Assessing the role of individual and neighbourhood characteristics in HIV testing: evidence from a population based survey.

Abstract

Individuals living in deprived neighbourhoods have poor health outcomes, including human immunodeficiency virus (HIV) infection mortality. We assessed the association between individual and neighbourhood characteristics, and HIV testing across Canada.
We used logistic regression modelling to evaluate this association in 2219 men and 2815 women, aged 18-54 years, in Canada, using data from the National Population Health Survey (1996/7),. Socio-economic characteristics and presence of a sexually transmitted infection (STI) were the individual level characteristics. Small area of residence was classified according to categories of material and social deprivation; these were the 'neighbourhood' variables in the model.
Ethnic minority women were less likely to report an HIV test than white women (OR 0.44, 95% CI: 0.23 to 0.86). Women without a regular doctor were significantly less likely to report ever having had an HIV test (OR 0.57, 95% CI: 0.35 to 0.93). Adjusting for individual level characteristics, we found that men and women living in the most materially deprived neighbourhoods were slightly less likely to report HIV testing than those living in the least deprived neighbourhoods (Men - OR 0.61, 95% CI: 0.34 to 1.08; Women - OR 0.62, 95% CI: 0.38 to 1.00).
Thus, living in poor neighbourhoods was associated with poor uptake of an HIV test. These economic disparities should be taken in account while designing future prevention strategies. Ethnic minority women were less likely to go for HIV testing and culturally appropriate messages may be required for prevention in ethnic minorities.

Full text

Page 1

46 The Open AIDS Journal, 2009, 3, 46-54


1874-6136/09 2009 Bentham Open
Open Access
Assessing the Role of Individual and Neighbourhood Characteristics in
HIV Testing: Evidence from a Population Based Survey
Maninder Singh Setia*,1, Amelie Quesnel-Vallee1,2, Sarah Curtis3 and John Lynch4
1Department of Epidemiology, Biostatistics, and Occupational Health, McGill University, Montreal, Quebec, Canada
2Department of Sociology, McGill University, Montreal, Quebec, Canada
3Department of Geography, University of Durham, Durham, UK
4Division of Health Sciences, University of South Australia, Adelaide, Australia
Abstract: Objectives: Individuals living in deprived neighbourhoods have poor health outcomes, including human
immunodeficiency virus (HIV) infection mortality. We assessed the association between individual and neighbourhood
characteristics, and HIV testing across Canada.
Methods: We used logistic regression modelling to evaluate this association in 2219 men and 2815 women, aged 18-54
years, in Canada, using data from the National Population Health Survey (1996/7),. Socio-economic characteristics and
presence of a sexually transmitted infection (STI) were the individual level characteristics. Small area of residence was
classified according to categories of material and social deprivation; these were the ’neighbourhood’ variables in the
model.
Results: Ethnic minority women were less likely to report an HIV test than white women (OR 0.44, 95% CI: 0.23 to 0.86).
Women without a regular doctor were significantly less likely to report ever having had an HIV test (OR 0.57, 95% CI:
0.35 to 0.93). Adjusting for individual level characteristics, we found that men and women living in the most materially
deprived neighbourhoods were slightly less likely to report HIV testing than those living in the least deprived
neighbourhoods (Men - OR 0.61, 95% CI: 0.34 to 1.08; Women - OR 0.62, 95% CI: 0.38 to 1.00).
Discussion: Thus, living in poor neighbourhoods was associated with poor uptake of an HIV test. These economic
disparities should be taken in account while designing future prevention strategies. Ethnic minority women were less
likely to go for HIV testing and culturally appropriate messages may be required for prevention in ethnic minorities.
Keywords: HIV testing, ethnic minorities, neighbourhoods.
INTRODUCTION

of 68,604 human immunodeficiency virus (HIV) infections
until 2007 [1]. Men who have sex with men accounted for
62% of total new HIV infections in the period from 1985-
1991, however, this proportion had reduced to 41% in 2007.
There was a simultaneous increase in the number of
infections due to heterosexual transmission (7% in 1985-
1991 and 12% in 2007) [1]. The introduction of highly active
antiretroviral therapy has resulted in reduction of
complications and mortality in HIV infected individuals.
However, HIV infected individuals will access the therapy
and related prevention services only if they are aware of their
HIV status. Thus HIV testing forms a core component of
HIV prevention and care [2].
The Public Health Agency of Canada has reported a total

testing facilities include nominal or name-based testing,

Canada has a universal health care system; the HIV

*Address correspondence to this author at the Department of Epidemiology
and Biostatistics, 1020 Pine Avenue West, Montreal, QC, H3A 1A2,
Canada; Tel: 001-514-398-1269; Fax: 001-514-398-8851;
E-mail: maninder.setia@mail.mcgill.ca
non-nominal testing, and anonymous testing in various
provinces [3]. However, it is estimated that about 27% of
HIV infected Canadians are not aware of their status [4].
HIV testing may depend on various factors – perception of
risk, clinical indicators, and access to health services. The
testing pattern may also vary according to cultural practices;
it may be different among those from ethnic minorities
compared with the White Canadian population, and women
may be particularly disadvantaged [5-7]. A recent review
identified HIV testing patterns among women in Canada to
be an important research area [8]. This may be useful to
design prevention and care programmes for these
communities in Canada. Another factor often discussed in
health care access is the role of the neighbourhood in which
individuals live. Studies have demonstrated that AIDS
incidence and mortality is higher in economically deprived
areas in non-industrialised as well as industrialised countries
[9-14]. Although geographical mapping has shown that HIV
services are less accessible in economically disadvantaged
neighbourhoods in Toronto [15] an analysis of predictors of
uptake of an HIV test in different neighbourhoods across
Canada has never been done.

