Series www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 1 HIV and related infections in prisoners 1 Global burden of HIV, viral hepatitis, and tuberculosis in prisoners and detainees Kate Dolan, Andrea L Wirtz, Babak Moazen, Martial Ndeffo-mbah, Alison Galvani, Stuart A Kinner, Ryan Courtney, Martin McKee, Joseph J Amon, Lisa Maher, Margaret Hellard, Chris Beyrer, Fredrick L Altice The prison setting presents not only challenges, but also opportunities, for the prevention and treatment of HIV, viral hepatitis, and tuberculosis. We did a comprehensive literature search of data published between 2005 and 2015 to understand the global epidemiology of HIV, hepatitis C virus (HCV), hepatitis B virus (HBV), and tuberculosis in prisoners. We further modelled the contribution of imprisonment and the potential impact of prevention interventions on HIV transmission in this population. Of the estimated 10·2 million people incarcerated worldwide on any given day in 2014, we estimated that 3·8% have HIV (389 000 living with HIV), 15·1% have HCV (1 546 500), 4·8% have chronic HBV (491 500), and 2·8% have active tuberculosis (286 000). The few studies on incidence suggest that intraprison transmission is generally low, except for large-scale outbreaks. Our model indicates that decreasing the incarceration rate in people who inject drugs and providing opioid agonist therapy could reduce the burden of HIV in this population. The prevalence of HIV, HCV, HBV, and tuberculosis is higher in prison populations than in the general population, mainly because of the criminalisation of drug use and the detention of people who use drugs. The most effective way of controlling these infections in prisoners and the broader community is to reduce the incarceration of people who inject drugs. Introduction From the beginning of the AIDS epidemic in 1981, the association between HIV, tuberculosis, and prisons was apparent, 1 with HIV responsible for a steep rise in tuberculosis in US prison populations. 2 This is important because the prevalence of HIV in prisons in many countries is high, with one review reporting levels greater than 10% in 20 low-income and middle-income countries. 3 Several factors have a role in the epidemics of HIV, tuberculosis, and related infections in prisons. 4 Many individuals who are most likely to be incarcerated are at greatest risk of these infections, whether because of injection drug use for HIV and viral hepatitis or poverty and overcrowding for tuberculosis. Drug injection is common in prison inmates, ranging from 2% to 38% in Europe, 34% in Canada, and up to 55% in Australia, in stark contrast with the percentage in the general population, estimated at 0·3% in the European Union and 0·2% in Australia. 5 Prisons provide many opportunities both for the spread 4 and prevention of these infections. 6 The situation is complicated further by the expansion of parallel prison systems for those suspected of drug use in at least 27 countries. These compulsory drug detention centres operate extrajudicially and often under the guise of drug treatment (panel 1). 20 Punishment and inhumane conditions are widespread, but evidence-based treatment for drug dependence and infectious diseases is rare or non-existent. 7,21 However, prisons not only pose a threat to the health of people incarcerated within them. They also pose a risk to staff and to the population at large, because detainees are not a static population, but move around the prison system and back and forth from the outside world. The risks particularly lie at the interface between prisons and society outside. In the USA, HIV incidence is highest in detainees who were released and re-incarcerated compared with continuously incarcerated prisoners, people who inject drugs with no history of incarceration, and men who have sex with men (MSM; panel 2). 4 The period immediately after release is especially risky for receptive syringe sharing, acquisition of HIV and hepatitis C virus (HCV), and mortality. 29–32 Thus, the transition between the prison and community settings represents a high-risk Published Online July 14, 2016 http://dx.doi.org/10.1016/ S0140-6736(16)30466-4 This is the first in a Series of six papers on HIV and related infections in prisoners National Drug and Alcohol Research Centre (Prof K Dolan PhD, R Courtney PhD), and Kirby Institute (Prof L Maher PhD), University of New South Wales, Sydney, NSW, Australia; Center for Public Health and Human Rights, and Department of Epidemiology, Johns Hopkins School of Public Health, Baltimore, MD, USA (A L Wirtz PhD, Prof C Beyrer MD); Non-Communicable Diseases Research Center, Endocrinology and Metabolism Population Sciences Institute, Tehran University of Medical Sciences, Tehran, Iran (B Moazen MScIH); Center for Infectious Disease Modeling and Analysis (M Ndeffo-mbah PhD, Prof A Galvani PhD), and Epidemiology and Public Health (Prof F L Altice MD), Yale University, New Haven, CT, USA; Griffith Criminology Institute & Menzies Health Institute Queensland, Griffith University, Brisbane, QLD, Australia (Prof S A Kinner PhD); Department of Health Services Research and Policy, London School of Hygiene & Tropical Medicine, London, UK (Prof M McKee DSc); Health and Human Rights Division, Human Rights Watch, New York, NY, USA (J J Amon PhD); and Centre for Population Health, Burnet Institute, Melbourne, VIC, Australia (Prof M Hellard PhD) Correspondence to: Prof Kate Dolan, National Drug and Alcohol Research Centre, University of New South Wales, Sydney, NSW 2052, Australia [email protected]Key messages • Prevalence of HIV, HCV, HBV, and tuberculosis is higher in prison populations than in the general population, mainly because of the criminalisation of drug use and the detention of people who inject or use drugs • We strongly support the UN’s 2012 call to close compulsory drug detention centres and expand voluntary, evidence-based treatment in the community • Mathematical modelling suggests that incarceration and re-incarceration of people who inject drugs contributes to the overall HIV epidemic and a reduction in incarceration of this population will reduce the incidence of HIV • Evidence-based prevention and treatment such as opioid agonist therapy and antiretroviral therapy can substantially reduce the incidence of HIV, HCV, and HBV, and reduce drug dependence in this population • Responses to co-infection with HIV and tuberculosis should include an integrated, patient-centred model of prevention and care, with systematic screening of high-risk groups and equitable access to effective treatment • The most effective way of controlling infection in prisoners and the broader community is to reduce mass incarceration of people who inject drugs
79
Embed
Series HIV and related infections in prisoners 1 Global burden ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Series
www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 1
HIV and related infections in prisoners 1
Global burden of HIV, viral hepatitis, and tuberculosis in prisoners and detaineesKate Dolan, Andrea L Wirtz, Babak Moazen, Martial Ndeffo-mbah, Alison Galvani, Stuart A Kinner, Ryan Courtney, Martin McKee, Joseph J Amon, Lisa Maher, Margaret Hellard, Chris Beyrer, Fredrick L Altice
The prison setting presents not only challenges, but also opportunities, for the prevention and treatment of HIV, viral hepatitis, and tuberculosis. We did a comprehensive literature search of data published between 2005 and 2015 to understand the global epidemiology of HIV, hepatitis C virus (HCV), hepatitis B virus (HBV), and tuberculosis in prisoners. We further modelled the contribution of imprisonment and the potential impact of prevention interventions on HIV transmission in this population. Of the estimated 10·2 million people incarcerated worldwide on any given day in 2014, we estimated that 3·8% have HIV (389 000 living with HIV), 15·1% have HCV (1 546 500), 4·8% have chronic HBV (491 500), and 2·8% have active tuberculosis (286 000). The few studies on incidence suggest that intraprison transmission is generally low, except for large-scale outbreaks. Our model indicates that decreasing the incarceration rate in people who inject drugs and providing opioid agonist therapy could reduce the burden of HIV in this population. The prevalence of HIV, HCV, HBV, and tuberculosis is higher in prison populations than in the general population, mainly because of the criminalisation of drug use and the detention of people who use drugs. The most effective way of controlling these infections in prisoners and the broader community is to reduce the incarceration of people who inject drugs.
IntroductionFrom the beginning of the AIDS epidemic in 1981, the association between HIV, tuberculosis, and prisons was apparent,1 with HIV responsible for a steep rise in tuberculosis in US prison populations.2 This is important because the prevalence of HIV in prisons in many countries is high, with one review reporting levels greater than 10% in 20 low-income and middle-income countries.3 Several factors have a role in the epidemics of HIV, tuberculosis, and related infections in prisons.4 Many individuals who are most likely to be incarcerated are at greatest risk of these infections, whether because of injection drug use for HIV and viral hepatitis or poverty and overcrowding for tuberculosis. Drug injection is common in prison inmates, ranging from 2% to 38% in Europe, 34% in Canada, and up to 55% in Australia, in stark contrast with the percentage in the general population, estimated at 0·3% in the European Union and 0·2% in Australia.5 Prisons provide many opportunities both for the spread4 and prevention of these infections.6
The situation is complicated further by the expansion of parallel prison systems for those suspected of drug use in at least 27 countries. These compulsory drug detention centres operate extrajudicially and often under the guise of drug treatment (panel 1).20 Punishment and inhumane conditions are widespread, but evidence-based treatment for drug dependence and infectious diseases is rare or non-existent.7,21 However, prisons not only pose a threat to the health of people incarcerated within them. They also pose a risk to staff and to the population at large, because detainees are not a static population, but move around the prison system and back and forth from the outside world.
The risks particularly lie at the interface between prisons and society outside. In the USA, HIV incidence is highest
in detainees who were released and re-incarcerated compared with continuously incarcerated prisoners, people who inject drugs with no history of incarceration, and men who have sex with men (MSM; panel 2).4 The period immediately after release is especially risky for receptive syringe sharing, acquisition of HIV and hepatitis C virus (HCV), and mortality.29–32 Thus, the transition between the prison and community settings represents a high-risk
Published Online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
This is the first in a Series of six papers on HIV and related infections in prisoners
National Drug and Alcohol Research Centre (Prof K Dolan PhD, R Courtney PhD), and Kirby Institute (Prof L Maher PhD), University of New South Wales, Sydney, NSW, Australia; Center for Public Health and Human Rights, and Department of Epidemiology, Johns Hopkins School of Public Health, Baltimore, MD, USA (A L Wirtz PhD, Prof C Beyrer MD); Non-Communicable Diseases Research Center, Endocrinology and Metabolism Population Sciences Institute, Tehran University of Medical Sciences, Tehran, Iran (B Moazen MScIH); Center for Infectious Disease Modeling and Analysis (M Ndeffo-mbah PhD, Prof A Galvani PhD), and Epidemiology and Public Health (Prof F L Altice MD), Yale University, New Haven, CT, USA; Griffith Criminology Institute & Menzies Health Institute Queensland, Griffith University, Brisbane, QLD, Australia (Prof S A Kinner PhD); Department of Health Services Research and Policy, London School of Hygiene & Tropical Medicine, London, UK (Prof M McKee DSc); Health and Human Rights Division, Human Rights Watch, New York, NY, USA (J J Amon PhD); and Centre for Population Health, Burnet Institute, Melbourne, VIC, Australia (Prof M Hellard PhD)
Correspondence to: Prof Kate Dolan, National Drug and Alcohol Research Centre, University of New South Wales, Sydney, NSW 2052, Australia [email protected]
Key messages
• PrevalenceofHIV,HCV,HBV,andtuberculosisishigherinprison populations than in the general population, mainly because of the criminalisation of drug use and the detention of people who inject or use drugs
• WestronglysupporttheUN’s2012calltoclosecompulsory drug detention centres and expand voluntary, evidence-based treatment in the community
• Mathematicalmodellingsuggeststhatincarcerationandre-incarceration of people who inject drugs contributes to the overall HIV epidemic and a reduction in incarceration of this population will reduce the incidence of HIV
• Evidence-basedpreventionandtreatmentsuchasopioidagonist therapy and antiretroviral therapy can substantially reduce the incidence of HIV, HCV, and HBV, and reduce drug dependence in this population
• Responsestoco-infectionwithHIVandtuberculosisshouldinclude an integrated, patient-centred model of prevention and care, with systematic screening of high-risk groups and equitable access to effective treatment
• Themosteffectivewayofcontrollinginfectioninprisoners and the broader community is to reduce mass incarceration of people who inject drugs
Series
2 www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
environment, especially for people with substance use disorders.31 This is important because, although an estimated 10·2 million people were incarcerated at any time in 2014, over 30 million individuals transition from prison to the community each year.33 Prisons act as incubators for tuberculosis and HIV, because they are associated with higher levels of infection than in the surrounding populations,3,34 yet many countries have parallel and vertical systems, with fragmented policy responses to these interlinked issues—prisons, HIV, viral hepatitis, and tuberculosis—and interruptions of surveillance and treatment during transitions. This Series paper encourages a coordinated response by reviewing the global epidemiology of HIV, HCV, HBV, and tuberculosis in prison populations.35,36
Disease burden in prisoners and detaineesWe did a comprehensive review of studies of prevalence and incidence data on HIV, HCV, HBV, tuberculosis, and co-infection with tuberculosis and HIV in prisoners and detainees, published between Jan 1, 2005, and Nov 30, 2015, for 196 countries in 2015 (appendix p 2, 4).37
We searched for studies with biological markers of each infection in general prisoners and in people who inject drugs, MSM, female sex workers, and transgender people, in prisons, jails, and compulsory drug detention centres.
Of 11 000 publications identified, 299 met inclusion criteria for the meta-analysis of infections (appendix p 7). These data show substantial heterogeneity in disease burden across regions (figure 1).
