HIV prevention and risk reduction interventions in prison settings are a major public health concern and a critical political issue 1 . Infectious diseases are more prevalent in prison than in the general population and many reports indicate that prison stay is an independent risk factor for the transmission of blood-borne viruses' 2 3 4 . Risky behaviors are frequent while the dramatic prevalence of psychiatric disorders and the prison context (lack of hygiene, promiscuity, sexual violence, overpopulation and violence) exacerbate the risks 4 5 6 7 . Many inmates cycle in and out of prison repeatedly, increasing the likelihood that any 4 infections contracted in prison could soon affect the general community 2 8 .

In France, the prison population has increased by one third in the last 10 years, partly due to the criminalization of drug use. In 2008, 14.3% of sentences were for drug-related offences, 36% of which involved prison sentences 9 .

The Ministry of Health has been responsible for health in French prisons since 1994. National guidelines 10 recommend equity in access to prevention measures and continuity of care between the community and prisons and back to the community. However, the question of access to prevention measures for HIV and other infectious diseases is not referred to in any great detail in these guidelines. Such access is referred to in a 1996 Ministry of Health/Ministry of Justice joint circular 11 regulating HIV prevention in prison settings in which several harm reduction (HR) measures available in the general community (including needle syringe exchange programs) are not permitted.

The main objective of this national survey is to estimate the level of infectious risk by measuring to what extent French prisons adhere to international guidelines for HIV prevention in prison and to national policies, as regulated by the 1996 circular 11 and national guidelines 10 . A secondary objective is to identify which particular characteristics of prison settings predict non-adherence to national and WHO guidelines.

A nationwide survey targeting all French prisons was conducted between November 2009 and May 2010. A questionnaire focusing primarily on access to infectious disease prevention and HR measures for French prisoners was first sent to the heads of medical (all prisons) and psychiatric (26 prisons) units (see additional file 1: ANRS-PRI² DE Inventory Questionnaire). Additional detailed information about issues regarding access to HIV prevention was gathered through a structured phone interview. More specifically the questionnaire collected data about characteristics of the prisons with 10 sections, accounting for a total of 46 items, each exploring access to a specific prevention measure.

We defined "international guidelines for HIV prevention in prison" as those recommendations provided by WHO, in collaboration with UNAIDS and UNODC in their document entitled "Effectiveness of interventions to address HIV in prisons" which includes interventions for preventing not only HIV but other infectious diseases in prison settings 4 . This WHO document was chosen from among several other tools and reports as it is the most comprehensive document on HIV prevention in prison settings.

Two scores of adherence to national and international guidelines were built in order to both evaluate the implementation of HIV prevention and other HR measures in French prisons and to estimate the level of infectious risk. Table 1 reports items corresponding to national and international guidelines and how adherence to each specific recommendation was scored. Finally, two global scores were built, for national and international guidelines respectively, by summing the subscores corresponding to each recommendation. Each subscore was dichotomous (0 = non- adherence, 1 = adherence) with the exception of the scores for opioid substitution treatment and access to condoms. The corresponding scores for these latter items varied between 0 and 2 in order to give them a higher weighting, as their effectiveness in HIV prevention has already been established 4 .

Potential score ranges for measuring adherence were as follows: 0 to 10 for national guidelines and 0 to 9 for international guidelines. For each subscore, we computed the proportion of prisons adherent to national and WHO guidelines as well as their 95% confidence intervals.

A linear regression model was used to assess the relationship between each prison characteristic and the level of adherence to national and WHO guidelines.

The National Agency for Research on Aids and Viral Hepatitis (ANRS) committee granted approval for the project. As only aggregated and openly available data were used in the study, authorization from the Commission Nationale Informatique et Liberté was not required.

No significant difference was found between the structural characteristics of prisons included in the study and those excluded.

Among the 103 selected prisons, covering 43 365 incarcerated subjects (69%), 62% are remand centers (RC), 13% are both RC and prisons for persons sentenced (PPS), and 25% are PPS only (including 5 security prisons). Seventy four (72%) are male only prisons, 1 is a female only prison, and 28 are mixed male and female prisons. Among the 103 prisons, 24% receive juvenile offenders. The mean number of prisoners per prison is 421, ranging from 36 to 3785.