Page 2

Assessing the Role of Individual and Neighbourhood Characteristics in HIV Testing The Open AIDS Journal, 2009, Volume 3 47

association between socioeconomic factors and HIV testing
in men and women in Canada, both at the individual and
neighbourhood level.
Thus, the present study was designed to evaluate the
METHODS

National Population Health Survey (NPHS; 1996/97) [16].
Though NPHS is a longitudinal survey with seven data
waves with 12 years of follow-up, the variable on HIV
testing was available only in the 1996/7 wave. It is nationally
representative dataset and provides information on the social,
demographic, economic, occupational, environmental, and
health characteristics of the Canadian population. We linked
these data from the NPHS to information relating to
Dissemination areas (DA) of residence in 1996 - the smallest
unit of disseminating census data – using composite small
area classifications for corresponding areas derived from the
Canadian Census data (2001). The data included 2219 men
and 2815 women, aged 18-54 years, residing in 1945 (mean
1.1) and 2460 (mean 1.1) DAs respectively.
This study is a cross-sectional analysis of data from the

have had an HIV test other than for insurance or blood
donation. The individual variables use as explanatory of this
outcome were for 1996: socio-demographic variables - age,
the living condition of individuals, educational level, income
category, the province of residence, and ethnicity; clinical
indicator - presence of a sexually transmitted infection (STI)
in the past two years; and access indicator - whether they
have a regular doctor or not. Ethnic minority status was
determined by self-reporting of ethnicity. Two geographic
variables describing the individual’s 1996 area of residence
were also used as explanatory variables in the model; the
material deprivation index and the social deprivation index,
developed by Pampalon and colleagues, using 2001 census
data for each of the DAs in Canada [17]. The material
deprivation indicator takes into account the education level,
employment status, and income; thus, reflects the economic
poverty in the concerned populations. The social deprivation
measure, takes into account the living condition of
individuals (alone, separated, single parent families) and
represents the level of social isolation or social cohesion in
the population in the respective areas. As discussed earlier,
economically deprived neighbourhoods reported higher
AIDS cases and related mortality [9-14]. Based on the Postal
Code correspondence file [18] we identified the DA in which
the respondent lived in 1996. The DA identifier was used to
link the individual data with geographical information on
local material and social deprivation indicators.
The outcome variable of interest was whether individuals

Station, Texas, USA) to conduct the logistic regression
analysis for the present study. The ‘boot strap’ method with
500 replications was used for calculating standard errors to
allow for representativeness of this complex survey design.
We initially analysed the proportion of HIV testing in the
various individual and geographic level predictor categories;
these were population weighted proportions. The modelling
was conducted in the following sequence: 1) we initially

We used STATA (version 10) (StataCorp, College
performed analysis between the outcome and each of the
explanatory variables (individual and geographic); 2) a
multivariate analysis of the outcome and each of the
individual level explanatory variables; 3) The next group of
models were to test if the geographic level variables
(separately for material and social deprivation indices)were
associated with the outcome after controlling for individual
level attributes. These data were not analysed in a multi-level
model because the average number of individuals in each
DA was 1.1, but options in STATA were used to adjust
errors for any effects of clustering of some individuals in the
same areas. We performed the linear contrast tests for trend
to asses for any trends in the material and social deprivation
quintiles [19]. All the models were initially built for the
whole cohort, followed by models for men and women
separately. The latter was done to assess if any of the
explanatory variables had different effects in these two
genders.

the Institutional Review Board of McGill University.
The study was approved for secondary data analysis by
RESULTS
Descriptive data

were 37.1 (+ 9.6) and 36.3 (+ 9.6) years respectively.
Overall, more women had tested for HIV than men (20%
versus 15%, p<0.01) in our sample. Of the 146 ethnic
minority men, 25% were Chinese, 23% were black, and 17%
were South Asian. However, among the 139 ethnic minority
women, 22% were South Asian, 21% were Chinese, and
16% were Black. The proportion of HIV testing was highest
in men (21%) and women (27%) aged 25 to 34 years.
However, 36% of men and 44% of women reporting a STI in
the past two years had tested for HIV (Table 1). The most
common reason reported for HIV testing was ‘peace of
mind’ in both men (43%) and women (32%). About 19% of
women reported ‘pregnancy’ as a reason for getting tested.
However, only 5% of men and women reported ‘risky sexual
behaviours’ as the reason for the HIV test.
The mean ages (standard deviation) of men and women
Complete Cohort Models

various individual and neighbourhood level characteristics
are presented in Table 2. We are only referring to the
adjusted estimates in the subsequent discussion.
Crude associations and adjusted associations between

characteristics we found that women were more likely to
have had an HIV test than men. People who had an STI in
the past two years were more likely to report an HIV test
than those who had not. After adjusting for all individual
level variables, we found that people living in the most
materially deprived neighbourhoods reported a lower HIV
test uptake than those living in the least deprived
neighbourhoods. However, the reverse was true for social
deprivation. Thus, people living in the most socially
deprived neighbourhoods were more likely to report an HIV
test than those living in the least socially deprived

In the complete model with individual level

Page 3

48 The Open AIDS Journal, 2009, Volume 3 Setia et al.

Table 1. Proportion of People Reporting HIV Testing (Total = 5034; 2219 Males and 2815 Females) from the National Population
Health Survey, Canada*