Global estimates of prisoners with HIV, HCV, HBV, and tuberculosis infectionOf an estimated 10·2 million people incarcerated worldwide on any given day in 2013,22 we estimated the midpoint of the number infected was 389 000 with HIV (3·8%), 1 546 500 with HCV (15·1%), 491 500 with chronic HBV infection (4·8%), and 286 000 with active tuberculosis (2·8%; appendix pp 9, 15, 20, 24).
Overall, prevalence of all infections was substantially higher in prison populations than in surrounding communities, especially where there are generalised HIV epidemics, such as in sub-Saharan Africa,38 and where there is a high prevalence of injection drug use, such as in eastern Europe and central Asia (figure 1, 2, appendix p 9).39
HIV infectionPooled estimated HIV prevalence in prisoners74 of 196 countries37 had HIV prevalence data in 2015 (200 datapoints) in prisoners (appendix p 8). The regions most affected were the two African regions (east and southern Africa and west and central Africa), which have a high prevalence in the general population, and the two European regions (eastern Europe and central Asia and west Europe), reflecting the over-representation of people who inject drugs in prison—a group with a high prevalence of HIV infection. Transmission via injection drug use also contributes to the HIV epidemic in the Middle East and north Africa and Asia Pacific regions.
While injection drug use is rare in the Caribbean and Latin America, HIV prevalence in prisoners is generally higher than in North America, possibly reflecting the concentrated epidemic of HIV in MSM40 and cocaine users in Latin America.41
Female inmates had a slightly higher prevalence of HIV than male prisoners in six regions (both African regions, Asia Pacific, both European regions, and North America) and lower than male inmates in Latin America and the Middle East and north Africa (appendix p 9). Notable differences were found in west and central Africa, where the prevalence of HIV in women was almost double that of men (13·1% vs 7·1%), and in eastern Europe and central Asia, where it was almost three times higher than in men (22·1% vs 8·5%). HIV prevalence was higher in prisoners than in the general population in eastern Europe and central Asia (4·1% vs 0·5%) and western Europe (4·6% vs 0·2%).38
Panel 1: Compulsory drug detention centres
Some 27 countries detain drug users, or suspected drug users, in compulsory drug detention centres for the purpose of treatment or rehabilitation. In east and southeast Asia, an estimated 600 000 drug users are detained in roughly 1000 (mostly government-run) centres.7 In Latin America,8,9 and sub-Saharan Africa,10–12 an unknown number of individuals are detained in hundreds of faith-based and unregulated residential treatment centres. Abuses in these facilities include being shackled to trees and starved.13 Similar unregistered treatment centres operate in central Asia.14
Detainees are typically held in forms of administrative detention, often without due legal process,assessmentofdrugdependency,orinformedconsent.Evidence-baseddrugdependency treatment such as opioid agonist therapy is rarely provided,15 while physical abuse—including torture, forced prayer, forced exercise to sweat drugs out of the body, and manual labour (sometimes forced labour)—are common.16Estimatesofdiseaseburden in compulsory drug detention centres are given in the appendix (p 31).
In Vietnam, one formerly detained child described punishment in a compulsory drug detention centre, where the staff beat him on the arm and back with a truncheon. He reported being held in a small punishment room for 3 months. Another ex-detainee served his 2-year sentence only to have it extended by 5 years with no reason given.16
In Cambodia, one former detainee reported having his head bashed against a wall until he lost consciousness.16 One child, detained in a compulsory drug detention centre in Cambodia, said he saw a doctor after he was beaten. After the doctor treated the child, the doctor told the child to not try and escape again.17 In Thailand, individuals detained in compulsory drug detention centres reported being beaten or made to roll on gravel as a punishment.18
Ex-detaineeshavelimitedornoaccesstohealthcare.InChina,oneindividualdetainedinGuangxi province reported that he was unable to continue taking his antiretroviral drugs once placed in (compulsory) detoxification. Another person, detained in Yunnan province, reported that many incarcerated people have tuberculosis and many acquire tuberculosis while incarcerated, yet there is no treatment.19
These unregulated and abusive forms of detention in the name of drug treatment do not meet minimum health or human rights standards and should be closed.
See Online for apppendix
Series
www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 3
HIV prevalence in imprisoned people who inject drugsOf the 200 HIV datapoints, only 47 were related to people who inject drugs in 16 countries. Prevalence estimates from Iran ranged from 0·7% in 2002 to 18·2% in 2003 and fell to 2·3% in 2007. After two large outbreaks of HIV in prisons (panel 3), Iran increased opioid agonist therapy coverage from 100 to over 25 000 drug users, provided condoms and conjugal visits, and piloted six needle and syringe programmes in prisons from 2002.47,49 Australia’s HIV infection prevalence of almost zero in people who inject drugs in and out of prison can be traced back to very early introduction of community-based needle and syringe programmes in 1986, which prevented an estimated 25 000 HIV cases in people who inject drugs.58
Hepatitis C infectionPooled estimated HCV prevalence in prisoners46 of 196 countries had HCV prevalence data (171 data-points; hepatitis C antibody) in prisoners from 2005 to 201537 (appendix p 8). HCV infection in prisoners is high worldwide, exceeding 10% in six regions (figure 1, 3, appendix p 15). This finding reflects the increased infectivity and earlier HCV entry into populations of people who inject drugs compared with HIV.59
Epidemic patterns of HCV infection related to injection drug use persist in Europe. HCV prevalence estimates were high in the eastern Europe and central Asia region at 20·2% (95% CI 11·8–30·1), in west Europe at 15·5% (12·2–19·1), and in North America at 15·3% (13·1–17·7). Injection drug use is rare in the Caribbean, which probably accounts for the absence of HCV data (and possibly infection) in prison populations. In Latin America, HCV prevalence was 4·7% (3·1–6·7), also reflecting the low level of drug injection in this region. Insufficient data, particularly for female inmate populations, precluded gender comparisons of the prevalence of HCV.
HCV prevalence in imprisoned people who inject drugsOf the 171 datapoints on HCV infection, only 53 related to people who inject drugs in 19 countries. As expected, the prevalence of HCV in imprisoned people who inject drugs was high or very high, ranging from 8% to 95%, most of which were above 40% (appendix p 28). Only six countries had multiple datapoints from which trends could be surmised. With ten datapoints, Iran showed a clear reduction in prevalence from 78·3% (2001) to 43·4% (2009).49 No data were identified for HCV in people who inject drugs for either African regions or Latin America.
Hepatitis B infectionPooled estimated HBV prevalence in prisoners43 of 196 countries had HBV prevalence data (56 data-points) in prisoners from 2005 to 2015 (figure 3, appendix pp 8, 20).37 The prevalence of HBsAg in
prisoners in west and central Africa was very high at 23·5% (95% CI 19·8–27·5)—the highest burden of all four types of infection and across the nine regions. High levels of chronic HBV infection were also reported in east and southern Africa (5·7%, 95% CI 2·9–9·4) and in eastern Europe and central Asia (10·4%, 1·9–24·6). In other regions, HBV prevalence was below 5% and unsafe injection drug use and unprotected sex remain important modes of transmission.60
HBsAg prevalence in prisoners relative to the general populationThe burden of HBV in the general population is highest in sub-Saharan Africa, with a meta-analysis estimate of HBsAg prevalence at 8·8%, which was primarily attributed to vertical transmission from mother to child,
Panel 2: Mass incarceration and the HIV epidemic in the USA
The US incarceration rate of 716 per 100 000 population is almost five times the global average of 146 per 100 000.22 This translates into 2·2 million people, or almost 1% of the US population behind bars on any given day in 2013. In the 1970s, this rate was a meagre 75 per 100 000. However, the Rockefeller Drug Laws, introduced in 1973, caused the incarceration rate to increase by five times in a decade.23 Specifically designed to target heroin and crack cocaine users, these laws mandated lengthy minimum sentences for many drug offences, including the possession and sale of small quantities of drugs, which matched sentences for rape, assault, and robbery. These laws took account of previous offences and culminated in the three-strikes law (ie, three strikes and you are in for life), for minor, but multiple offences.
By 1985, drug offenders comprised one-third of all inmates in New York state. The burden of incarceration fell heavily on young black men, who were 40 times more likely to be incarcerated than their white peers, and Hispanic men, who were 30 times more likely to beimprisonedthantheirwhitepeers.NewYorkCity’sRikersIslandprisononcehadthelargest concentration of HIV-positive people in the USA.
HIV prevalence is roughly three times higher in incarcerated individuals than in the general US population.24 HIV prevalence in the state prisons of Florida, Maryland, and New York exceeds 3%, which is higher than the national prevalence of any country outside of sub-Saharan Africa.25 At least one in six HIV-positive American people passes through a correctional centre each year.23 HIV infection has been linked to a history of incarceration; however, proving the infection occurred in prison is difficult. Although cases of intraprison HIV and HBV transmission have been recorded (panel 3), evidence suggests that most HIV-positive people were infected before prison entry. For people in detention with HIV infection, treatment outcomes have been good, but retention in treatment after release is more challenging.26
The post-release period can also be especially dangerous for treatment interruptions, fatal andnon-fataloverdose,andinfectionacquisition.TheALIVEstudy27 reported that the risk of HIV transmission increases, rather than decreases, on release from prison. Incarceration plays a role in delaying recovery from drug dependence. Incarceration also has an impact on those in the community, most probably through the return of prisoners who have been released and are infectious, who have interruptions in antiretroviral treatment. The main risk behaviour for newly diagnosed, heterosexually acquired HIV infection in African-American women was having sex with a partner who had a history of incarceration.28 The approach to drug policy and the mass incarceration of drug users in the USA is counterproductive, expensive, and increases the risk of HIV treatment interruption and subsequent transmission.23
Series
4 www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
low HBV vaccine coverage (including birth dose), and unsafe medical products and procedures.61 In Australia, the prevalence of HBsAg was 2·3% in prison and 1·0% in the community.62
HBsAg prevalence in imprisoned people who inject drugsOf the 56 HBV datapoints, 24 related to people who inject drugs in ten countries. Iran and Australia were the only countries with multiple datapoints (appendix p 28). In Iran, HBV prevalence was significantly and positively associated with the frequency and duration of imprisonment.49 Despite a reduction in the proportion of Australian prisoners reporting injection drug use, evidence of HBV infection remains high in
prison entrants, with many unaware of their infection status.62
Tuberculosis and co-infection with HIVPrevalence of active tuberculosis and co-infection with HIV in prisonersEstimates of active tuberculosis in prison populations were higher than in the general population in almost all settings, with investigators of one study, from the Dourados prison in Brazil, estimating that prevalence of tuberculosis was 40 times higher than in the general population.63
25 of 196 countries reported tuberculosis prevalence data (46 datapoints) and 17 of 196 countries prevalence data on co-infection with HIV and tuberculosis (25 data-points) in prisoners from 2005 to 2015 (figure 4, appendix pp 8, 27).37 In the three regions with the highest prevalence of HIV and viral hepatitis, active tuberculosis was also common (appendix pp 24, 27): 5·3% (95% CI 2·1–10·0) in east and southern Africa, 2·9% (2·4–3·6) in west and central Africa, and 4·9% (1·8–9·3) in eastern Europe and central Asia. In the African regions, active tuberculosis infection was associated with overcrowding, high inmate turnover, and increased frequency or duration of incarceration.64–66 Disease burden was lower in other regions, though only one study was identified in the Middle East and north Africa, two in the western European region, and none in the Caribbean.
A systematic review67 found that the incidence of tuberculosis is 23 times higher in prison populations than in the general community, and that the prevalence of drug-resistant tuberculosis is also substantially higher. Increased prevalence of HIV infection exacerbates the effect of incarceration on tuberculosis incidence, highlighting the importance of a coordinated approach to prevent and treat these two infections.34 Important for tuberculosis control in prisons is case-finding to ensure early detection and treatment, case isolation and infection control measures,
0–1·01·1–5·05·1–10·010·1–15·015·1–20·0>20·0No data meeting inclusion criteria
North America1·3% (95% CI 1·0–1·7)
Latin America2·3% (95% CI 1·5–3·4)
Caribbean3·3% (95% CI 2·7–4·0)
East and southern Africa15·6% (95% CI 11·8–19·8)
Eastern Europe and central Asia4·1% (95% CI 1·4–8·0)
Middle East and north Africa1·3% (95% CI 0·2–3·3)
West andcentral Africa8·2% (95% CI 6·2–10·5)
Asia Pacific1·4% (95% CI 0·9–1·9)
Western Europe4·2% (95% CI 2·7–6·1)
Figure 2: Global and regional burden of HIV in prison inmates, published between 2005 and 2015
East andsouthern
Africa
West andcentralAfrica
Middle Eastand north
Africa
AsiaPacific
EasternEurope andcentral Asia
WesternEurope
NorthAmerica
Caribbean LatinAmerica
0
5
10
15
20
25
35
30
40
Prev
alen
ce (%
)
HIVHCVHBsAgActive tuberculosis
Figure 1: Regional prevalence of HIV, hepatitis C, HBsAg, and active tuberculosis in prisoners, published between 2005 and 2015HCV=hepatitis C antibodies.
Series
www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 5
initiation of tuberculosis treatment followed soon after by HIV treatment, and treatment of latent tuberculosis infection, especially in patients with HIV.
