV infected subjects from the CASCADE
Collaboration have well estimated dates of HIV seroconversion, which
by definition exclude people with a late HIV diagnosis. Using the new
variables provided by the CASCADE Collaboration, and reclassifying
this information in two newly created categories “Geographical origin
and ethnicity” we aim to examine differences in time to HAART from HIV
seroconversion, as well as to time to AIDS and death, between people
of different geographical origins and ethnicity. We will estimate the
effect of geographical origin and ethnicity on time from HIV
seroconversion to HAART, AIDS and death from any cause using
Kaplan-Meier methods and Cox proportional hazard models. Patterns of
hazard will be obtained by parametrical models.
Study aims and objectives
The overall objectives are:
1.
To examine differences in time to HAART from HIV seroconversion
between people of different geographical origins and ethnicity.
2.
To examine differences in time from HIV seroconversion to AIDS
(overall) and to each specific AIDS-defining disease as initial
AIDS-defining event, before and after the introduction of HAART,
between people of different geographical origins and ethnicity.
3.
To examine differences in time from HIV seroconversion to death
before and after the introduction of HAART between people of
different geographical origins and ethnicity.
4.
To explore differences in the above outcomes between cohorts
within CASCADE.
5.
To compare CD4 counts in African, Black West Indies and European
subjects according to sex and subtype at the time of
seroconversion and under HAART, taking into account sex and viral
subtype
6.
To compare rates of clinical progression to clinical Aids and
death in African, Black West Indies and European subjects, overall
and at different strata of CD4 cells
7.
To characterize different patterns of hazard of AIDS and death
between people of different geographical origins and ethnicity
Data requirements and classifications of main exposure variables
After the Rome workshop in 2006, we decided to ask the cohorts the
data reflecting geographical origin, nationality, ethnicity, and/or
race were they collecting. By 2007 most cohorts responded and we
presented the collected data items in the Berlin workshop in 2007. We
then decided to explore how best to classify the information provided.
Data collected by the cohorts
The following variables are collected by the cohorts. This information
was presented in Berlin.

The classifications used for race and ethnicity are described below:
Classifications used for race and ethnicity in CASCADE Cohorts
*
Southern Alberta
*
White, Aboriginal, Inuit, Metis, Black, Oriental, Hispanic,
Indo-Asian, Other, unknown
*
Serocco
*
White, Black, Asiatic, Others, Missing
*
Netherlands IDU&MSM
*
Western-European (incl Southern-Europe), Suriname / Antillean,
Moroccan, Turkish, Moluccan, Suriname creole, Suriname other
(Hindo, Chinese, Javanese), Antillean, African, Asia, Eastern
Europe, Other,Unknown
*
UK Reg
*
White, black African, Black Caribbean, Indian subcontinent,
Other, Black unspecified, Unknown or not recorded
*
Madrid
*
White, Black, Gypsy, Other, Not known
*
Swiss
*
White, black, hispano-american, asian, other
*
Danish
*
White, black, asia, inuit and other
*
Norway
*
Caucasian,oriental,black, other
Data received
We have received some data from 13 out the 19 cohorts. Of the 6 which
have not provided data, only one has no data available (Royal Free).
1.
Australia Sydney AIDS Prospective Study Sydney Primary HIV
Infection cohort - no data provided
2.
Canada - South Alberta clinic
3.
Denmark - Copenhagen HIV Seroconverter Cohort - no data provided
but available in the original CASCADE dataset
4.
France - Aquitaine cohort
5.
French Hospital Database
6.
Lyon Primary Infection cohort - no data provided but available in
the original CASCADE dataset
7.
SEROCO cohort
8.
Germany - German cohort - no data provided
9.
Greece - Greek Haemophilia cohort – all white greeks
10.
Italy - Italian Seroconversion Study
11.
Netherlands - Amsterdam Cohort Studies among homosexual men and
drug users
12.
Norway - Oslo and Ulleval Hospital cohorts
13.
Madrid – Sandoval
14.
Barcelona
15.
Badalona
16.
Valencia
17.
Switzerland - Swiss HIV cohort – no data provided but available in
the original CASCADE dataset
18.
Royal Free haemophilia cohort – no data available
19.
UK Register of HIV Seroconverters
Based on what we have, and acknowledging all possible solutions will
inevitably carry some degree of misclassification, we suggest creating
two variables called Geographical origin and Ethnicity.
Geographical origin
Given the data available, we suggest to create a new variable called
“geographical origin” which would include the information collected in
the variables “country of birth, country of origin and nationality”.
However, looking at the way the variable “ethnicity” has been filled
in the UK, we could also make some assumptions regarding geographical
origin except for “black-unspecified/other” (n=13), “mixed
race”(n=27). For “whites” in UK, we can assume they are National. This
will fit better classification 2.
For the FHDB, the variables “stay foreign country > 6 months since 78”
is difficult to classify, but we could assimilate it to country of
origin.
Given the numbers, we could make the following classifications for
“geographical origin”:
Classification 1
1.
National
2.
Other European country
3.
North Africa and Middle East
4.
Sub-Saharan Africa
5.
Caribbean
6.
South and Central America
7.
North America (USA & Canada)
8.
Indian Subcontinent
9.
South East Asia
10.