Separated by Gender
Total
N=5034
[Population = 11540425]
Males
N=2219
[population=5756447]
Females
N=2815
[population = 5783978]
Characteristics
n Proportion HIV Tested
n Proportion HIV Tested (%) n Proportion HIV Tested
All 5034 18 2219 15 2815 20
Individual level Variables
Age groups (years)
18-24 658 19 264 13 394 24
25-34 1452 24 618 21 834 27
35-44 1666 17 736 13 930 20
45-54 1258 10 601 11 657 9
p < 0.00 p < 0.00 p < 0.00
Living conditions
Living with partner/children 3815 16 1733 14 2082 19
Single parent with
dependent children
773 20 387 21 386 21
Living alone/unattached 446 25 99 0.18 347 27
p < 0.00 p=0.03 p=0.02
Education
Less than secondary 619 12 304 12 315 13
Secondary education 703 12 315 9 388 15
Some post-secondary 1406 22 624 19 782 24
College/University 2306 18 976 15 1330 21
p < 0.00 p=0.01 p < 0.00
Income group
Lowest 704 23 240 19 464 26
Low-Mid 1399 17 591 12 808 20
Upper-Mid 2141 18 1011 17 1130 19
Upper 790 15 377 12 413 19
p = 0.02 p=0.05 p=0.18
Province
Others 2486 13 1071 12 1415 15
Ontario 1163 21 519 17 644 25
Quebec 904 16 425 16 479 17
British Columbia 481 18 204 15 277 20
p < 0.00 p=0.15 p < 0.00
Ethnicity
White 4749 18 2073 15 2676 21
Ethnic minorities 285 14 146 15 139 13
p=0.18 p=0.94 p=0.05
Has a regular doctor
Yes 4288 18 1746 15 2542 21
No 746 15 473 16 273 13
p = 0.13 p=0.68 p=0.01

Page 4

Assessing the Role of Individual and Neighbourhood Characteristics in HIV Testing The Open AIDS Journal, 2009, Volume 3 49

neighbourhoods. The tests for trend were significant for the
material and social deprivation quintiles. The correlation
between the material and social deprivation indices
represented in our data was -0.07 in our data.
Models in Men and Women

men and women separately in Table 3. We refer to the
adjusted association (Table 3, Models II and III) in the
subsequent discussion.
We have described the crude and adjusted associations in

report an HIV test than white women, although this
difference was not seen in ethnic minority men. Women
without a regular doctor were less likely to report ever
having had an HIV test, a feature again not seen in men.
Women living in Ontario were more likely to have ever had
an HIV test than other Canadian provinces. After adjusting
for individual level characteristics, we found that men and
women living in most materially deprived neighbourhoods
reported lower HIV testing than those living in least
deprived neighbourhoods, although the OR in the males was
not significant. However, men and women living in most
socially deprived neighbourhoods were significantly more
We found that ethnic minority women were less likely to
likely to report an HIV test than those living in the least
socially deprived neighbourhoods.
DISCUSSION

men in Canada in 1996/7. However, ethnic minority women
reported lower HIV testing than white women. Further,
women who did not have a regular doctor were less likely to
report ever having an HIV test. About 36% of men and 44%
of women who reported having an STI in the past two years
had been tested for HIV. Adjusting for all the individual
level predictor variables, people living in the most materially
deprived neighbourhoods in Canada were about 40% less
likely to have ever had an HIV test compared with those
living in the least deprived neighbourhoods.
Women were more likely to have had an HIV test than

factors: perception of risk, access to testing services, and
perception of social stigma associated with the infection as
well as practical implications of testing positive, emotional
trauma, and fear of social rejection [5]. HIV testing is higher
in individuals who perceive themselves to be at risk [20].
Ethnic minorities often present late to HIV clinics as they are
less likely to undergo HIV testing, a feature common to most

HIV testing by individuals may depend on various
(Table 1) contd…..
Separated by Gender
Total
N=5034
[Population = 11540425]
Males
N=2219
[population=5756447]
Females
N=2815
[population = 5783978]
Characteristics
n Proportion HIV Tested
n Proportion HIV Tested (%) n Proportion HIV Tested
Had an STI in the past
Yes 114 41 37 36 77 44
No 4920 17 2182 15 2738 19
p < 0.00 p < 0.00 p < 0.00
Geographic Variables
Material deprivation
quintile

1st (least deprived) 799 20 317 20 482 21
2nd 917 19 411 15 506 23
3rd 1041 17 492 14 549 20
4th 1116 18 513 14 603 23
5th (most deprived) 1161 13 486 13 675 13
p = 0.05 p=0.29 p=0.03
Social deprivation quintile
1st (least deprived) 1019 13 450 10 569 16
2nd 1104 16 491 13 613 18
3rd 1018 17 463 16 555 18
4th 955 19 443 14 512 24
5th (most deprived) 938 24 372 23 566 25
p < 0.00 p < 0.00 p = 0.03
* = The proportions are weighted for the population, hence only proportions are provided.

Page 5

50 The Open AIDS Journal, 2009, Volume 3 Setia et al.

Table 2. Models Showing Crude and Adjusted Association between Individual and Geographic Characteristics, and Outcome
(HIV Testing) in 5034 Individuals from the National Population Health Survey, Canada

Model III: Adjusted Association for Individual and
Geographic Characteristics
Characteristics
Model I:
Crude Association for
Individual and Geographic
Characteristics
Model II:
Adjusted Association for
Individual
Characteristics
Material Deprivation Social Deprivation
Individual level Variables
Gender
Male
Females
Age groups (years)
18-24
25-34
35-44
45-54
Living with partner/children
Single parent with dependent
children
Living alone/unattached
Education
Less than secondary
Secondary education
Some post-secondary
College/University
Income group
Lowest
Low-Mid
Upper-Mid
Upper
Province
Others
Ontario
Quebec
British Columbia
Ethnicity
White
Ethnic minorities
Has a regular doctor
Yes
No
Had an STI in the past
No
Yes
Geographic Variables
Material deprivation quintile
1st (least deprived)
2nd
3rd
4th
5th (most deprived)
Social deprivation quintile
1st (least deprived)
2nd
3rd
4th
5th (most deprived)
* Test for trend p=0.01, ** Test for trend p=0.004.