Studies investigating HIV and tuberculosis co-infection are of two types (appendix p 27). The first study type measures HIV and tuberculosis co-infection in all prisoners, which typically finds prevalence estimates that are less than 10% of the prison population, ranging from 1·6% (Zambia) and 5·1% (South Africa) in east and southern Africa to 0·3% (Guinea) and 8·9% (Nigeria) in west and central Africa. The figures were less than 1% in studies from Latin America and western Europe. The second type of study measures HIV prevalence in inmates with tuberculosis infection and vice versa. Studies of HIV in individuals with tuberculosis done in Africa typically report a prevalence of 40% and higher (Burkina Faso 75%, Uganda 57%, Zambia 45·8%, and South Africa 42–58%). Prevalence estimates of tuberculosis in people with HIV infection tend to be lower (Burkina Faso 20%, Ethiopia 47%, and Malaysia 17%; appendix p 27).
Incidence of HIV, HCV, HBV, and tuberculosis in prisonersFew incidence studies were found on HIV (three countries), HCV (four), HBV (two), and tuberculosis (14; appendix p 32). Spain had detailed annual HIV incidence data from 2000 to 2014 that revealed a steady decline from 0·70% to 0·04% per year over 14 years.68 A large prospective study of hepatitis C incidence in prisons in Australia reported an annual incidence of 14·1%,69 which was high compared with Scotland (0·9%),70 the USA (0·4%),71 and Spain (1·7%).72
Of the 14 countries with tuberculosis data, only seven had multiple incidence rates. Countries that had a decrease in incidence were Colombia,73,74 Argentina,75 Hong Kong,76 Macedonia,77 and Romania.78 An increase in tuberculosis cases in prison was observed in the USA (figure 5).79
The potential effect of increased incarceration on tuberculosis in Europe and central Asia was modelled.34 For each percentage point increase in incarceration rates, there was a corresponding increase in tuberculosis incidence of 0·34% (population attributable risk; 95% CI 0·10–0·58, p<0·01). Mathematical modelling of tuberculosis transmission in Ukraine further suggests that incarceration contributes to 75% of new tuberculosis infections in people who inject drugs.80
Infectious disease prevalence in key populations HIV, HCV, HBV, and tuberculosis in imprisoned MSM, female sex workers, and transgender peopleEstimates of infectious diseases in key populations who become imprisoned (eg, MSM, female sex workers, and transgender people) are less frequently reported than in people who inject drugs. In 2008, HIV prevalence in Ghana in imprisoned female sex
workers (n=73) was 11% and 8·9% in MSM (n=403).81 HIV prevalence in imprisoned MSM was 43·8% in Nigeria (n=32 in 2009), 2·6% in Iran (n=113), and 5·5–34·0% in the USA.82–90 The only study of transgender prisoners was done in Argentina, in which half of transgender prisoners (six of 11 people) were HIV positive.75 US studies provided estimates of HIV prevalence in female sex workers from 2·0% to 5·6%.89,91
Imprisoned MSM in the USA and Iran had high levels of exposure to HCV (50%)82,92 compared with imprisoned MSM in Nigeria (12·5%).83 Imprisoned MSM had high levels of HBV in Ghana (40·2%)81 and Nigeria (56·3%).83 Female sex workers in Ghana also had high levels of HBV infection at 37·0%.81
Panel 3: Outbreaks of infectious diseases in prisoners
One of the first recognised HIV outbreaks occurred in a Bangkok prison in 1988.42 The outbreak was only detected once infected inmates were released and HIV in people who inject drugs escalated from 2% to 43% over the course of several months. An investigation revealed that HIV incidence in prison was very high at 35 per 100 person-years.43,44 In Lithuania, a policy of segregation saw a prisoner who was seroconverting to HIV infection housed with individuals who were HIV negative. At least 284 prisoners were identified as infected with HIV in a few months, which doubled the number of diagnosed HIV cases in Lithuania.3,45 Two outbreaks in Russian prisons saw more than 400 (of 1824) prisoners infected in Nizhnekamsk and 260 inmates infected in the Tatarstan region in 2001.3,46 Large outbreaks have occurred in prisons in Iran and Ukraine, with hundreds of prisoners infected.47,48 Iran increased opioid agonist therapy places from 100 to over 25 000, provided condoms and conjugal visits, and piloted six needle and syringe programmes in prison from 2002.47,49 Small HIV outbreaks have even occurred in areas where prevalence was very low, such as in Australia (four prisoners) and Scotland (ten prisoners), which also had intraprison hepatitis B virus (HBV) transmission.50–52 In 2000, two Australian prisoners were caught sharing syringes and an investigation found that four inmates had acquired hepatitis C virus infection, but none had acquired HIV in prison.53
IntheUSA,Georgia’sstateprisonsystemrecordedanoutbreakofsevencasesofhepatitisBin 2000 and 2001, which were attributed to sex. 52% of 907 susceptible inmates who completed a questionnaire reported at least one risk behaviour for hepatitis B virus infection (including 48% who reported tattooing, sharing a razor [8%], having sex [4%], or injecting drugs [2%]). HBV vaccination was offered only to inmates in the dormitory where the outbreak originated, leaving over 90% of inmates potentially at risk. In the second outbreak, an audit of prison medical records revealed 41 (72%) of 57 HBV cases had occurred in prison. Routine HBV vaccination was recommended for all new prisoners.54–56
In 2006, two ex-inmates in the USA were diagnosed with tuberculosis sparking an investigation.57 Despite one inmate having prolonged symptoms and abnormal chest radiographs, he went undiagnosed while incarcerated. Of 910 exposed inmates who were tested, 53 inmates (5·8%) had newly positive tuberculin skin tests. This included 11 (of 204) new cases of tuberculosis who had been released and re-incarcerated. Of 485 prison employees tested, ten (2·1%) were identified as tuberculin skin test converters.
Although the study of the transmission of these infections prospectively in prison populations has been challenging, these outbreaks illustrate the role that prisons might play in continuing these epidemics beyond the prison wall. It is good public health policy to screen at-risk new prisoners for infectious diseases and to provide inmates with free, easy, and confidential access to prevention programmes such as opioid agonist therapy, antiretroviral therapy, condoms, and HBV vaccination to reduce the risk of outbreaks.6
Series
6 www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
Prevalence of HIV, HCV, and HBV in compulsory drug detention centre detainees in China, Taiwan, Vietnam, and IranIn 2012, an estimated 235 000–600 0007,20 drug users were detained for compulsory drug treatment in over 1000 compulsory drug detention centres.20 While the number of compulsory drug detention centres remained the same, the numbers of drug users detained in those centres had increased by 2015. In China, HIV prevalence ranged from 0·2% to 9·5% in general detainees and was 10·3% in people who inject drugs, whereas HCV prevalence was 44% in general detainees (appendix p 31). In Taiwan, HIV prevalence was 6·9% in general detainees and 25·5% in people who inject drugs in detention. Viral hepatitis was also common with HCV prevalence estimated at 30·5% in general detainees and 89·6% in people who inject drugs; HBsAg prevalence was estimated at 16·9% in general detainees and 16·7% in people who inject drugs in detention. In Vietnam, HIV prevalence was 19·8% and
HCV prevalence was 76·9% in people who inject drugs in detention. Iran reported that 24·4% of detained people who inject drugs were HIV positive, 80% were anti-HCV antibody positive, and 5·8% were HBsAg positive (appendix p 31).
Mathematical modelling of HIV transmission and prevention in people who inject drugsWe modelled the contribution of incarceration and re-incarceration on HIV incidence in people who inject drugs and examined the effects of a reduced rate of incarceration, introduction of prison-based opioid agonist therapy followed by post-release opioid agonist therapy, and post-release antiretroviral treatment retention on HIV incidence.
In our model, the increased risk for HIV transmission was captured by behaviour change in people who inject drugs who temporarily receptively share syringes in prison and a combination of interruption of antiretroviral treatment and behaviour change in people who inject
A
B
North America15·3% (95% CI 13·1–17·7)
Latin America4·7% (95% CI 3·1–6·7)
Caribbean
East and southern Africa1·8% (95% CI 0·2–9·5)
Eastern Europe and central Asia20·2% (95% CI 11·8–30·1)
Middle East and north Africa11·9% (95% CI 5·8–19·8)
West and central Africa16·9% (95% CI 13·1–21·1)
Asia Pacific20·6% (95% CI 15·4–26·4)
Western Europe15·5%(95% CI 12·2–19·1)
North America1·4% (95% CI 0·3–3·1)
Latin America2·3% (95% CI 0·1–8·3)
Caribbean
East and southern Africa5·7% (95% CI 2·9–9·4)
Eastern Europe and central Asia10·4% (95% CI 1·9–24·6)
Middle Eastand north Africa3·3% (95% CI 2·4–4·5)
West and central Africa23·5% (95% CI 19·8–27·5)
Asia Pacific4·4% (95% CI 1·4–9·0)
Western Europe2·4% (95% CI 1·6–3·3)
0–5·05·1–10·010·1–15·015·1–20·020·1–30·0>30·0No data meeting inclusion criteria
0–1·01·1–5·05·1–10·010·1–15·015·1–20·0>20·0No data meeting inclusion criteria
Figure 3: Global and regional prevalence of viral hepatitis in prison inmates, published between 2005 and 2015(A) Prevalence of HCV antibodies. (B) Prevalence of HBsAg.
Series
www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 7
drugs who temporarily receptively share syringes during transient post-release period, which was assumed to last between 1 and 6 months (panel 4, appendix p 33). We quantified the contribution of discontinuation of antiretroviral treatment and temporary syringe sharing to cumulative HIV incidence in people who inject drugs for varying proportions of people who continue this high-risk behaviour for a transient period after release (figure 6). Antiretroviral treatment discontinuation was shown to contribute to less than 5% of cumulative HIV incidence during the epidemic, from epidemic onset to 2015 (figure 7A). This small contribution of discontinuation of antiretroviral treatment to HIV
incidence was due to the small proportion of HIV cases discontinuing antiretroviral treatment at a given time and the short duration of discontinuation.
Specifically, our results show that for people who inject drugs from communities with high HIV prevalence (over 20%), the relative contribution of antiretroviral treatment discontinuation to HIV transmission varies from 0·12% (95% credible interval [CrI] 0·02–0·62) to 0·15% (0·02–1·0), which depends on whether individuals who engage in temporary syringe sharing behaviour only do so during incarceration or continue to do so after release (figure 7A). In communities with moderate HIV prevalence (5–20%), the relative contribution of antiretroviral
North America
Latin America1977(95% CI 1347–2726)
Caribbean
WesternEurope
10–10001001–20002001–50005001–80008001–10 000>10 000No data meeting inclusion criteria
East and southern Africa5330 (95% CI 2066–9998)
Middle East and north AfricaWest and
central Africa2946(95% CI 2362–3591)
Asia Pacific1173 (95% CI 712–1747)
Eastern Europe and central Asia4903 (95% CI 1846–9321)
Figure 4: Global and regional prevalence of active tuberculosis in prison inmates, published between 2005 and 2015Data are per 100 000.
Figure 5: Incidence of tuberculosis in prisons and closed settings*Cumulativeoverthisperiod.†Averageoverthisperiod.‡Estimatedannualrisk1996–2001.
Series
8 www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
treatment discontinuation varies from 0·10% (0·01–1·1) to 0·23% (0·03–3·3), which also depends on whether individuals who engage in temporary sharing behaviour only do so while incarcerated (figure 7A). The contribution of temporary syringe sharing to HIV transmission varies from 7·0% (0·1–41·6) to 12·0% (0·5–51·8) in communities with high HIV prevalence, and from 5·8% (0·2–31·7) to
21·1% (0·4–53·0) in communities with moderate prevalence (figure 7B). The contribution of temporary syringe sharing to HIV transmission was higher in communities with lower HIV prevalence than in communities with high HIV prevalence (figure 7B).
To identify the contribution of each model parameter to the variability of temporary syringe sharing to HIV transmission, we calculated the partial rank correlation coefficients (PRCCs), which quantify the magnitude of increase or decrease of the value of an outcome measure predicted by the model as a given input model parameter varies.99
PRCCs indicated that the effect of temporary syringe sharing increases with the rate of re-incarceration for people who inject drugs, the proportion of people who inject drugs who share syringes in prison, and the risk of HIV transmission per episode of syringe sharing (appendix p 41). Conversely, the effect of temporary syringe sharing decreases with the proportion of people who inject drugs who share syringes outside prison, the mortality rate of people who inject drugs who do not have HIV, and the proportion of people who inject drugs and have never been incarcerated (appendix p 41). Given that the rate of first incarceration in people who inject drugs was defined as a non-monotonic functional relationship between the average non-HIV mortality rate for people who inject drugs and the proportion of people who inject drugs who have never been incarcerated (appendix p 33), we deduced from the PRCC results that the rate of first incarceration has a substantial contribution to the effect of temporary syringe sharing. However, that effect could not be captured by the PRCC analysis because of the non-monotonic relationship.