This study, the first conducted in France on access to HIV prevention interventions and availability of HR tools in prison settings clearly shows a wide gap between national & international policies and local practices. Adherence to national guidelines on availability of information and access to HIV prevention and HR measures is very poor, and preventive measures like NSP promoted by WHO guidelines are absent, even as local initiatives.

The lack of structural predictors of adherence suggests that practices depend on the ideology of the local administration even if they are not evidence-based.

When focusing on each specific tool for HIV prevention or HR, the heterogeneity of practices and gaps between national and international policies becomes more striking: such is the case for bleach distribution, NSP, PEP and HBV vaccination.

Bleach effectiveness as a disinfectant for inactivating HIV and particularly HCV is limited in prison settings 4 . Bleach distribution does not meet the regulations expressed by the French 1996 circular in two thirds of prisons and the information about its use as a HR tool is often absent. This may be due to the shared responsibility with the criminal justice administration regarding the distribution of information, to insufficient training in health care units but also to a lack of a comprehensive and clear HR policy in French prisons 12 .

Condoms are almost always available in French prison settings but in nearly half of the cases without lubricants and in three quarters of the prisons, access is restricted to medical units where confidentiality is limited. Sexuality in prison is still a taboo because it refers to homosexuality, which is an especially stigmatized practice in prison. Although prison OST coverage is progressively scaling-up (2% of the prison population in 1998 - 2 years after the Marketing Authorization Application in France - reaching 8.9% in 2010), this is more probably a reflection of increased coverage in the French opioid dependent population. OST initiation remains problematic and the use of a ceiling dosage prescription which is not consistent with official guidelines is often observed. Buprenorphine crushing or diluting practices to reduce the risk of diversion, underscore the difficulties met by health staff when trying to balance the risk of diversion with adequate care for opioid dependence, and suggest that fear of buprenorphine diversion in prisons may represent a major barrier to its access.

BBV testing at prison entry is widely proposed but according to a previous survey of prison settings, only a small percentage of prisoners agree to be tested 12 . The proportion of inmates vaccinated against HBV on prison entry increased from 13.7% in 1997 to 31.3% in 2003, probably thanks to the HBV vaccination campaign in the general population 12 . As a history of incarceration has consistently been found to be a major predictor of HBV seropositive status 13 14 , accelerated strategies based on injection at days 0, 10 and 21 which are easy to perform and remain effective, need to be urgently introduced in prison settings.

The currently poor access to and implementation of PEP interventions in prisons is unfortunate and this is one of the fields, together with HBV vaccination, which deserves major attention. Information for prisoners about the existence of such HR tools is rarely available and no such treatment is prescribed for risky injection practices.

At the international level, only a few studies on the accessibility of HIV prevention and HR measures in prison settings are available. In terms of bleach, NSPs and condoms, most associated studies are from Australia and Canada, where confidentiality or anonymity are considered as crucial points 15 16 17 18 . NSPs are available in more than 50 prisons in 12 countries worldwide, mostly in Europe 3 . In Germany, one evaluation reported only a small reduction in syringe sharing in one program, due to insufficient supply of needles and syringes, access without anonymity and inadequate provision of correct sized syringes 4 . Condoms, in theory, are widely accessible in many countries, except in the United States where they are available in less than 1% of prisons 4 . In reality, condoms are often not accessible for different reasons 19 , lack of anonymity being a frequently cited example 4 .

One major advantage of our survey is its capacity to investigate not only accessibility to tools for reducing the risk of infection in prison but also the general information transmitted to inmates about the availability of HR tools. It is important to underline that wide scale care assessment in prison settings is rare, because it is particularly difficult to carry out. Official authorization, funding and agreement from ethics committees are only rarely obtained 20 . A second advantage is that the study can be easily replicated in other countries for international comparisons and guide a global policy reform to improve HIV prevention and general care in prison settings.