1.00 (reference)
1.42 (1.18 - 1.71)

1.00 (reference)
1.36 (1.01 – 1.85)
0.86 (0.61 – 1.21)
0.47 (0.33 – 0.67)
1.00 (reference)


1.00 (reference)
1.33 (1.10 - 1.61)

1.00 (reference)
1.49 (1.10 - 2.03)
0.97 (0.69 - 1.36)
0.53 (0.37 - 0.77)
1.00 (reference)


1.00 (reference)
1.32 (1.09 - 1.61)

1.00 (reference)
1.49 (1.10 - 2.02)
0.96 (0.68 - 1.34)
0.52 (0.36 - 0.75)
1.00 (reference)


1.00 (reference)
1.32 (1.08 - 1.60)

1.00 (reference)
1.47 (1.08 - 2.00)
0.98 (0.70 - 1.38)
0.54 (0.37 - 0.77)
1.00 (reference)
1.69 (1.31 – 2.18) 1.57 (1.19 - 2.07) 1.56 (1.18 - 2.05) 1.42 (1.06 - 1.90)
2.01 (1.44 – 2.81)

1.00 (reference)
1.00 (0.64 - 1.56)
1.96 (1.35 - 2.84)
1.58 (1.09 - 2.30)

1.00 (reference)
0.65 (0.47 - 0.92)
0.71 (0.52 - 0.95)
0.60 (0.41 - 0.87)

1.00 (reference)
1.74 (1.38 - 2.19)
1.27 (0.99 - 1.64)
1.40 (1.04 - 1.89)

1.00 (reference)
0.76 (0.49 - 1.18)

1.00 (reference)
0.80 (0.59 - 1.08)

1.00 (reference)
3.38 (2.12 - 5.37)


1.00 (reference)
0.91 (0.66 - 1.25)
0.78 (0.58 - 1.06)
0.89 (0.65 - 1.22)
0.61 (0.45 - 0.83)

1.00 (reference)
1.26 (0.92 - 1.72)
1.34 (0.94 - 1.91)
1.54 (1.10 - 2.14)
2.09 (1.51 - 2.89)
1.71 (1.21 - 2.42)

1.00 (reference)
0.91 (0.57 - 1.43)
1.73 (1.17 - 2.56)
1.38 (0.93 - 2.06)

1.00 (reference)
0.70 (0.50 - 0.99)
0.79 (0.58 - 1.07)
0.68 (0.46 - 1.02)

1.00 (reference)
1.81 (1.42 - 2.32)
1.38 (1.05 - 1.80)
1.41 (1.04 - 1.93)

1.00 (reference)
0.65 (0.41 - 0.1.01)

1.00 (reference)
0.81 (0.58 - 1.12)

1.00 (reference)
2.73 (1.71 - 4.36)













1.73 (1.22 - 2.43)

1.00 (reference)
0.87 (0.55 - 1.39)
1.64 (1.11 - 2.44)
1.29 (0.86 - 1.92)

1.00 (reference)
0.69 (0.49 - 0.97)
0.75 (0.54 - 1.03)
0.62 (0.41 - 0.95)

1.00 (reference)
1.75 (1.36 - 2.25)
1.36 (1.04 - 1.79)
1.39 (1.02 - 1.89)

1.00 (reference)
0.66 (0.42 - 1.03)

1.00 (reference)
0.79 (0.57 - 1.09)

1.00 (reference)
2.73 (1.69 - 4.40)


1.00 (reference)
0.91 (0.65 - 1.28)
0.73 (0.52 - 1.01)
0.85 (0.61 - 1.19)
0.61 (0.43 - 0.88)*






1.62 (1.15 - 2.30)

1.00 (reference)
0.91 (0.57 - 1.44)
1.71 (1.15 - 2.53)
1.38 (0.92 - 2.05)

1.00 (reference)
0.71 (0.50 - 0.99)
0.81 (0.59 - 1.09)
0.72 (0.48 - 1.07)

1.00 (reference)
1.81 (1.41 - 2.31)
1.31 (0.99 - 1.73)
1.38 (1.00 - 1.89)

1.00 (reference)
0.65 (0.41 - 1.02)

1.00 (reference)
0.78 (0.56 - 1.09)

1.00 (reference)
2.65 (1.67 - 4.22)








1.00 (reference)
1.17 (0.85 - 1.62)
1.31 (0.91 - 1.88)
1.42 (1.01 - 2.01)
1.66 (1.17 - 2.34)**

Page 6

Assessing the Role of Individual and Neighbourhood Characteristics in HIV Testing The Open AIDS Journal, 2009, Volume 3 51

Table 3. Models Showing Crude and Adjusted Association between Individual and Geographic Characteristics, and Outcome
(HIV Testing) in 2219 Males and 2815 Females Separately from the National Population Health Survey, Canada