We assessed the effectiveness of three interventions in reducing HIV incidence after 5 years of initial implementation: the incarceration rate of people who inject drugs, prison-based opioid agonist therapy followed by post-release retention, and prison-based opioid agonist therapy followed by post-release opioid agonist therapy combined with antiretroviral treatment retention. We showed that lowering incarceration rates (both first-time incarceration and re-incarceration) by 10–50%, corresponding to a 4·6% (95% CrI 2·6–6·6) to 25·9% (16·9–35·4) reduction of incarcerated people who inject drugs over 5 years, could potentially reduce community-wide cumulative HIV incidence in people who inject drugs by 0·9% (–0·1 to 3·6) to 7·6% (0–26·5) over a 5-year period in communities with high HIV prevalence, and by 1·1% (0·1–3·1) to 15·4% (1·4–29·9) in communities with moderate HIV prevalence (appendix p 42). We found that prison-based opioid agonist therapy followed by post-release retention in treatment during the transitory period of elevated risk behaviour could reduce cumulative HIV incidence by 1·6% (95% CrI 0·1–7·3) for a 20% coverage and by 12·0% (1·4–45·3) as an upper bound of complete (100%) coverage in communities with high HIV prevalence. In
Panel 4: Mathematical model for HIV transmission in people who inject drugs
To assess the fundamental dynamics that govern the interplay between HIV transmission in people who inject drugs in the community and in prison around the world, we developed a deterministic compartmental model of drug injection transmission of HIV in the general community and incarcerated populations (figure 6). Individuals were structured in the model according to their HIV infection history, disease progression, and antiretroviral treatment status (1=susceptible, 2=acute infection, 3=latent infection with CD4 cell count ≥350 cells per µL, 4=latent infection with CD4 cell count <350 cells per µL, 5=pre-AIDS with elevated viraemia, 6=AIDS, 7=initiated antiretroviral treatment with CD4 cell count ≥350 cells per µL, 8=initiated antiretroviral treatment with low CD4 cell count or pre-AIDS, 9=initiated antiretroviral treatment with AIDS), incarceration status (1=currently incarcerated, 2=recently released [in the past 6 months], 3=previously incarcerated, 4=never incarcerated), and drug injection risk behaviour (1=people who do not inject drugs, 2=people who inject drugs who do not share syringes, 3=people who inject drugs who share syringes, 4=people who inject drugs and share syringes temporarily). As our analysis focused on HIV epidemics in people who inject drugs, we only considered injection drug use as transmission mode. Rather than focusing on a specific setting, we considered a wide range of scenarios by stratifying the rates of HIV transmission in people who inject drugs into communities with moderate HIV prevalence in people who inject drugs (ranging from 5% to 20%), and communities with high HIV prevalence in people who inject drugs (over 20%).93 Our model accounted for the elevated frequency of needle sharing in prison as well as discontinuation of antiretroviral treatment during a transitory period after prison release.31,94–97 We parameterise our model using epidemiological and clinical data from the scientific literature about HIV infectiousness, disease progression, demographic, injection drug behaviour data on the incarcerated population, proportion of people who inject drugs, syringe sharing in and out of prison, and incarceration history in people who inject drugs and people who do not share syringes (appendix p 33). To address empirical uncertainty regarding model parameters, we used a Bayesian inference approach to fit our model to epidemiological data of HIV prevalence in people who inject drugs in and out of prisons and HIV incidence in prison.
Given the scarcity of data on the impact of this post-release behaviour change on HIV transmission, we did a sensitivity analysis to evaluate the contribution of the frequency of post-release needle sharing and percentage discontinuation of antiretroviral treatment on HIV incidence in people who inject drugs in the community and in prison. We denoted people who inject drugs who share syringes while incarcerated as temporary sharers, a proportion of whom share for a transient duration after release, but not beyond. This post-release transient period of temporary syringe sharing and antiretroviral treatment discontinuation was assumed to vary between 1 and 6 months.31,95
Finally, we used the fitted model to evaluate the impact of demand reduction programmes, such as prison-based opioid agonist therapy followed by post-release retention, reduced rates of incarceration for non-violent drug users, and antiretroviral treatment retention in care post-release for curtailing HIV incidence in people who inject drugs. Consistent with empirical studies, we assumed that people who inject drugs who continue opioid agonist therapy after incarceration have a 20% (95% CI 10–30) reduced risk of re-incarceration while they are given treatment.98
Series
www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 9
communities with moderate HIV prevalence, the reduction in HIV incidence is predicted to vary from 3·9% (0·9–13·4) for 20% coverage to 28·5% (9·4–57·5) for complete coverage (appendix p 42). Combining opioid agonist therapy with post-release antiretroviral treatment retention could achieve a further reduction of HIV incidence ranging from 0·3% to 1·8% in communities with high HIV prevalence and from 0·7% to 4·2% in communities with moderate HIV prevalence as coverage is expanded from 20% to 100% (appendix p 42).
Our estimates of the contribution of incarceration to HIV transmission and the potential effect of interventions depends on the general HIV prevalence in communities of people who inject drugs, which varies substantially between and within countries (figure 7, appendix p 42). Our results show that prison-based opioid agonist therapy followed with post-release retention could be an effective strategy for the reduction of HIV in people who inject drugs, at least in the short term (appendix p 42). Lowering the rate of incarceration of people who inject drugs might be effective in reducing HIV transmission in addition to opioid agonist therapy. The contribution of incarceration and effectiveness of intervention were higher in communities of lower HIV prevalence, where transmission because of incarceration is most substantial relative to that in the community, compared with communities with high HIV prevalence.
Summary of findingsThis Series paper provides clear evidence that the findings, replicated in many individual studies, that higher prevalence of HIV in individuals who are detained than in the population from which they arise, is almost universal. However, it also shows that the detailed epidemiology of HIV in people who are detained varies
A HIV infection diagram B Incarceration dynamics diagram C Subpopulation and incarceration impact diagram
Susceptible
Acute infection
Pre-AIDS
AIDS
Latent infection CD4 cell count ≥350
Latent infection CD4 cell count <350
Never incarcerated
Currently incarcerated
Recently released
Previously released
ART treatedPre-AIDS
ART treatedAIDS
ART treated
Non-PWID PWID non-sharers
PWID sharers
Community
Prison
Non-PWID PWID non-sharers
PWID sharers
Prison-induced riskProportion of non-sharers temporarily sharing in prison and in a 1–6 month transient post-release period
Incarceration rate for non-PWIDElevated incarceration rate for PWIDTemporal transition between sharers and non-sharers due to incarceration
Figure 6: Deterministic compartmental model of drug injection transmission of HIV in the general community and incarcerated populationsART=antiretroviral therapy. PWID=people who inject drugs.
0
1
2
3
4
Cont
ribut
ion
to H
IV in
ciden
ce (%
)
0 0·25 0·50 0·75 1·000
10
20
40
60
A
B
50
30
Cont
ribut
ion
to H
IV in
ciden
ce (%
)
Proportion of temporary sharers sharing syringe post-release
Moderate HIV prevalenceHigh HIV prevalence
Figure 7: Contribution of antiretroviral therapy discontinuation and temporary sharers to cumulative HIV incidence in people who inject drugs(A) Contribution of ART discontinuation. (B) Contribution of temporary sharers. Contribution accounts for transmission in prison and during the transitory period of elevated risk behaviour after release from prison. Contributions were evaluated relative to the scenario of no antiretroviral therapy discontinuation or no temporary sharers, and were assessed for a varying proportion of temporary sharers who share syringes after incarceration versus those who only share syringes while incarcerated, from epidemic onset to 2015. Bars show median (middle cross line) and range of contribution estimated for different HIV epidemics modelled (error bars show upper and lower 95% credible interval). Prevalence of HIV in people who inject drugs was varied from 5–20% (moderate prevalence communities) and 20–60% (high prevalence communities).
Series
10 www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
considerably, reflecting the disease burden, the dominant mode of transmission in the population as a whole, and the role of incarceration. In most regions of the world, HIV prevalence is higher in detained women than detained men. However, although we have shown great variation in the pattern of HIV infection in prisons, the ability to draw general conclusions is limited by the scarcity of data, with data in some regions restricted to one or a few countries. The studies that do exist are often restricted to individual prisons (even subgroups of detainees in them) and are heterogeneous in reporting quality.
There are several reasons for the scarcity of data. First, many countries with the highest disease burden have very weak systems for health research and surveillance. Second, where health systems in prisons are organisationally separate from systems serving the general population, capacity that exists in prison systems might be deployed elsewhere. Consistent with the UN’s Mandela Rules, the WHO recommends that prisoner health care be consistent with community standards of care, and under the direction of the ministry of health.100 Third, secrecy and fear of exposing human rights abuses might cause governments to be unwilling to permit studies in their prisons, or if they are done, researchers are often prohibited from published results.101,102 Compulsory drug detention centres are a particular concern. In 2012, 12 UN agencies called for their immediate closure15 and since then, international funding for compulsory drug detention centres has stopped. However, they remain in operation in several regions, although largely hidden from those undertaking HIV surveillance.7
A particular problem is the scarcity of data on women detainees, even though they comprise about 10% of prison populations, they are more likely to have a substance abuse disorder, and are less likely to receive treatment than male prisoners.103
Most studies are cross-sectional, providing only snapshots of the situation at a particular moment in time. Cross-sectional studies, however, fail to capture the dynamic nature of the prison population, because the number of individuals passing through prisons each year can be up to three times higher than the estimated prison census. This discrepancy causes an especially important gap in our knowledge, given the role that the transition in and out of prison has in interruption of both antiretroviral treatment and opioid agonist therapy, with implications for both transmission and the emergence of drug-resistant HIV and tuberculosis infections.104
So what should be done to address the problem of HIV and related infections in prisons? One obvious response is to reduce the prison population, as our model showed. Measures that can reduce the population in detention include evidence-based treatment in and out of prison. Non-violent drug offenders and especially women15 should particularly be offered treatment, as an alternative, where appropriate.105 Our model further showed that
opioid agonist therapy given in prison can reduce HIV transmission within prison and reduce post-release HIV transmission.
Other infectious diseases might also be mitigated in prison settings. Interferon-free HCV therapies using direct-acting antiviral agents are rapidly becoming available and HCV treatment as prevention strategies, particularly in the prison setting, are now feasible.106 These treatments can produce 90–95% rates of HCV eradication.107
Routine screening and vaccination for HBV in prison settings108 is likely to produce considerable savings to the broader health system through prevention of cirrhosis of the liver and hepatocellular carcinoma,109 even though only around 5% of adults infected with HBV develop chronic infection110 because of the comparatively high prevalence in prison settings. Such programmes should be linked to a wider strategy to reduce HBV infection in countries where it is prevalent, through early childhood immunisation. A safe and efficacious vaccine for HBV has existed since 1986.111
A review of the evidence base for the management of tuberculosis and HIV and tuberculosis co-infection in prisons provides a useful framework including universal drug susceptibility testing, systematic screening of contacts in high-risk groups, access to high-quality treatment, and adequate coordination.112 Other specific measures include identification, isolation, and treatment of patients who are infected with tuberculosis as early as possible. In many countries, prisoners with symptoms suggestive of tuberculosis experience long delays in obtaining a diagnosis.113 Isolation of infectious prisoners is rare; many prisoners who are isolated might not receive treatment, and conditions are frequently appalling.114
Tuberculosis treatment completion rates in prisoners are often low, exacerbated by their movement within and in and out of the prison system. Prisoners in eastern Europe are unlikely to receive adequate treatment for HIV and tuberculosis.80 A study in Uganda found a default treatment prevalence of 12% in people staying in the same prison and 53% for those transferred to another prison, and 81% of prisoners who were on treatment and released were subsequently lost to follow-up.115
Pooled estimates of HIV prevalence and incidence, collected for the purpose of this review, were used to develop and calibrate a model for people who inject drugs for HIV transmission in and out of prison. Given that our focus was the evaluation of the contribution of incarceration to injection-related HIV transmission in people who inject drugs and the potential effect of interventions on the HIV epidemic in people who inject drugs, we did not address sexual, tattoo-related, or mother-to-child transmission of HIV. Inclusion of these transmission routes would further increase the predicted effect of the interventions considered by accounting for the substantial indirect benefits to people who do not
Series
www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 11
inject drugs. Another limitation of our modelling analysis is the fact that the model was not developed for a specific setting; rather we investigated a range of plausible scenarios. This approach resulted in large uncertainty about model parameter values and model outcomes. Additionally, our model assumed constant antiretroviral treatment coverage during the course of the epidemic, which does not capture the effect of increased antiretroviral treatment coverage on HIV transmission. However, the effect of this assumption was mitigated by using a wide range of values for antiretroviral treatment coverage. Finally, our model was restricted to inmates who inject heroin, because most research and prevention has focused on these individuals.