Other
Classification 2
We could further classify this in:
1.
National, Other European country North America (USA & Canada)
2.
North Africa and Middle East
3.
Sub-Saharan Africa
4.
Caribbean, South and Central America
5.
Indian Subcontinent, South East Asia
6.
Other
Ethnicity
We could also create another variable reflecting “ethnicity” (white,
black, other) looking at the data from UK, Madrid-Sandoval, Seroco,
Netherlands, Norway and Canada which record ethnicity or race. At a
second stage, we could make some assumptions in the other cohorts
based on the commonest ethnicity group for a given country of
origin/country of birth or nationality.
Censoring strategies
Special attention will be paid to losses to follow-up by geographical
origin to assess the “unhealthy remigration bias” (salmon bias) of
people with a different geographical origin who may return to their
countries to die. We will examine the proportions of losses to
follow-up and measure the last CD4 count of the people lost to
follow-up according to their geographical origin.
Analyses
We will perform all analyses stratified by geographical origin or
ethnicity so we will not assume that the distribution of measured and
unmeasured confounding is uniform for each stratum of the main
exposure variable. We will assess all possible interactions to support
this approach.
Before carrying out the analyses with the newly created variables, we
will asses the effect on the various outcomes of interest for each of
the variables collected in the cohorts (nationality, race, country of
birth, country of origin, etc…).
We will estimate the effect of geographical origin and ethnicity on
time from HIV seroconversion to HAART, AIDS and death from any cause
using Kaplan-Meier methods and Cox proportional hazard models. Also
parametric survival models (Weibull, lognormal, generalised gamma)
will be fitted to characterize different patterns of hazard for each
level of the considered varibles
Differences in time from seroconversion to each of the outcomes will
be analysed using both a cause-specific proportional hazard model and
a competing risk proportional hazard model. We will allow late entry
of individuals at the time of enrolment into the original cohort.
Calendar period at risk will be divided in different time-bands
reflecting availability of antiretroviral therapy, and modelled as a
time-dependent covariate.
Parametric survival models will be considered also to show different
changes in pattern of hazard by calendar period. Interaction between
geographical origin or ethnicity and calendar period will be checked
Differences in outcome by cohort/country will be tested using
homogeneity tests.
References
1.
Anastos K, Gange SJ, Lau B, Weiser B, Detels R, Giorgi JV.
Association of race and gender with HIV-1 RNA levels and
immunologic progression. Acquir Immune Defic Syndr 2000;24:218-26
2.
Bhopal R Glossary of terms relating to ethnicity and race: for
reflection and debate. J Epidemiol Community Health 2004;58:441-5.
3.
Bhopal R Revisiting race/ethnicity as a variable in health
research. Am J Public Health 2002 92:156-7.
4.
Blaxhult A, Mocroft A, Phillips A, van Lunzen J, Bentwich Z,
Stergiou G, et al. Does European or non-European origin influence
health care and prognosis for HIV patients in Europe? The EuroSIDA
Study Group. HIV Med 1999;1(1):2-9
5.
Bröring, Canter, Schinaia, Teixeira (eds.): Access to Care:
Privilege or Right? Migration and HIV Vulnerability in Europe,
NIGZ (Woerden, NL), November 2003
6.
Brown AE, Malone JD, Zhou SY, Lane JR, Hawkes CA. Human
immunodeficiency virus RNA levels in US adults: a comparison based
upon race and ethnicity. J Infect Dis. 1997;176(3):794-7
7.
Burns FM, Fakoya AO, Copas AJ, French PD. Africans in London
continue to present with advanced HIV disease in the era of highly
active antiretroviral therapy. AIDS 2001;15:2453-5
8.
Cooper RS, Kaufman JS, Ward R.Race and genomics.N Engl J Med
2003;348:1166-70.
9.
Del Amo J, Broring G, Hamers F, Infuso A, Fenton K. Monitoring
HIV/AIDS in migrants communities in Europe.AIDS
2004;18(14):1867-73
10.
Del Amo J, Petruckevitch A, Phillips AN, Johnson AM, Stephenson J,
Desmond N, et al. Disease progression and survival in
HIV-1-infected Africans in London. AIDS 1998;12:1203-9
11.
Ed: Louhenapessy M. 7th European Migrants Meeting: Access to HIV
Care and Support for Migrants and Ethnic Minorities in Europe.
http://www.aidsmobility.org/am5a1.htm
12.
Fiscella K, Franks P, Gold MR & Clancy CM. Inequalities in racial
access to health care. JAMA 2000; 284:2579–2584
13.
Goodman A. Why genes don't count (for racial differences in
health).Am J Public Health 2000 ;90:1699-702.
14.
Kaufman J and Cooper R. Commentary: considerations for use of
racial/ethnic classification in etiologic research. Am J Epidemiol
2001;154:291-8.
15.
Kaufman JS, Cooper RS. Seeking causal explanations in social
epidemiology. Am J Epidemiol. 1999;150:113-20
16.
Staehelin C, Rickenbach M, Low N, Egger M, Ledergerber B, Hirschel
B, et al. Migrants from Sub-Saharan Africa in the Swiss HIV Cohort
Study: access to antiretroviral therapy, disease progression and
survival. AIDS 2003;17:2237-44
6