Model I:
Crude Association between Individual
and Geographic Characteristics
Model II:
Adjusted Association for Individual
Level Characteristics
Characteristics
Males
OR (95% CI)
Females
OR (95% CI)
Males
OR (95% CI)
Females
OR (95% CI)
Individual level Variables
Age groups (years)
18-24 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
25-34 1.69 (1.01 - 2.83) 1.32 (0.88 - 1.99) 1.90 (1.11 – 3.25) 1.32 (0.88 - 1.99)
35-44 1.00 (0.58 - 1.72) 0.88 (0.57 - 1.35) 1.22 (0.69 – 2.16) 0.88 (0.57 - 1.35)
45-54 0.74 (0.43 - 1.28) 0.35 (0.22 - 0.56) 0.90 (0.51 - 1.61) 0.35 (0.22 - 0.56)
Living conditions
Living with partner/children 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Single parent with dependent children 1.89 (1.32 - 2.71) 1.59 (1.11 - 2.28) 1.58 (1.06 - 2.36) 1.59 (1.11 - 2.28)
Living alone/unattached 1.70 (0.90 - 3.22) 1.70 (1.13 - 2.55) 1.73 (0.87 - 3.43) 1.70 (1.13 - 2.55)
Education
Less than secondary 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Secondary education 0.75 (0.37 - 1.52) 0.95 (0.49 - 1.86) 0.78 (0.38 - 1.60) 0.95 (0.49 - 1.86)
Some post-secondary 1.72 (1.00 - 2.94) 1.68 (0.92 - 3.08) 1.66 (0.94 - 2.96) 1.68 (0.92 - 3.08)
College/University 1.30 (0.74 - 2.28) 1.39 (0.77 - 2.53) 1.25 (0.68 - 2.27) 1.39 (0.77 - 2.53)
Income group
Lowest 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Low-Mid 0.58 (0.33 - 1.00) 0.78 (0.50 - 1.21) 0.59 (0.33 - 1.00) 0.78 (0.50 - 1.21)
Upper-Mid 0.82 (0.50 - 1.32) 0.73 (0.49 - 1.09) 0.85 (0.50 - 1.44) 0.73 (0.49 - 1.09)
Upper 0.54 (0.31 - 0.95) 0.76 (0.44 - 1.30) 0.59 (0.32 - 1.11) 0.76 (0.44 - 1.30)
Province
Others 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Ontario 1.53 (1.08 - 2.17) 2.15 (1.55 - 2.99) 1.45 (1.00 - 2.11) 2.15 (1.55 - 2.99)
Quebec 1.47 (1.03 - 2.12) 1.36 (0.94 - 1.97) 1.44 (0.97 - 2.13) 1.36 (0.94 - 1.97)
British Columbia 1.37 (0.84 - 2.23) 1.56 (1.06 - 2.29) 1.27 (0.76 - 2.11) 1.56 (1.06 - 2.29)
Ethnicity
White 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Ethnic minorities 1.02 (0.59 - 1.76) 0.44 (0.23 - 0.86) 0.97 (0.56 - 1.70) 0.44 (0.23 - 0.86)
Has a regular doctor
Yes 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
No 1.09 (0.72 - 1.64) 0.57 (0.35 - 0.93) 0.99 (0.65 - 1.53) 0.57 (0.35 - 0.93)
Had an STI in the past
No 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Yes 3.21 (1.49 - 6.93) 3.16 (1.76 - 5.69) 2.56 (1.07 - 6.14) 3.16 (1.76 - 5.69)
Geographic Variables
Material deprivation quintile
1st (least deprived)
2nd
3rd
4th
5th (most deprived)
Social deprivation quintile
1st (least deprived)
2nd
3rd
4th
5th (most deprived)

1.00 (reference)
0.73 (0.43 – 2.24)
0.65 (0.38 – 1.10)
0.64 (0.38 – 1.08)
0.63 (0.38 – 1.03)

1.00 (reference)
1.43 (0.86 – 2.40)
1.65 (0.93 – 2.94)
1.48 (0.86 – 2.51)
2.63 (1.61 – 4.30)

1.00 (reference)
1.12 (0.76 – 1.66)
0.93 (0.65 – 1.33)
1.16 (0.79 – 1.70)
0.60 (0.41 – 0.89)

1.00 (reference)
1.13 (0.73 – 1.75)
1.15 (0.73 – 1.80)
1.62 (1.05 – 2.50)
1.73 (1.12 – 2.66)