ConclusionWe found a consistently higher prevalence of HIV, HCV, HBV, and tuberculosis in prisoners than in the general population across all regions and especially in imprisoned people who inject drugs.38,61,116 Interventions to prevent, identify, and treat these infections in prisons are poorly implemented, particularly in low-income and middle-income countries and in populations such as people who inject drugs, where care and treatment remain challenging in community settings. Investment in surveillance infrastructure is needed to improve country-level data on the prevalence of these infections and to inform policy and programmatic responses. This is particularly important in regions where injection drug use is increasing and the burden of HIV is already high. Mass incarceration of people who inject drugs is a key driver of the ever-growing population of prisoners. Decriminalising drug use and possession or providing alternatives to imprisonment for people who use drugs, and ensuring access to opioid agonist therapy, tuberculosis treatment, and antiretroviral treatment for prisoners will reduce the burden of these infections in the world’s prison population.ContributorsKD and CB developed the manuscript concept. ALW led the comprehensive search and data extraction and did the meta-analysis. KD developed the initial draft of the manuscript. BM wrote the sections on imprisonment data. BM and ALW calculated the number of inmates with each infection. JJA wrote the Compulsory Drug Detention Centres section. LM provided a revised structure for the manuscript, edited it, and addressed reviewers’ comments. MH analysed and wrote the section on incidence data. ALW, AG, MN-m, KD, and FLA conceived of the mathematical modelling analysis, MN-m conducted the mathematical modelling, and MN-m and AG wrote the modelling section. MM wrote the section on tuberculosis data and provided the final edit. SAK wrote the section on hepatitis B data. RC wrote the section on hepatitis C data. All authors had full access to the data, reviewed and edited the manuscript, and take responsibility for its integrity and the accuracy of the analysis.
Declaration of interestsWe declare no competing interests.
AcknowledgmentsThis paper was supported by grants to the Center for Public Health and Human Rights at Johns Hopkins Bloomberg School of Public Health from the National Institute on Drug Abuse; the Open Society Foundations; the United Nations Population Fund; Mac AIDS Fund;
the Bill & Melinda Gates Foundation; and the Johns Hopkins University Center for AIDS Research, a National Institute of Health (NIH)-funded programme 1P30AI094189. KD was supported by Australia’s National Drug and Alcohol Research Centre. RC is supported by a Cancer Institute New South Wales Early Career Research Fellowship (GNT14/ECF/1–46). SAK is supported by Australian National Health and Medical Research Council (NHMRC) Senior Research Fellowship (APP1078168). LM is supported by an NHMRC Senior Research Fellowship (APP1060443). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. We thank Lori Rosman, Informationist at Johns Hopkins School Medicine, who assisted with the development of the systematic review; Teresa Ping Yeh and Natalie Flath, Johns Hopkins School of Public Health, who reviewed the identified articles and extracted data for the systematic review and meta-analysis; and Mary Kumvaj, Librarian, National Drug and Alcohol Research Centre, University of New South Wales who assisted with the grey literature search. We thank colleagues at the UN Office on Drugs and Crime, the European Monitoring Centre for Drugs and Drug Addiction, WHO, Glasgow Caledonian University, Health Protection Surveillance Centre (Ireland), Luxembourg Division de l’Inspection Sanitaire, National Advisory Committee on Drugs and Alcohol (Ireland), The National AIDS Centre (Poland), Public Health Agency of Catalonia, Trimbos Instituut, The University of Texas at Austin, Brandon Marshall, Brown University School of Public Health, Anna Bowring, Burnet Institute, and The UK Collaborating Centre for WHO Health in Prisons who contributed relevant reports and articles to the comprehensive review.
Cunningham-Rundles S. Acquired immunodeficiency syndrome in male prisoners. New insights into an emerging syndrome. Ann Intern Med 1983; 98: 297–303.
2 Braun MM, Truman BI, Maguire B, et al. Increasing incidence of tuberculosis in a prison inmate population. Association with HIV infection. JAMA 1989; 261: 393–97.
3 Dolan K, Kite B, Black E, et al. HIV in prison in low-income and middle-income countries. Lancet Infect Dis 2007; 7: 32–41.
4 Gough E, Kempf MC, Graham L, et al. HIV and hepatitis B and C incidence rates in US correctional populations and high risk groups: a systematic review and meta-analysis. BMC Public Health 2010; 10: 777.
5 European Monitoring Centre for Drugs and Drug Addiction. Prisons and drugs in Europe: the problem and responses. Luxembourg: Publications Office of the European Union, 2012. http://www.emcdda.europa.eu/attachements.cfm/att_191812_EN_TDSI12002ENC.pdf (accessed Sept 1, 2015).
6 Kamarulzaman A, Reid S, Schwitters A, et al. Prevention of transmission of HIV, hepatitis B virus, hepatitis C virus, and tuberculosis in prisoners. Lancet 2016; published online July 14. http://dx.doi.org/10.1016/S0140-6736(16)30769-3.
7 Dolan K, Worth H, Wilson D. Compulsory treatment of drug users in Asia: designed to torture? Int J Prison Health 2015; 11: 255–68.
8 Colectivo por una Política Integral Hacia las Drogas. ¡Ay Padrino, no me ayudes! Abusos en centros de tratamiento con internamiento para usuarios de drogas in México. Mexico City, 2015. http://www.cupihd.org/portal/publicaciones_documentos/abuso-en-vez-de-asistencia-realidad-centros-tratamiento (accessed Sept 15, 2015).
9 Wolfe D. Death, drug treatment, and Christ’s love. Feb 14, 2012. https://www.opensocietyfoundations.org/voices/death-drug-treatment-and-christ-s-love (accessed Sept 15, 2015).
10 Open Society Foundation. Human rights abuses in the name of drug treatment: reports from the field. March, 2009. https://www.opensocietyfoundations.org/publications/human-rights-abuses-name-drug-treatment-reports-field (accessed Sept 15, 2015).
11 YouthRise, OSIWA, CiSHRWIN. We are people: the unintended consequences of the Nigerian drug law on the health and human rights of young people who use drugs. January, 2015. https://dl.dropboxusercontent.com/u/16336789/We-Are-People-Final.pdf (accessed Sept 15, 2015).
12 Messersmith L, Beard J, Adjei R, et al. HIV vulnerability of Men and women who inject drugs in Kumasi, Ghana. July 23, 2014. http://www.bu.edu/cghd/files/2014/10/PWID_Report_Final.pdf (accessed Sept 15, 2015).
Series
12 www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
13 Human Rights Watch. “Like a death sentence”: abuses against persons with mental disabilities in Ghana. New York, NY: Human Rights Watch, 2012.
14 Vagenas P, Azbel L, Polonsky M, et al. A review of medical and substance use co-morbidities in Central Asian prisons: implications for HIV prevention and treatment. Drug Alcohol Depend 2013; 132 (suppl 1): S25–31.
15 United Nations. Joint statement on compulsory drug detention and rehabilitation centres. March, 2012. https://www.unodc.org/documents/southeastasiaandpacific/2012/03/drug-detention-centre/JC2310_Joint_Statement6March12FINAL_En.pdf (accessed Sept 15, 2015).
16 Amon J, Pearshouse R, Cohen J, Schleifer R. Compulsory drug detention centers in China, Cambodia, Vietnam, and Laos: health and human rights abuses. Health Hum Rights 2013; 15: 124–37.
17 Human Rights Watch. Skin on the cable: the illegal arrest, arbitrary detention and torture of people who use drugs in Cambodia. New York, NY: Human Rights Watch, 2010.
18 Pearshouse R. “Patients, not criminals”? An assessment of Thailand’s compulsory drug dependence treatment system. HIV AIDS Policy Law Rev 2009; 2009: 11–17.
19 Cohen JE, Amon JJ. Health and human rights concerns of drug users in detention in Guangxi Province, China. PLoS Med 2008; 5: e234.
20 Amon JJ, Pearshouse R, Cohen JE, Schleifer R. Compulsory drug detention in East and Southeast Asia: evolving government, UN and donor responses. Int J Drug Policy 2014; 25: 13–20.
21 Fu JJ, Bazazi AR, Altice FL, Mohamed MN, Kamarulzaman A. Absence of antiretroviral therapy and other risk factors for morbidity and mortality in Malaysian compulsory drug detention and rehabilitation centers. PLoS One 2012; 7: e44249.
22 Walmsley R. World prison population list (tenth addition). London: International Centre for Prison Studies, 2013. http://www.prisonstudies.org/sites/default/files/resources/downloads/wppl_10.pdf (accessed Dec 21, 2015).
23 Drucker E. Population impact of mass incarceration under New York’s Rockefeller drug laws: an analysis of years of life lost. J Urban Health 2002; 79: 434–35.
24 Maruschak LM. HIV in prisons, 2001–2010. US Department of Justice, Bureau of Justice Statistics, September, 2012. http://www.bjs.gov/content/pub/pdf/hivp10.pdf (accessed Sept 1, 2015).
25 UNAIDS. UNAIDS report on the global AIDS epidemic, 2010. Geneva: Joint United Nations Programme on HIV/AIDS, 2010. http://www.unaids.org/globalreport/documents/20101123_GlobalReport_full_en.pdf (accessed Sept 1, 2015).
26 Westergaard RP, Kirk GD, Richesson DR, Galai N, Mehta SH. Incarceration predicts virologic failure for HIV-infected injection drug users receiving antiretroviral therapy. Clin Infect Dis 2011; 53: 725–31.
27 Genberg BL, Astemborski J, Vlahov D, Kirk GD, Mehta SH. Incarceration and injection drug use in Baltimore, Maryland. Addiction 2015; 110: 1152–59.
28 Adimora A, Schoenbach V, Stancil T, et al. Incarceration and heterosexual HIV infection among rural African Americans 7th Conference on Retroviruses and Opportunistic Infections; San Fransico, CA; Jan 30–Feb 2, 2000. 486.
29 Tyndall MW, Currie S, Spittal P, et al. Intensive injection cocaine use as the primary risk factor in the Vancouver HIV-1 epidemic. AIDS 2003; 17: 887–93.
30 Allen EJ, Palmateer NE, Hutchinson SJ, Cameron S, Goldberg DJ, Taylor A. Association between harm reduction intervention uptake and recent hepatitis C infection among people who inject drugs attending sites that provide sterile injecting equipment in Scotland. Int J Drug Policy 2012; 23: 346–52.
31 Cepeda JA, Niccolai LM, Lyubimova A, Kershaw T, Levina O, Heimer R. High-risk behaviors after release from incarceration among people who inject drugs in St. Petersburg, Russia. Drug Alcohol Depend 2015; 147: 196–202.
32 Spaulding AC, Sharma A, Messina LC, Zlotorzynska M, Miller L, Binswanger IA. A comparison of liver disease mortality with HIV and overdose mortality among Georgia prisoners and releasees: a 2-decade cohort study of prisoners incarcerated in 1991. Am J Public Health 2015; 105: e51–57.
33 Open Society Foundations. Improving health in pretrial detention: pilot interventions and the need for evaluation. May, 2011. https://www.opensocietyfoundations.org/sites/default/files/pretrial-detention-health-20110531.pdf (accessed June 8, 2016).
34 Stuckler D, Basu S, McKee M, King L. Mass incarceration can explain population increases in TB and multidrug-resistant TB in European and central Asian countries. Proc Natl Acad Sci USA 2008; 105: 13280–85.
35 Rubenstein LS, Amon JJ, McLemore M, et al. HIV, prisoners, and human rights. Lancet 2016; published online July 14. http://dx.doi.org/10.1016/S0140-6736(16)30663-8.
36 Telisinghe L, Charalambous S, Topp SM, et al. HIV and tuberculosis in prisons in sub-Saharan Africa. Lancet 2016; published online July 14. http://dx.doi.org/10.1016/S0140-6736(16)30578-5.
37 UNAIDS. Countries. 2013. http://www.unaids.org/en/regionscountries/countries/ (accessed July 1, 2014).
38 UNAIDS. UNAIDS report on the AIDS global epidemic 2013 (appendix). Geneva: Joint United Nations Programme on HIV/AIDS, 2013.
39 Azbel L, Wickersham JA, Grishaev Y, Dvoryak S, Altice FL. Burden of infectious diseases, substance use disorders, and mental illness among Ukrainian prisoners transitioning to the community. PLoS One 2013; 8: e59643.
40 Beyrer C, Baral SD, van Griensven F, et al. Global epidemiology of HIV infection in men who have sex with men. Lancet 2012; 380: 367–77.
41 Aceijas C, Stimson GV, Hickman M, Rhodes T. Global overview of injecting drug use and HIV infection among injecting drug users. AIDS 2004; 18: 2295–303.
42 Dolan K, Moazen B, Noori A, Rahimzadeh S, Farzadfar F, Hariga F. People who inject drugs in prison: HIV prevalence, transmission and prevention. Int J Drug Policy 2015; 26 (suppl 1): S12–15.
43 Choopanya K, Des Jarlais DC, Vanichseni S, et al. Incarceration and risk for HIV infection among injection drug users in Bangkok. J Acquir Immune Defic Syndr 2002; 29: 86–94.
44 Choopanya K, Vanichseni S, Des Jarlais DC, et al. Risk factors and HIV seropositivity among injecting drug users in Bangkok. AIDS 1991; 5: 1509–13.
45 Caplinskiene I, Caplinskas S, Griskevicius A. Narcotic abuse and HIV infection in prisons. Medicina (Kaunas) 2003; 39: 797–803 (in Lithuanian).
46 Bobrik A, Danishevski K, Eroshina K, McKee M. Prison health in Russia: the larger picture. J Public Health Policy 2005; 26: 30–59.
47 Farnia M, Ebrahimi B, Shams A, Zamani S. Scaling up methadone maintenance treatment for opioid-dependent prisoners in Iran. Int J Drug Policy 2010; 21: 422–24.
