Objective: To determine prevalence of HIV infection and risk behaviour
s among male inmates of Her Majesty's Prison (HMP) Barlinnie, Glasgow,
Scotland on 8-9 September 1994. Design: Cross-sectional study: volunt
ary, anonymous HIV surveillance (using saliva samples) of all inmates
and linked self-completion risk-factor questionnaire. Subjects: Of 107
3 prisoners available to participate, 985 (92%) completed a risk-facto
r questionnaire and 982 salivettes were received for testing, of which
978 were tested for HIV antibodies (four were dry samples); 928 quest
ionnaires passed logical checks for consistency. Main outcome measures
: HIV prevalence on saliva testing, related risk behaviours and ratio
of overall-to-disclosed HIV prevalence. Proportion of all inmates who
have ever injected drugs, ever injected inside prison, started injecti
ng inside, ever had acute hepatitis, had a recent personal HIV test (s
ince January 1993). Results: Nine saliva samples [eight injecting drug
users (IDU), one recognized other risk] out of 978 were HIV-antibody-
positive (three presumptively from known HIV-infected inmates). Overal
l HIV prevalence was estimated at 1% compared with a known prevalence
of 0.4%, giving an overall-to-disclosed HIV prevalence ratio of 2.6 in
HMP Barlinnie in September 1994. A higher proportion of prisoners fro
m Glasgow (48%) than elsewhere (19%) were IDU. Year of first injection
was also different by residence with 23% of Glasgow IDU having first
injected after 1988 compared with 45% of IDU from elsewhere, mainly We
st and South Scotland. Half the IDU inmates reported having injected w
hile incarcerated and 6% had started to inject while incarcerated. Ten
per cent of all prisoners and 20% of IDU inmates had had a personal H
IV test since January 1993. Logistic regression showed that there was
a significant deficit of recent HIV test uptake by Glasgow residents (
odds ratio, 0.5; 95% confidence interval, 0.27-0.89), that IDU were mo
re likely to have had treatment for a sexually transmitted disease, an
d that IDU who had injected inside and those whose injecting career be
gan prior to 1989 were more likely to have had acute hepatitis. Conclu
sion: A consistent harm-reduction policy is needed across prisons in t
he United Kingdom to avoid transmission of blood-borne viral infection
s. Drug injecting inside prison is common, a proportion of IDU inmates
having first injected drugs while in prison, and much higher rates of
hepatitis have been reported in association with injecting while inca
rcerated compared with that for IDU who only injected outside prison.