Page 7

52 The Open AIDS Journal, 2009, Volume 3 Setia et al.

(Table 3) contd…..
Model I:
Crude Association between Individual
and Geographic Characteristics
Model II:
Adjusted Association for Individual
Level Characteristics
Characteristics
Males
OR (95% CI)
Females
OR (95% CI)
Males
OR (95% CI)
Females
OR (95% CI)
Individual level Variables
Age groups (years)
18-24 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
25-34 1.89 (1.10 - 3.25) 1.33 (0.88 - 1.99) 1.90 (1.10 - 3.26) 1.32 (0.88 - 1.99)
35-44 1.20 (0.68 - 2.15) 0.86 (0.56 - 1.33) 1.24 (0.70 - 2.21) 0.89 (0.58 - 1.37)
45-54 0.87 (0.49 - 1.58) 0.34 (0.21 - 0.55) 0.93 (0.52 - 1.67) 0.35 (0.22 - 0.56)
Living conditions
Living with partner/children 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Single parent with dependent children 1.56 (1.04 - 2.33) 1.61 (1.12 - 2.30) 1.39 (0.90 - 2.13) 1.46 (0.99 - 2.16)
Living alone/unattached 1.77 (0.88 - 3.55) 1.67 (1.12 - 2.49) 1.67 (0.85 - 3.29) 1.58 (1.04 - 2.39)
Education
Less than secondary 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Secondary education 0.76 (0.37 - 1.57) 0.91 (0.46 - 1.79) 0.76 (0.36 - 1.57) 0.96 (0.49 - 1.86)
Some post-secondary 1.59 (0.90 - 2.81) 1.56 (0.85 - 2.86) 1.70 (0.95 - 3.03) 1.63 (0.89 - 2.98)
College/University 1.14 (0.63 - 2.06) 1.29 (0.70 - 2.37) 1.24 (0.68 - 2.26) 1.38 (0.76 - 2.51)
Income group
Lowest 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Low-Mid 0.60 (0.32 - 1.11) 0.74 (0.48 - 1.15) 0.59 (0.33 - 1.15) 0.78 (0.51 - 1.20)
Upper-Mid 0.84 (0.49 - 1.46) 0.68 (0.45 - 1.01) 0.87 (0.51 - 1.46) 0.74 (0.50 - 1.10)
Upper 0.54 (0.28 - 1.05) 0.70 (0.40 - 1.22) 0.61 (0.33 - 1.15) 0.78 (0.46 - 1.34)
Province
Others 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Ontario 1.39 (0.95 - 2.04) 2.08 (1.49 - 2.90) 1.44 (0.99 - 2.10) 2.17 (1.56 - 3.02)
Quebec 1.44 (0.97 - 2.16) 1.33 (0.92 - 1.92) 1.38 (0.91 - 2.07) 1.30 (0.89 - 1.92)
British Columbia 1.28 (0.76 - 2.13) 1.50 (1.02 - 2.21) 1.23 (0.73 - 2.08) 1.54 (1.04 - 2.27)
Ethnicity
White 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Ethnic minorities 1.00 (0.56 - 1.77) 0.45 (0.23 - 0.87) 1.00 (0.56 - 1.79) 0.43 (0.22 - 0.82)
Has a regular doctor
Yes 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
No 0.96 (0.62 - 1.50) 0.56 (0.34 - 0.92) 0.96 (0.62 - 1.50) 0.55 (0.33 - 0.89)
Had an STI in the past
No 1.00 (reference) 1.00 (reference) 1.00 (reference) 1.00 (reference)
Yes 2.72 (1.07 - 6.92) 3.09 (1.69 - 5.66) 2.44 (1.02 - 5.85) 3.08 (1.71 - 5.56)
Geographic Variables Material deprivation quintile
1st (least deprived)
2nd
3rd
4th
5th (most deprived)

1.00 (reference) 1.00 (reference)
0.70 (0.39 - 1.26) 1.09 (0.71 - 1.66)
0.60 (0.33 - 1.08) 0.83 (0.56 - 1.24)
0.59 (0.35 - 1.02)
0.61 (0.34 - 1.08)*
1.11 (0.73 - 1.71)
0.62 (0.38 - 1.00)**


Social deprivation quintile
1st (least deprived)
2nd
3rd
4th
5th (most deprived)

1.00 (reference) 1.00 (reference)
1.39 (0.82 - 2.35) 1.08 (0.68 - 1.70)
1.72 (0.96 - 3.09) 1.11 (0.68 - 1.79)
1.35 (0.77 - 2.37)
2.12 (1.25 - 3.61)†
1.57 (1.00 - 2.49)
1.42 (0.89 - 2.27)††

*Test for trend p=0.07, ** Test for trend p=0.09.
†Test for trend p=0.02, †† Test for trend p=0.04.

Page 8

Assessing the Role of Individual and Neighbourhood Characteristics in HIV Testing The Open AIDS Journal, 2009, Volume 3 53
industrialised countries [21-23]. In our population, although
ethnic minority men had similar HIV testing rates as white men,
the ethnic minority women reported significantly lower testing
rates than white women. Gardezi and coworkers found that
ethnic minority women in Toronto did not consider themselves
at risk partly because of their religious beliefs and cultural
norms [24]. Although we did not have information on
perception of risk, HIV testing was lowest in ethnic minority
women. These quantitative findings may echo qualitative
reports risk perception may differ in ethnic minority men and
women. Another aspect of healthcare access is the cultural
relevance of these services for ethnic minorities [25]. The
Toronto Public Health Department, for example, provides HIV
counselling services in 16 languages other than the two official
languages [26]. Culturally sensitive programmes may
potentially increase the comfort and eventually the access for
HIV related services [27, 28].

testing during pregnancy (19%). Indeed, women not having
access to a regular doctor had lower HIV testing than those
having a regular doctor, although a similar effect was not seen
in men. Thus, potentially most of the HIV testing in these
women appears to be physician driven. Massive community
outreach programmes including HIV testing may increase the
proportion of women undergoing an HIV test [29].
The overall higher HIV testing in women may be due to

structural access barriers may potentially be responsible for low
rates of testing. Spatial analysis of HIV services in Toronto
neighbourhoods demonstrated that preventive services were
concentrated in downtown areas and were less accessible in
other areas [30]. Similarly another study from Toronto found
fewer HIV services in economically disadvantaged and
immigrant neighbourhoods [15]. These findings support our
results that men and women living in most materially deprived
neighbourhoods were less likely to report HIV testing. It has
also been reported that people who have strong community ties
may not access these services to avoid the ‘gossip’ in these
communities [24]. Particularly, ethnic minorities may not go for
HIV testing due to the fear of the stigma associated with it [5].
In our study, those living in most socially deprived
neighbourhoods (increased social isolation) were more likely to
report a test than those living in least socially deprived
neighbourhoods. There is weak negative correlation at area level
between social and material deprivation. These measures are
proxies for different aspects of socio-economic conditions at the
local level, and have been demonstrated to vary independently
[17]. Thus improving the services in deprived neighbourhoods
may potentially increase the HIV testing in individuals living in
these areas. Stronger social cohesion and community ties in less
socially deprived neighbourhoods should be used to improve
the HIV prevention and care messages among individuals living
in these areas.
Apart from individual access to physician services;

high risk behaviour is also a useful strategy to improve HIV
testing. In our study, although, individuals with an STI had a
higher proportion of HIV tests than the general population, only
36% of male and 44% of the females had had an HIV test.;
however, the overall numbers for presence of an STI were small
Targeting individuals who access health care services with
and these results should be evaluated in this context. Though the
Canadian policy is to offer an HIV test to anyone with known
risk behaviour [31] this is potentially an important area of
intervention to improve HIV testing services among individuals
at risk. Further, providing easy access to anonymous testing
facilities may help to improve the testing in indiviudals at risk
[3].