48 Gunchenko A, Kozhan N. HIV infection in the penitentiaries of Ukraine. Zh Mikrobiol Epidemiol Immunobiol 1999; 1: 31–33.
49 Dana D, Zary N, Peyman A, Behrooz A. Risk prison and hepatitis B virus infection among inmates with history of drug injection in Isfahan, Iran. ScientificWorldJournal 2013; 2013: 735761.
50 Dolan K, Hall W, Wodak A, Gaughwin M. Evidence of HIV transmission in an Australian prison. Med J Aust 1994; 160: 734.
51 Taylor A, Goldberg D, Emslie J, et al. Outbreak of HIV infection in a Scottish prison. BMJ 1995; 310: 289–92.
52 Hutchinson SJ, Goldberg DJ, Gore SM, et al. Hepatitis B outbreak at Glenochil prison during January to June 1993. Epidemiol Infect 1998; 121: 185–91.
53 O’Sullivan BG, Levy MH, Dolan KA, et al. Hepatitis C transmission and HIV post-exposure prophylaxis after needle-and syringe-sharing in Australian prisons. Med J Aust 2003; 178: 546–49.
54 Centers for Disease Control and Prevention. Hepatitis B outbreak in a state correctional facility, 2000. MMWR Morb Mortal Wkly Rep 2001; 50: 529–32.
55 Centers for Disease Control and Prevention. Transmission of hepatitis B virus in correctional facilities—Georgia, January 1999–June 2002. MMWR Morb Mortal Wkly Rep 2004; 53: 678–81.
56 Khan AJ, Simard EP, Bower WA, et al. Ongoing transmission of hepatitis B virus infection among inmates at a state correctional facility. Am J Public Health 2005; 95: 1793–99.
57 Sosa LE, Lobato MN, Condren T, Williams MN, Hadler JL. Outbreak of tuberculosis in a correctional facility: consequences of missed opportunities. Int J Tuberc Lung Dis 2008; 12: 689–91.
58 Commonwealth of Australia. Return on investment in needle and syringe programs in Australia. Canberra: Commonwealth Department of Health and Aging, 2002.
Series
www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4 13
59 Hagan H, Des Jarlais DC. HIV and HCV infection among injecting drug users. Mt Sinai J Med 2000; 67: 423–28.
60 Alter MJ. Epidemiology of viral hepatitis and HIV co-infection. J Hepatol 2006; 44 (suppl 1): S6–S9.
61 Schweitzer A, Horn J, Mikolajczyk RT, Krause G, Ott JJ. Estimations of worldwide prevalence of chronic hepatitis B virus infection: a systematic review of data published between 1965 and 2013. Lancet 2015; 386: 1546–55.
62 Reekie JM, Levy MH, Richards AH, et al. Trends in HIV, hepatitis B and hepatitis C prevalence among Australian prisoners—2004, 2007, 2010. Med J Aust 2014; 200: 277–80.
63 Sacchi F, Praça R, Tatara M, et al. Prisons as reservoir for community transmission of tuberculosis, Brazil. Emerg Infect Dis 2015; 21: 452–55.
64 Noeske J, Kuaban C, Amougou G, Piubello A, Pouillot R. Pulmonary tuberculosis in the Central Prison of Douala, Cameroon. East Afr Med J 2006; 83: 25–30.
65 Noeske J, Mbondi Mfondih S, Kuaban C. Surveillance of HIV infection in new prison entries in Cameroon, 2008–10. 6th IAS Conference on HIV Pathogenesis and Treatment; Rome; July 17–20, 2011. MOPE296.
66 Diendere EA, Tieno H, Bognounou R, et al. Prevalence and risk factors associated with infection by human immunodeficiency virus, hepatitis B virus, syphilis and bacillary pulmonary tuberculosis in prisons in Burkina Faso. Med Trop (Mars) 2011; 71: 464–67 (in French).
67 Baussano I, Williams BG, Nunn P, Beggiato M, Fedeli U, Scano F. Tuberculosis incidence in prisons: a systematic review. PLoS Med 2010; 7: e1000381.
68 Ministerio del Interior, Secretaria General de Instituciones Penitenciarias. Prevalencia de las infecciones VIH y VHC en instituciones penitentciarias 2014. Madrid: Ministerio del Interior, 2015.
69 Luciani F, Bretana NA, Teutsch S, et al. A prospective study of hepatitis C incidence in Australian prisoners. Addiction 2014; 109: 1695–706.
70 Taylor A, Munro A, Allen E, et al. Low incidence of hepatitis C virus among prisoners in Scotland. Addiction 2013; 108: 1296–304.
71 Macalino GE, Vlahov D, Sanford-Colby S, et al. Prevalence and incidence of HIV, hepatitis B virus, and hepatitis C virus infections among males in Rhode Island prisons. Am J Public Health 2004; 94: 1218–23.
72 Marco A, Gallego C, Cayla JA. Incidence of hepatitis C infection among prisoners by routine laboratory values during a 20-year period. PLoS One 2014; 9: e90560.
73 Castaneda-Hernandez DM, Martinez-Ramirez JE, Bolivar-Mejia A, Rodriguez-Morales AJ. Differences in TB incidence between prison and general populations, Pereira, Colombia, 2010–2011. Tuberculosis (Edinb) 2013; 93: 275–76.
74 Rueda ZV, Lopez L, Velez LA, et al. High incidence of tuberculosis, low sensitivity of current diagnostic scheme and prolonged culture positivity in four colombian prisons. A cohort study. PLoS One 2013; 8: e80592.
75 Hariga F. Evaluation and recommendations for the improvement of the health programmes, including for the prevention and treatment of drug dependence and of HIV and AIDS, implemented in the establishments under the responsibility of the Federal Penitentiary Service in Argentina: a follow-up of the 2008 assessment. Vienna: United Nations Office on Drugs and Crime, 2011.
76 Wong MY, Leung CC, Tam CM, Kam KM, Ma CH, Au KF. TB surveillance in correctional institutions in Hong Kong, 1999–2005. Int J Tuberc Lung Dis 2008; 12: 93–98.
77 Poposka BI, Zakoska M, Atanasova S, Milanovski N, Doneva P. Tuberculosis in prisons in the Republic of Macedonia. Eur Respir J 2013; 42 (suppl 57): 3903 (abstr).
78 Cioran N, Popescu G, Mihailescu L, Didilescu C, Cocei H, Chiotan D. TB among prisoners in Romania, 2009–2011. Eur Respir J 2013; 42 (suppl 57): 4039 (abstr).
79 Miller TL, Hilsenrath P, Lykens K, McNabb SJ, Moonan PK, Weis SE. Using cost and health impacts to prioritize the targeted testing of tuberculosis in the United States. Ann Epidemiol 2006; 16: 305–12.
80 Altice F, Azbel L, El-Bassel N, et al. The perfect storm: incarceration and the high-risk environment perpetuating transmission of HIV, hepatitis C virus, and tuberculosis in Eastern Europe and Central Asia. Lancet 2016; published online July 14. http://dx.doi.org/10.1016/S0140-6736(16)30856-X.
81 Adjei AA, Armah HB, Gbagbo F, et al. Correlates of HIV, HBV, HCV and syphilis infections among prison inmates and officers in Ghana: a national multicenter study. BMC Infect Dis 2008; 8: 33.
82 Pourahmad M, Javady A, Karimi I, Ataei B, Kassaeian N. Seroprevalence of and risk factors associated with hepatitis B, hepatitis C, and human immunodeficiency virus among prisoners in Iran. Infect Dis Clin Pract 2007; 15: 368–72.
83 Adoga MP, Banwat EB, Forbi JC, et al. Human immunonodeficiency virus, hepatitis B virus and hepatitis C virus: sero-prevalence, co-infection and risk factors among prison inmates in Nasarawa State, Nigeria. J Infect Dev Ctries 2009; 3: 539–47.
84 Beckwith C, Rick J, Flanigan T, et al. Routine jail-based HIV testing—Rhode Island, 2000–2007. MMWR Morb Mortal Wkly Rep 2010; 59: 742–45.
85 Begier EM, Bennani Y, Forgione L, et al. Undiagnosed HIV infection among New York City jail entrants, 2006: results of a blinded serosurvey. J Acquir Immune Defic Syndr 2010; 54: 93–101.
86 Burt RD, Thiede H, Barash ET, Sabin K. Recent condom use by arrested injection drug users in King County, Washington, USA. Int J Drug Policy 2006; 17: 222–29.
87 Javanbakht M, Murphy R, Harawa NT, et al. Sexually transmitted infections and HIV prevalence among incarcerated men who have sex with men, 2000–2005. Sex Transm Dis 2009; 36 (suppl 2): S17–21.
88 MacGowan R, Margolis A, Richardson-Moore A, et al. Voluntary rapid human immunodeficiency virus (HIV) testing in jails. Sex Transm Dis 2009; 36 (suppl 2): S9–13.
89 Rosen DL, Schoenbach VJ, Wohl DA, White BL, Stewart PW, Golin CE. Characteristics and behaviors associated with HIV infection among inmates in the North Carolina prison system. Am J Public Health 2009; 99: 1123–30.
90 VanHandel M, Beltrami JF, MacGowan RJ, Borkowf CB, Margolis AD. Newly identified HIV infections in correctional facilities, United States, 2007. Am J Public Health 2012; 102 (suppl 2): S201–04.
91 Javanbakht M, Boudov M, Anderson LJ, et al. Sexually transmitted infections among incarcerated women: findings from a decade of screening in a Los Angeles County Jail, 2002–2012. Am J Public Health 2014; 104: e103–09.
92 Fox RK, Currie SL, Evans J, et al. Hepatitis C virus infection among prisoners in the California State correctional system. Clin Infect Dis 2005; 41: 177–86.
93 Beyrer C, Malinowska-Sempruch K, Kamarulzaman A, Kazatchkine M, Sidibe M, Strathdee SA. Time to act: a call for comprehensive responses to HIV in people who use drugs. Lancet 2010; 376: 551–63.
94 Milloy MJ, Buxton J, Wood E, Li K, Montaner JS, Kerr T. Elevated HIV risk behaviour among recently incarcerated injection drug users in a Canadian setting: a longitudinal analysis. BMC Public Health 2009; 9: 156.
95 Wood E, Li K, Small W, Montaner JS, Schechter MT, Kerr T. Recent incarceration independently associated with syringe sharing by injection drug users. Public Health Rep 2005; 120: 150–56.
96 Althoff AL, Zelenev A, Meyer JP, et al. Correlates of retention in HIV care after release from jail: results from a multi-site study. AIDS Behav 2013; 17 (suppl 2): S156–70.
97 Iroh PA, Mayo H, Nijhawan AE. The HIV care cascade before, during, and after incarceration: a systematic review and data synthesis. Am J Public Health 2015; 105: e5–16.
98 Larney S, Toson B, Burns L, Dolan K. Effect of prison-based opioid substitution treatment and post-release retention in treatment on risk of re-incarceration. Addiction 2012; 107: 372–80.
99 Marino S, Hogue IB, Ray CJ, Kirschner DE. A methodology for performing global uncertainty and sensitivity analysis in systems biology. J Theor Biol 2008; 254: 178–96.
100 United Nations Office on Drugs and Crime, WHO (Europe). Good governance for prison health in the 21st century. A policy brief on the organization of prison health. Copenhagen: World Health Organization, 2013.
101 Thomson N, Reid G, Dolan K. Examining HIV, drug use and risk behaviours: a case study in the custodial settings of Thailand and Indonesia. Int J Prison Health 2009; 5: 180–91.
102 Avelino F, Feliciano J, Macabaya A. HIV, hepatitis B, hepatitis C and syphilis among inmates in Cebu City Jails, Philippines: seroprevalence and risk behaviours. Manila: Philippine National AIDS Council, Department of Health, 2014.
Series
14 www.thelancet.com Published online July 14, 2016 http://dx.doi.org/10.1016/S0140-6736(16)30466-4
103 Strathdee SA, West BS, Reed E, Moazan B, Azim T, Dolan K. Substance use and HIV among female sex workers and female prisoners: risk environments and implications for prevention, treatment, and policies. J Acquir Immune Defic Syndr 2015; 69: S110–17.
104 Beyrer C, Kamarulzaman A, McKee M, for the Lancet HIV in Prisoners Group. Prisoners, prisons, and HIV: time for reform. Lancet 2016; published online July 14. http://dx.doi.org/10.1016/S0140-6736(16)30829-7.
105 United Nations Office on Drugs and Crime. Handbook of basic principles and promising practices on Alternatives to Imprisonment. Vienna: United Nations Office on Drugs and Crime, 2007.
106 Martin NK, Vickerman P, Dore GJ, Hickman M. The hepatitis C virus epidemics in key populations (including people who inject drugs, prisoners and MSM): the use of direct-acting antivirals as treatment for prevention. Curr Opin HIV AIDS 2015; 10: 374–80.
107 Naggie S, Cooper C, Saag M, et al. Ledipasvir and sofosbuvir for HCV in patients coinfected with HIV-1. N Engl J Med 2015; 373: 705–13.
108 Shamliyan TA, Johnson JR, MacDonald R, et al. Systematic review of the literature on comparative effectiveness of antiviral treatments for chronic hepatitis B infection. J Gen Intern Med 2011; 26: 326–39.