Some limitations need to be acknowledged. Although self-reports and interviews serve as the only feasible methods to study access to prevention, they may be affected by social desirability bias. However, while the risk of over-reporting availability of preventive measures was high in our study, adherence rates were rather low, suggesting that true rates are likely to be even lower. Although no significant difference was found between the structural characteristics of prisons included in the study and those excluded, it is possible that the former could be more concerned by prevention/harm reduction issues. Accordingly, we can presume that we selected those more likely to be adherent to national or international guidelines. This means that the level of adherence to national and WHO guidelines is likely to have been overestimated, which in turn strengthens our conclusions.

Secondly, it is also possible that physicians may not have been sufficiently informed about the interventions promoted by the prison administration in which they worked, but this could only be the case for bleach distribution, whose access depends on the penitentiary administration. Although the provision of care to reduce harms from injecting practices is frequently reported, drug injection remains a taboo practice in prison settings because drug use is criminalized both in prison and in the general community. This criminalization contributes to making the introduction of NSPs useless and hindering the full implementation of initiatives for reducing harm in the injecting drug inmate population. Despite these issues, in a French survey conducted among drug users in the community, 12% reported injecting practices in prison settings 21 .

The failure in the implementation of a global HR strategy in prison settings can be attributed to many causes, but the primary one is that criminal justice laws and drug prohibition still prevail over the public health objectives of equity of access to prevention and care, both in prison and in the general community.

In the last ten years, the reinforcement of the criminalization of drug use has had a twofold effect. First it constitutes a barrier to the implementation of an effective HR policy in the community, decriminalization being associated with a reduced HIV prevalence among drug users 22 .

Secondly, the criminalization of drug use has contributed to overpopulation in prisons. This has not been counterbalanced by the adequate implementation of preventive measures to control the increased risk of infection due to promiscuity or by an increase in health care funds which could improve the coordination and delivery of existing HR interventions. More globally with regard to health and hygiene, prison conditions for French inmates are so deplorable that interventions to reduce harm from drug use are probably not considered as a priority for health care providers and administrative professionals working in prison settings.

On the other hand, national and international guidelines are mainly designed for HIV prevention and are rapidly becoming obsolete in terms of the emerging risk of other infectious diseases like tuberculosis, hepatitis or sexually transmissible diseases. The control of HCV in correctional settings needs implementation of additional preventive measures and/or of a HR package. In terms of health policy, it would be more valuable if national and international guidelines could anticipate policies to control emerging public health problems in correctional settings.

A policy reform for HIV prevention and HR in prison settings is a priority but needs to be incorporated into a wider national health policy reform to improve the general health and quality of life of French prisoners as well as equalizing access to care and prevention in prisons and the general community. A drug policy reform decriminalizing drug use seems to be the sine-qua-non condition for replacing incarcerations for the use of drugs with improved access to prevention and care for drug dependence.

HR will not be more effective in prison settings unless innovative interventions and adequate funding are introduced to motivate prison health and administrative staff, improve their working environment and finally relieve the debilitating somatic and psychological conditions of French prisoners.

ANRS: National Agency for Research on Aids and Viral Hepatitis; WHO: World Health Organization; OST: Opioid Substitution Treatment; HBV: Hepatitis B Virus; PEP: Post-Exposure Prophylaxis; HIV: Human Immunodeficiency Virus; HR: Harm Reduction; UNAIDS: United Nations programme on AIDS; UNODC: United Nations Office on Drugs and Crime; RC: Remand Center; PPS: Prison for Persons Sentenced; HCV: Hepatitis C Virus; IST: Infection Sexually Transmissible; NGO: Non-Governmental Organization; BBV: Blood Born Viruses; NSP: Needle Syringe Program;

The authors declare that they have no competing interests.

LM designed the study, supervised the design of the questionnaire and data collection, and wrote and revised the manuscript. MJR contributed to the design of the study and the questionnaire and revised the final version of the manuscript. OM and CC contributed to the design of the questionnaire, collected the data and contributed to the data analysis. JB and JC conducted the data analysis in and contributed to the results section. MPC contributed to the design of the study, wrote the manuscript and edited the final version. All authors read and approved the final manuscript.