limitations. Though we used the presence of an STI in the past
two years as a marker of high risk behaviour, we did not have
information on the sexual behaviours or sexual preferences of
individuals; a previous study though has demonstrated that HIV
testing is higher in individuals who were more at risk [32]. Lack
of information on risk behaviours other than STIs may be
another potential limitation of the study. Ethnic minority status
is not a homogenous entity - they represent multiple groups
which may be missed in this category [33]. The area data
applied to respondents’ place of residence in 1996/7 were
collected 5 years later than the survey data, which may have
resulted in some inaccuracy in our estimate of likely area
conditions in 1996/7. Data on HIV testing were collected in
1996/7, making them of historic rather than contemporary
relevance. However, more recent data do suggest that ethnic
minority women do not consider themselves to be at risk for
HIV and fewer HIV services are available in economic deprived
areas; this may influence testing [15, 24]. Since, NPHS is an
ongoing survey; it will be useful to collect recent information on
HIV testing again, and assess the changes, if any, in the testing
patterns. Interestingly, it was around 1996 that highly active
antiretroviral therapy was introduced in care of HIV patients,
and massive information and treatment campaign was initiated
globally. Further, Rapid tests for HIV were approved in 2005 in
Canada for “Point of Care settings” [34]. This may have
resulted in changes in the awareness and testing levels. Thus,
recent information on HIV testing could then be used to
compare the testing behaviours over time and design our
prevention and care programmes accordingly.
As is the case with many studies, this study also had its

information on HIV testing in Canada. First, the results show
that living in economically deprived neighbourhoods was
associated with lower HIV testing. These economic disparities
should be taken in account while designing future prevention
strategies and improving access for these individuals. Second,
though HIV positivity has increased in ethnic minorities in the
past few years as per National statistics in Canada [1], testing
for HIV in these population groups was relatively low,
particularly for women. Thus, ethnic minorities, particularly
women, according to our results should have been the focus of
active public health interventions. Assuming similar conditions
still prevail now, this suggests that culturally appropriate and
relevant messages be developed for HIV prevention in ethnic
minorities and campaigns may need to be focused in deprived
areas where minority groups tend to be concentrated. With
increased immigration from Asian and African nations, these
results assume added significance in current HIV prevention
programmes in Canada. Further, though international studies
have discussed disparities in AIDS cases and mortality
according to neighbourhood; [9-14] testing has not been
adequately addressed. Thus, these findings should also be
In spite of the above limitations, the study provides useful

Page 9

54 The Open AIDS Journal, 2009, Volume 3 Setia et al.
explored in nations with high immigration; results from these
neighbourhood analyses will potentially help us formulate our
prevention programmes through better access to HIV testing
facilities.
ACKNOWLEDGEMENTS