109 Pisu M, Meltzer MI, Lyerla R. Cost-effectiveness of hepatitis B vaccination of prison inmates. Vaccine 2002; 21: 312–21.
110 Nelson PK, Mathers BM, Cowie B, et al. Global epidemiology of hepatitis B and hepatitis C in people who inject drugs: results of systematic reviews. Lancet 2011; 378: 571–83.
111 WHO. Hepatitis B vaccines WHO position paper. Geneva: World Health Organization, 2009.
112 Dara M, Acosta CD, Melchers NV, et al. Tuberculosis control in prisons: current situation and research gaps. Int J Infect Dis 2015; 32: 111–17.
113 Todrys KW, Amon JJ. Criminal justice reform as HIV and TB prevention in African prisons. PLoS Med 2012; 9: e1001215.
114 Todrys KW, Amon JJ, Malembeka G, Clayton M. Imprisoned and imperiled: access to HIV and TB prevention and treatment, and denial of human rights, in Zambian prisons. J Int AIDS Soc 2011; 14: 8.
115 Schwitters A, Kaggwa M, Omiel P, Nagadya G, Kisa N, Dalal S. Tuberculosis incidence and treatment completion among Ugandan prison inmates. Int J Tuberc Lung Dis 2014; 18: 781–86.
116 WHO. Global tuberculosis report. Geneva: World Health Organization, 2015.
Supplementary appendixThis appendix formed part of the original submission and has been peer reviewed. We post it as supplied by the authors.
Supplement to: Dolan K, Andrea L W, Moazen B, et al. Global burden of HIV, viral hepatitis, and tuberculosis in prisoners and detainees. Lancet 2016; published online July 14. http://dx.doi.org/10.1016/S0140-6736(16)30466-4.
1
Global burden of HIV, viral hepatitis and tuberculosis among prisoners and detainees
Authors: Kate Dolan, PhD; Andrea L Wirtz, PhD; Babak Moazen, MsCIH; Alison Galvani,
PhD; Martial Ndeffo-mbah, PhD; Stuart Kinner, PhD; Ryan Courtney, PhD; Martin Mckee, DSc;
Joseph J Amon, PhD; Lisa Maher, PhD; Margaret Hellard, PhD; Chris Beyrer, MD.; Fredrick
Altice, MD
Web Supplementary Appendix
Table of Contents
A1. Comprehensive search methods and meta-analysis ................................................................................................ 2
A4. Prevalence data located for HIV, HCV, HBV TB and HIV/TB co infection among prison populations, published
between 2005-2015........................................................................................................................................................ 8
A5. National and regional HIV prevalence among prison populations, published between 2005-2015 ....................... 9
A6. National and regional estimates of anti-HCV prevalence among prison populations, published between 2005-
A12. Incidence of HIV and viral hepatitis among prisoners, published 2005 – 201592,97,162,279,280 .............................. 32
A13. Modeling the interplay between prison and community HIV transmission among PWIDs ............................... 33
A14. Partial rank correlation coefficients (PRCCs) of model parameters. .................................................................. 41
A15. Model Results: Effect of different incarceration-based interventions on HIV incidence after 5 years. .............. 42
Web Appendix References: ......................................................................................................................................... 43
2
A1. Comprehensive search methods and meta-analysis
Our search focused on identifying recent studies that utilized biologic measures to assess prevalence and/or
incidence of HIV, HCV, HBV, TB, and co-infections among people living in prisons and other closed facilities. The
search strategy was developed in collaboration with an informationist at Johns Hopkins University. Searches
followed controlled vocabulary and keywords, which were used in combination for the concepts of HIV, HCV,
HBV, TB, prisons, jail, and compulsory drug detention. Index specific terms are provided in Web Appendix 2
(below). We searched Pubmed, Embase, CINAHL, and Criminal Justice Abstracts with Full Text to identify peer
reviewed publications that published in any language from January 1, 2005 through July 15, 2015. Identified
publications were compiled in Refworks reference manager (version 2.0, ProQuest, Ann Arbor, MI) and duplicates
were removed. Additional reports published between 2005 and 2015 by the World Health Organization and the UN
Office on Drug and Crime were identified and also included in the literature review. Identified literature was then
independently reviewed by two research assistants, with faculty oversight and final determination, for eligibility
using a title and abstract review followed by a full text review of publications with data abstraction. Supplementary
Web Appendix 3 provides the PRISMA search flow diagram.
In the instance that no data was extracted from the systematic review search for countries on either HIV, HCV,
HBV, or TB in prisons and other closed settings, the authors called for papers and reports through potential grey
literature sources and via an email request sent to experts and researchers from the UN and/or working or affiliated
with prison systems. This invitation email requested further data on either HIV, HCV, HBV or TB from each
country. Emails were sent in October-November, 2015. In the first instance we searched the prisonstudies.org
website for email contact details of the Prison/Corrective Services Department for each country. If there was an
email contact listed, they were emailed an invitation letter. Additionally, when no Prison Department or Correct
Service contact was listed on the prisonstudies.org website we searched Google to ensure a Prison/Corrective
Services contact was ascertained and sent an email for each country, whenever possible. We also searched Google
for email contacts relevant to Health Department’s in each country and sent emails to not-for-profit/NGO agencies,
where relevant in each country. At a minimum one email was sent per country with an upper limit of five emails
sent. This call for data yielded only 17 countries that provided data and this was from the Barbados, Dominican
266; Montenegro 174; Morocco 222; Mozambique 61; Myanmar 113; Namibia 144; Nauru 140; Nepal 59; New
Zealand 194; Nicaragua 171; Niger 39; Norway 71; Oman 36; Palau 343; Panama 392; Papua New Guinea 61;
Paraguay 158; Peru 242; Philippines 121; Portugal 138; Qatar 53; Republic of Korea 101; Republic of Moldova 215;
Romania 143; Rwanda 434; Saint Kitts and Nevis 607; Saint Lucia 349; Saint Vincent and the Grenadines 387; Sao
Tome and Principe 101; Saudi Arabia 161; Senegal 62; Serbia 148; Seychelles 799; Sierra Leone 55; Singapore 227;
Slovakia 184; Slovenia 73; Solomon Islands 56; Somalia N/A; South Sudan c.65; Spain* 136; Sri Lanka 92; Sudan
50; Suriname 183; Swaziland 289; Sweden 55; Switzerland 84; Syrian Arab Republic 60; Taiwan* ; The Former
Yugoslav Republic of Macedonia 147; Timor-Leste 50; Togo 64; Tonga 166; Trinidad and Tobago 258; Tunisia 212;
Turkmenistan 583; Tuvalu 110; United Arab Emirates 229; United Kingdom: Northern Ireland 87; United Kingdom:
Scotland 143; United Republic of Tanzania 69; Uruguay 291; Uzbekistan 150; Vanuatu 87; Venezuela 174; Viet
Nam 154; Yemen 53; Zimbabwe 145
27
A9. HIV/TB Co-infection among prison populations, published 2005 - 2015
Region / Country
Year(s) of
data
collection
HIV/TB co-
infection
among total
prison sample
(%)
Total
prevalence of
TB among
HIV+
(%)
Total
prevalence of
HIV among
TB+
(%)
East and Southern Africa
Ethiopia207
2011 3.6% 47.4% 34.6%
South Africa7,8,209
2009-2010 1.6% 6.6% 42.4%
2009-2010 NR 6.2% 44.1%
2009-2010 5.1% NR 58.0%
Uganda235
2011-2012 NR NR 57.0%
Zambia16
2010-2011 1.6% 6.4% 45.8%
West and Central Africa
Burkina Faso21
2009 NR 20.0% 75.0%
Cameroon22,236
2003-2004 NR NR 25.0%
2009 NR NR 10.0%
Guinea211
2009-2010 0.3% NR 11.1%
Nigeria30
2006 8.9% NR NR
Middle East and North
Africa
No data identified or met inclusion criteria for the MENA
Region
Asia and Pacific
Australia237
2001 0.1% NR 0.9%
Malaysia217
2010 NR 16.9% NR
Eastern Europe Central Asia
Russia238
2001 NR 12.2% NR
Ukraine239,240
2005-2006 NR NR 17.7%
2006 NR NR 23.7%
Western Europe
Italy241
2001-2002 0.9% 7.4% NR
North America
USA225,242
1998 NR NR 23.1%
2003 NR NR 11.3%
Caribbean No data identified or met inclusion criteria for the Caribbean
Latin America
Brazil226,227,230
2002 0.6% NR 14.6%
2003-2004 0.0% NR NR
2007-2008 0.3% NR 6.9%
Colombia243,244
2010-2011 0.1% NR 14.3%
2010-2012 0.2% 11.5% 4.2%
Notes: Includes national and subnational data; includes latent and active TB infection
28
A10. Infectious disease prevalence among imprisoned people who inject drugs, published 2005-2015
Country Facility Year of
collection
HIV
(%)
Anti-HCV
(%)
HBsAg
(%)
TB
(%)
West and Central Africa
Ghana26
Central Prisons 2004-2005 11.4% 49.3%
East and Southern Africa No data identified or met inclusion criteria for the East and Southern Africa
Middle East and North Africa
Iran136,194,245-256
Prison 2001-2002 NR 78.3% 1.5%
Prison 2002 0.7% 31.5%
Prison 2002 15.1% 64.8% 4.7%
Central Prison 2002-2003 18.2% 13.6%
Central Prison 2002-2003 NR 95.4%
Prison 2003 88.9%
Prison 2003 14.0% 76.0% 4.0%
Detention center 2006 NR 80.0%
Jail 2007 6.6%
Prison 2007 2.3%
Correctional
center
2008-2009 NR 50.0% 3.3%
Prison 2009 6.4%
Prison 2009 NR 41.6%
Prisons 2009 13.9%
Prison 2009-2010 4.5%
Central Prison
and Drop in
Centers
NR NR 43.4%
Asia and Pacific
Australia46,48-50,52,257-260
Prison 2004 0.8% 56.0% 27.0%
Correctional
Center
2005 NR 40.0%
Prison 2006 NR 7.6% 7.6%
Correctional
center
2005-2009 22.6%
Prison 2008-2010 47.4%
Prison 2010 0.0% 51.0% 18.8%
Reception
Prisons
2010 0.0% 18.4%
Prison 2013 0.0% 58.2% 19.0%
Prisons NR 17.0%
Prison NR 20.7%
Prison NR 29.4%
Indonesia57
Prison 2009 13.8% 50.0% 1.1%
Pakistan145
Central Prison NR 84.2%
Taiwan199,261-263
Prisons 1994 26.4%
29
Correction center 1997 59.4% 22.2%
Correctional
institution
2005-2007 42.0%
Jail 2008-2010 52.8% 91.3% 15.3%
Eastern Europe and Central Asia
Bulgaria69,71
Prison 2009 61.4%
Juvenile
correctional
facility
2010 66.7%
Croatia264
Prison 2007 0.0% 44.0% 0.5%
Western Europe
England265
Prison NR 32.6%
Germany266
Prison 1998-1999 18.0% 83.0% 53.0%*
Ireland85
Prison 2011 6.0% 41.5%
Italy157
Prisons 2001-2002 18.2% 57.1%*
Prisons 2001-2002 24.5%
North Ireland86
Prison 2004 0.0% 2.8% 2.8%
Netherlands161
2014 58.0%
Portugal88
Prisons 2005 44.4% 68.9%
Scotland162
Prison 2010-2011 53%
Spain95
Prison 2001 53.2% 92.7%
North America
Canada99,100
Prison 2003 10.2% 55.6%
Remand
Facilities
2003-2004 5.7% 54.7%
USA104,106,108,113,114,118,170,172
,181,267,268
Sentenced and
Pretrial detainees
1994-1996 20.4%
Jail and Prison
facility
1998-2000 84.9%
Jails 1998-2002 2.5%
Juvenile
Detention Center
2000-2001 35.8%
Jail 2003-2006 1.4%
Prison 2004-2006 9.4%
Jail 2006 16.9%
Correctional
facility
2007 1.0%
County
Correctional
Facility
2009-2011 0.3% 38.0%
Correctional
facility
NR 66.2%
Caribbean
Belize121
Central Prison 2005 0.0%
Latin America
Brazil126,127,187,189,191
Prison 2003 34.5% 72.7%
Prison 2007 13.8% 34.5% 51.7%*
30
Prisons 2007-2008 66.7%
Prison 2009-2010 20.6%
Prisons 2010 0.4%
Mexico132
Correctional
facility
2001-2002 66.7%
Notes: NR: Not reported; * anti-HBc
31
A11. HIV and viral hepatitis among inmates in Compulsory Drug Detention Centers (CDDCs) published between 2005 - 2015
HIV Prevalence
anti-HCV Prevalence
HBsAg
Country Location Year(s) of
data
collection
Total CDDC
prevalence
(%)
PWID
prevalence
(%)
Total CDDC
prevalence (%)
PWID
Prevalence
(%)
Total CDDC
prevalence (%)
PWID
Prevalence
(%)
China269-
272
Beijing,
Shanghai,
Guangdong
2009-2010 0.2% NR 44.0% NR NR NR
Dongguan 2008 4.% NR NR NR NR NR
Beijing 2007 5.2% 10.3% NR NR NR NR
Guangxi 2011 9.5% NR NR NR NR NR
Taiwan273
Taoyuan 2005 6.9% 25.5% 30.5% 89.6% 16.9% 16.7%
Vietnam274
Hatay 2007-08 NR 19.8% NR 76.9% NR NR
Iran193,253,2
54,275-278
Tehran 2006 NR 24.4% NR 80.0% NR 5.8%
Tehran 2001-2002 17.% 7.8% NR NR 4.5% 1.5%
Sari 2001-2003 NR NR 30.8% 77.0% NR NR
Note: CDDCs include injecting and non-injecting people who use drugs
32
A12. Incidence of HIV and viral hepatitis among prisoners, published 2005 – 201592,97,162,279,280
Australia (South Australia)
Australia (New South Wales)*
Scotland USA (Rhode Island)
Australia (New South Wales)
Scotland
Australia (Queensland)*
Spain (Barcelona)
Estonia USA (Rhode Island)
Spain
USA (Georgia)
USA (Rhode Island)
0
2000
4000
6000
8000
10000
12000
14000
16000
1999 2001 2003 2005 2007 2009 2011 2013 2015
Inci
de
nce
pe
r 1
00
,00
0 p
ers
on
ye
ars
Year
HCV
HIV
HBV
33
A13. Modeling the interplay between prison and community HIV transmission among PWIDs
Summary:
To evaluate the fundamental dynamics that govern the interplay of HIV transmission within and between prison and
community settings, we developed a model parameterized by data of incarceration history and risk behavior among
PWIDs. Rates of HIV infection among PWIDs were stratified into communities with moderate PWID HIV
prevalence, ranging from 5 to 20%, and communities with high PWID HIV prevalence, over 20%.281
To address
empirical uncertainty regarding model parameters, we used a Bayesian inference approach to fit our model to
epidemiological data of HIV prevalence among PWIDs in and out of prisons, as well as HIV incidence in prison.
