Prison rites: starting to inject inside

BMJ 1995; 311 doi: (Published 28 October 1995) Cite this as: BMJ 1995;311:1135
  1. Sheila M Gore, senior statisticiana,
  2. A Graham Bird, consultantb,
  3. Amanda J Ross, statisticianc
  1. aMRC Biostatistics Unit, Cambridge CB2 2SR
  2. bDepartment of Immunology, Churchill Hospital, Oxford OX3 7LJ
  3. cMRC-BIAS, Centre for HIV Research, Edinburgh EH9 3JN
  1. Correspondence to: Dr Gore.

    In 1993 an outbreak of HIV infection occurred within Glenochil prison, caused by sharing of infected needles.1 To determine the nature of injecting behaviour within prison we performed surveys in two Scottish prisons, Glenochil and Barlinnie, which combined voluntary anonymous testing of saliva samples for HIV and completion of a linked questionnaire asking about risk factors.2

    Subjects, methods, and results

    The surveys were performed in Glenochil prison in July 19943 and in Barlinnie prison in September 1994.4 Seventy five questionnaires in Glenochil and 327 in Barlinnie were from injector-inmates; 25% of injectors in Glenochil (18/72, 95% confidence interval 15% to 35%) and 6% (20/319; 3% to 9%) of Barlinnie injectors reported that they had started to inject inside a prison.2 4 Half the prisoners, and three quarters of injectors, came from Glasgow. Barlinnie is a local prison for the Glasgow area, whereas Glenochil holds men serving longer sentences from throughout Scotland. Self reported information from injectorinmates was pooled to inquire into the characteristics of the 38 men who started to inject inside prison.

    A third (23/72) of Glenochil's injector-inmates in July 1994 had injected in Glenochil prison between January and June 1993.2 3 Starting to inject inside was acknowledged by 2/72 injector-inmates in Barlinnie who first injected before 1983; by 8/159 who began in 1983-8; and by 10/88 who first injected after 1988. Nine (17%) out of 53 Barlinnie injectors whose sentence began in 1993 or earlier had started to inject inside. Only 5% of 245 Glasgow injector-inmates had started to inject inside but 11% had from elsewhere (8/73). Four per cent of Barlinnie's injector-inmates (12/324) had injected in Glenochil prison during January to June 1993, five having started in prison.

    Injector-inmates from outside Glasgow were more likely than Glaswegians to have started to inject inside (table: 1n odds of −1.1, SE 0.4), as were those whose injecting career began most recently (after 1988) (1n odds trend: −0.66, SE 0.27). Injector-inmates who injected in Glenochil prison between January and June 1993 included disproportionately many who had started to inject in prison (odds ratio of 8:1). Injectorinmates whose sentences began in 1993 were the most likely to have started to inject in prison (odds ratio of 4:1) and remand prisoners were least likely. The 95% confidence interval (0.5 to 4.8) for the odds by prison on having started to inject inside included one. Covariate adjustment had thus largely accounted for the differential (25% v 6%) in starting to inject inside, but the wide interval does not rule out an establishment effect.

    Linear logistic risk score for an injector-inmate having started to inject in prison

    View this table:

    In the combined dataset 238/1212 men (20%) had never been inside prison before and were non-injectors when received into prison, but four (1 in 60, or 1.7%) started injecting during their first sentence. This disturbing rate of conversion to injecting predicts over 100 new injectors from 7000 estimated first admissions to prison per year of non-injecting males in Scotland.


    Inmates from a needle exchange area like Glasgow will continue to inject when imprisoned and may introduce other prisoners to injecting. Various scenarios about HIV transmission in Glenochil prison and starting to inject inside can be considered. Firstly, individuals who first injected inside and were likely to continue during future sentences may have been together in Glenochil. Secondly, non-injectors convicted in 1993 from outside Glasgow, held together in Glenochil with Glasgow men, might have begun to inject there. Thirdly, some change in policing or sentencing policy might have resulted in the conviction of men adept at introducing drugs and injecting equipment into prison and initiating the practice in a group of novices.

    Our data are hypothesis generating, not conclusive. They warn that each year more than 100 men in Scotland may start injecting drugs in the high risk, needle sharing environment of prison. The problem is not a new one5 but has recently worsened--17% of injector-inmates who first injected after 1988 had started while in prison (19/114)--and has serious implications for transmission of bloodborne viruses. Prisons need to understand how initiation occurs and to support non-injectors, particularly those never imprisoned before, so that they can avoid starting to inject inside.

    We pay tribute to inmates' willingness to contribute to public health information, and to the Scottish Prison Service's commitment to evidence based determination of health policy in prisons.


    • Funding The Scottish Centre for Infection and Environmental Health and the Scottish Prison Service.

    • Conflict of interest None.


    View Abstract