Research (MOP 77880, PI AQV). MSS is funded by CIHR-
IHSPR Fellowship and CIHR-RRSPQ Public Health Training
Programme for his doctoral studies at McGill University.
Additional funding was from CIQSS-Matching Grant for 2007.
The study is funded by Canadian Institutes of Health
REFERENCES
[1] Public Health Agency of Canada HIV AIDS (in Canada: Selected
Surveillance Tables to June 30, 2007). Available from: http://www.
phac-aspc.gc.ca/aids-sida/publication/survreport/pdf/tables0607.pdf
[Accessed on 13 August 2009].
UNAIDS, W. UNAIDS/WHO Policy on Statement on HIV
Testing. Available from: http://www.who.int/rpc/research_ethics/
hivtestingpolicy_en_pdf.pdf [Accessed on 13 August 2009] 2004.
Public Health Agency of Canada HIV Testing and Infection
Reporting on Canada. Available from: http://www.phac-aspc.gc.ca/
publicat/epiu-aepi/epi_update_may_04/3-eng.php [Accessed on 13
August 2009] 2004.
Public Health Agency of Canada, HIV/AIDS: Populations at Risk.
Fact Sheet: People Living With HIV/AIDS. Available from:
http://www.phac-aspc.gc.ca/aids-sida/populations_e.html#pl
[Accessed on 13 August 2009] 2008.
Fakoya I, Reynolds R, Caswell G, et al. Barriers to HIV testing for
migrant black Africans in Western Europe. HIV Med 2008; 9(2):
23-5.
Squires KE. Gender differences in the diagnosis and treatment of
HIV. Gend Med 2007; 4(4): 294-307.
Stein MD, Crystal S, Cunningham WE, et al. Delays in seeking
HIV care due to competing caregiver responsibilities. Am J Public
Health 2000; 90(7): 1138-40.
Gatali M, Archibald C. Women and HIV. BMC Women's Health
2004; 4(1): S27.
Brugal MT, Borrell C, Diaz-Quijano E, Pasarin MI, Garcia-Olalla
P, Villalbi JR. Deprivation and AIDS in a southern European city:
different patterns across transmission group. Eur J Public Health
2003; 13: 259-61.
Gabrysch S, Edwards T, Glynn JR. The role of context:
neighbourhood characteristics strongly influence HIV risk in young
women in Ndola, Zambia. Trop Med Int Health 2008; 13(2): 162-
70.
Mari-Dell'Olmo M, Rodriguez-Sanz M, Garcia-Olalla P, et al.
Individual and community-level effects in the socioeconomic
inequalities of AIDS-related mortality in an urban area of southern
Europe. J Epidemiol Community Health 2007; 61(3): 232-40.
Msisha WM, Kapiga SH, Earls FJ, et al. Place matters: multilevel
investigation of HIV distribution in Tanzania. AIDS 2008; 22(6):
741-8.
Wallace R, Wallace D. Socioeconomic determinants of health:
community marginalisation and the diffusion of disease and
disorder in the United States. BMJ 1997; 314(7090): 1341-5.
Zierler S, Fullilove M, Fullilove R, et al. Economic deprivation and
AIDS incidence in Massachusetts. Am J Public Health 2000; 90(7):
1064-73.
Kaukinen C, Fulcher C. Mapping the social demography and
location of HIV services across Toronto neighbourhoods. Health
Soc Care Community 2006; 14(1): 37-48.
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16] Statistics Canada, National Population Health Survey. Available
from: http://www.statcan.ca/cgi-bin/imdb/p2SV.pl?Function=get
Survey&SurvId=3225&SurvVer=0&InstaId=15280&InstaVer=5&
SDDS=3225&lang=fr&db=IMDB&dbl=E&adm=8&dis=2 [Acces-
sed on 13 August 2009] 2008.
Pampalon R, Hamel D, Raymond G. Indice de défavorisation pour
l’étude de la santé et du bien-être au Québec – Mise à jour 2001.
Institut National de Santé Publique au Québec: Québec. 2004,
Accessed from: http://www.inspq.qc.ca/pdf/publicat ions/295-
IndiceDefavorisation_2001.pdf
Gonthier D, Hotton T, Cook C, Wilkins R. Fusion des données de
recensement par region et des données d’enquête dans les centre de
données de recherché de Statistique Canada. Bull Tech et d’Inf
2006; 3(1): 21-40.
Vittingghoff E, Glidden DV, Shiboski SC, McCulloch CE.
Regression Methods in Biostatistics. 2004, New York, NY, USA:
Springer Science+Business Media, Inc 2004; pp. 1-333.
de Wit JB, Adam PC. To test or not to test: psychosocial barriers to
HIV testing in high-income countries. HIV Med 2008; 9(Suppl 2):
20-22.
Anderson J, Melville R, Jeffries DJ, et al. Ethnic differences in
women with HIV infection in Britain and Ireland. The study group
for the mrc collaborative study of HIV infection in women. AIDS
1996; 10(1): 89-93.
Coenen T, Lundgren J, Lazarus JV, et al. Optimal HIV testing and
earlier care: the way forward in Europe. HIV Med 2008; 9(Suppl
2): 1-5.
Saul J, Erwin J, Bruce JC, et al. Ethnic and demographic variations
in HIV/AIDS presentation at two London referral centres 1995-9.
Sex Transm Infect 2000; 76(3): 215.
Gardezi F, Calzavara L, Husbands W, et al. Experiences of and
responses to HIV among African and Caribbean communities in
Toronto, Canada. AIDS Care 2008; 20(6): 718-25.
Giger JN, Davidhizar R. Promoting culturally appropriate
interventions among vulnerable populations. Annu Rev Nurs Res
2007; 25: 293-316.
Toronto Public Health. AIDS & Sexual Health InfoLine. Available
from: http://www.toronto.ca/health/ai_index.htm [Accessed on 13
August 2009] 2008.
Korner H. Late HIV diagnosis of people from culturally and
linguistically diverse backgrounds in Sydney: the role of culture
and community. AIDS Care 2007 19(2): 168-78.
McMahon T, Fairley CK, Donovan B, et al. Evaluation of an ethnic
media campaign on patterns of HIV testing among people from
culturally and linguistically diverse backgrounds in Australia. Sex
Health 2004; 1(2): 91-4.
Center for Disease Control and Prevention, Rapid HIV testing in
outreach and other community settings--United States, 2004-2006.
MMWR Morb Mortal Wkly Rep 2007; 56(47): 1233-7.
Fulcher C, Kaukinen C. Mapping and visualizing the location HIV
service providers: an exploratory spatial analysis of Toronto
neighborhoods. AIDS Care 2005; 17(3): 386-96.
Public Health Agency of Canada, Human Immunodeficiency Virus
Infections. Ottawa, Canada. Available from: http://www.phac-
aspc.gc.ca/std-mts/sti_2006/pdf/508_HIV.pdf [Accessed on 13
August 2009] 2008; pp. 1-17.
Houston S, Archibald CP, Strike C, et al. Factors associated with
HIV testing among Canadians: results of a population-based
survey. Int J STD AIDS 1998; 9(6): 341-6.
Bauder H. Visible minorities and urban analysis. Can J Urban Res
2001; 10(1): 69-90.
Canadian AIDS Society: Société Canadienne du SIDA. Rapid HIV
Testing in Canada. 2007 [cited 2009 21 August]; Available from:
http://www.cdnaids.ca/web/backgrnd.nsf/ pages/cas-gen-0142.
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]

?
Received: June 1, 2009

Revised: August 12, 2009 Accepted: August 23, 2009
© Setia et al.; Licensee Bentham Open.

This is an open access article licensed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/
3.0/) which permits unrestricted, non-commercial use, distribution and reproduction in any medium, provided the work is properly cited.