We used the fitted model to quantify the contribution of post-release ART discontinuation and behavior change of
PWIDs who temporarily share syringes while incarcerated and during a transitory period on HIV incidence among
PWIDs. Furthermore, we evaluated the impact of harm reduction programs, such as opioid substitution treatment
(OST) and lower rates of incarceration for non-violent drug users, and ART retention in care post-release for
curtailing HIV incidence across a range of epidemics parameterized from different countries.
Our model accounted for the elevated frequency of needle sharing in prison as well as discontinuation of ART
during a transitory period following prison release.282-286
Given the sparsity of data regarding the impact of this post-
release behavior change on HIV transmission, we conducted a sensitivity analysis to evaluate the contribution of the
frequency of post-release needle sharing and percentage discontinuation of ART treatment on HIV incidence among
PWID within the community and prisons. We denoted as ‘temporary sharers’ PWIDs who share syringes while
incarcerated, a proportion of whom share needs a transient duration following release, but not beyond. A detailed
description of our modeling methodology, parameterization, assumptions and limitations are presented in the
following section. We limited our model to inmates who inject heroin, as the majority of research and prevention
has focused on these individuals. To assess the variability of the effectiveness of interventions in settings with
different HIV epidemics among injection drug users, we considered parameters from a range of settings with
communities of PWID.
Model Structure:
To evaluate generalized principles governing the interplay between HIV transmission among people who inject drug
(PWID) in the community and in prison around the world, we developed a deterministic compartmental model of
drug injection transmission of HIV among the general community and incarcerated populations. To determine the
posterior distributions of model parameters through a Bayesian analysis, we used a range of estimates from HIV
prevalence among PWID in the general population, HIV prevalence among PWID in prisons and overall HIV
incidence in prisons. We then used our parameterized model to quantify the relative contribution of incarceration to
HIV incidence among PWID in different communities. Additionally, we evaluated the impact on HIV transmission
of harm reduction programs (opioid substitution treatment), ART retention in care post-release, and lower rates of
incarceration for non-violent drug users.
Model Formulation:
Transmission dynamics among PWID within and between prisons and the community were modeled by differential
equations. The state variables are given by ,
h
k iX , where h is the HIV infection status and stage of progression (1 =
susceptible; 2 = acute infection; 3 = latent infection with CD4 greater than 350 cells / L ; 4 = latent infection with
CD4 less than 350 cells / L ; 5 = pre-AIDS with elevated viremia; 6 = AIDS; 7 = initiated ART with CD4> 350
cells / L ; 8 = initiated ART with low CD4 or pre-AIDS; 9= initiated ART with AIDS), i is the incarceration
status (1 = currently incarcerated, 2 = recently released (within six months), 3 = previously incarcerated, 4 = never
incarcerated), and k is injection risk behavior (1 = non-PWID, 2 = PWID who do not share syringes, 3 = PWID
who share syringes). We stratify incarcerated PWIDs who did not share syringes prior to incarceration between
permanent non-sharers and temporary sharers. Temporary sharers only share syringes in prison, consistent with the
greater proportion of PWIDs who share syringes in prison than out of prison.287-290
Hence, we add a temporary
syringe sharer group to the injection risk behavior ( 4k ). In the description that follows, a dot ( ) in the position
of these indices indicates a sum across that index. To account for a transitory elevation of injection drug risk
behavior and discontinuation of ART treatment following release from incarceration,282,285-287
we distinguish
34
between PWIDs who have been recently released within the last six months and PWIDs released further in the past
who have resumed the relatively lower risk behavior regarding injection drug use and ART treatment exhibited prior
to incarceration. 1
, 1 1 1
, , , , , ,
,
k i
k k i k i k i k k i k i k i
k i
dXR X X X
dt
(1.1)
2
, 1 2 2
, , 2 , , ,
,
( )k i
k i k i k k i k i k i
k i
dXX X X
dt
(1.2)
3
, 4 2 3 3
2 , 3 , , ,
,
(1 ( )) ( )k i hCD
k i k k i k i k i
k i
dXa t X X X
dt
(1.3)
4
, 4 3 4 4
3 , 4 , , ,
,
(1 ( )) ( )k i lCD
k i k k i k i k i
k i
dXa t X X X
dt
(1.4)
5
, 4 4 5 5
4 , 5 , , ,
,
(1 ( )) ( )k i lCD
k i k k i k i k i
k i
dXa t X X X
dt
(1.5)
6
, 5 6 6
5 , 6 , , ,
,
(1 ) ( )k i AIDS
k i k k i k i k i
k i
dXa X X X
dt
(1.6)
7
, 4 2 7 7
2 , 7 , , ,
,
( )k i hCD
k i k k i k i k i
k i
dXa X X X
dt
(1.7)
8
, 4 3 4 8 8
3 , 4 , 8 , , ,
,
( ) ( )k i lCD
i k i k i k k i k i k i
k i
dXa X X X X
dt
(1.8)
9
, 5 9 9
5 , 9 , , ,
,
( )k i AIDS
k i k k i k i k i
k i
dXa X X X
dt
(1.9)
We denote the rate of entering PWID status k as kR , which is set as to maintain a constant population size. For
k=1, 2, 3, 4, respectively, we have
9
1 1 1, 1,
6
j
j
j
R X X
,
3 9
2 , ,
2 6
(1 ) 1idu idu j
k k j k
k j
R X X
,
3 9
3 , ,
2 6
idu j
k k j k
k j
R X X
, and
3 9
2 , ,
2 6
1idu idu j
k k j k
k j
R n X X
, where
idu is the proportion of PWIDs who share syringes in the
general population. The proportion of temporary sharers is denoted as ( ) (1 )idu idu idu idu
inc , where
idu
inc is the proportion of PWIDs who share syringes while incarcerated. To prevent negative values of idu ,
idu
inc
was sampled under the condition of being greater or equal to idu . New individuals enter the model as never having
been incarcerated, with , 1k i for 4i and 0 otherwise. The background death rate unrelated to HIV for PWID
status k is denoted k . To model the impact of expanding coverage of antiretroviral therapy we incorporated
4 4( ), ( ),and ( )HCD LCD AIDSa t a t a t as the proportions of individuals with high CD4 cell count (more than 350
cells / L threshold), low CD4 cell count (less than 350 cells / L threshold), and AIDS (advanced HIV/pre-AIDS
35
stage), respectively, with incarceration status i that are initiated on treatment at that time. We assumed that ART is
initiated at CD4 count threshold of less than 350. Thus, for the base case the status quo analysis, we assumed that 4( ) 0HCDa t . For simplicity, we assumed a uniform ART initiation rate irrespective of incarceration status. The
force of infection for individuals of specific injection drug use status k , and incarceration status i is given by ,k i .
4 9, 0 2 5 6
, , , , , ,
3 7,
( )
idu
k i j j
k i acute k latent k pre aids k aids k art k
j jk
cX X X X X
X
,
where if 1i i (currently incarcerated) and *=sum of never incarcerated, recently released, and previously
incarcerated if 2,3,4i . The transmission rate of HIV per use of contaminated syringes is denoted 0
idu if the
syringe was last used by an individual infected with HIV. The relative infectiousness of individuals in HIV stage h
compared to those with latent infection is denoted h . The frequency of syringe sharing for those of incarceration
status i is denoted ,k ic , where , ,k i k i ic f and where ,k if is the frequency of injection and i is the percentage
of PWIDs sharing syringes. Given our focus on syringe sharing as the route for HIV transmission under
consideration, we set 1, 0i and 2, 0i , such that transmission occurs exclusively between sharers. This is a
conservative assumption given that a sharer could become infected from a contaminated needle from a non-sharer.
Further studies on HIV transmission among PWIDs could explore relaxing this assumption by accounting for
differential infection risk from sharers and non-sharers.
Consistent with epidemiological behavioral studies of PWIDs with a history of incarceration,282-284,291
we assumed
that a proportion idu of recently released PWIDs who shared syringes in prison but not prior to incarceration will
continue to share syringes for a transitory period following release. After this transitory period, the proportion of
PWIDs who share syringes returns to that of PWIDs who have never been incarcerated, given that PWIDs typically
resume the injection behavior they exhibited prior to incarceration.282,284
In addition, epidemiological studies have
shown that more than 40% of the incarcerated individuals on ART would have poor adherence to treatment or even
discontinue entirely within a period of the first six-months post-release.285,286,292,293
We define the movement in and out of prison through k , a transition matrix which includes the rate of first-time
incarceration that depends on PWID status,inc
kr , rate of discharge from prison that depends on PWID status,disc
kr ,
average duration of the post-release transitory period of elevated risk behavior,1 elp
kr , and rate of re-incarceration
that depends on PWID status,rinc
kr .
To assess the benefits that could be gained from expanding harm reduction measures such as opioid substitution
treatment (OST), we incorporated an intervention cofactor within the force of infection ,k i as
, ,(1 (1 ))idu idu
j i j k i , where 0j for no OST and 1j for OST. The proportion increase in intervention j
among individuals of incarceration status i is denoted ,
idu
j i and the relative decrease in the force of infection is
denoted idu
j . The relative force of infection under OST 1
idu was varied in the uncertainty analyses from 0.16 to
0.5, consistent with epidemiological studies.294-296
The mean duration of HIV infection phase h is denoted by 1 h .
Consistent with empirical studies, we assumed that PWIDs who continue OST after incarceration have a 20% (95%
CI: 10-30%) reduced risk of re-incarceration while they are under treatment.297
Baseline coverage of OST was not
explicitly modeled, but assumed to be implicitly included in epidemiological data on HIV prevalence and incidence.
Altogether, the population dynamics of incarceration at equilibrium is given by:
36
,1
inc
n nn ninc
n n
rP A
r
, ,2 ,1 ( )disc elp
n n n n nP P r r , ,3 ,1( ) ( ) ( )elp disc rinc elp
n n n n n n n nP P r r r r ,
,4 ( )inc
n n n nP r ,with ( )( )
( )( )( )
rinc elp
n n n nn disc rinc elp rinc elp disc
n n n n n n n n n
r rA
r r r r r r
,
where ,n iP is the proportion of individuals with injection drug status n ( 1n for non-PWID, and 2n for PWID)
and incarceration status i .The system of differential equations was numerically integrated in continuous time in
MATLAB 8.5® using a fourth-fifth order Runge-Kutta solver (“ode45”).
Model fitting and analysis:
To estimate posterior distributions for the model parameters governing HIV transmission, injection drug behavior
and incarceration, we using a Bayesian melding method to calibrate the model to epidemiological data of HIV
prevalence among PWIDs, HIV prevalence among incarcerated PWIDs and HIV incidence in prison.298
Data were
divided between communities with high PWID HIV prevalence that ranged from 20 to 60% and communities with
moderate PWID HIV prevalence that ranged from 5 to 20%.281,299-301
Empirical estimates of HIV prevalence among
incarcerated PWID ranged from 22 to 80%302,303
for communities with high PWID HIV prevalence and 5.7 to
28.6%106,118,304-306
for communities with moderate PWID HIV prevalence. For annual HIV incidence in prisons, we
derived a pooled estimates from the literature that ranged between 0.01 and 1.8%307,308
for communities with
moderate PWID HIV prevalence, and 0.01–4% 72,307,308
for communities with high PWID HIV prevalence. Model
parameters were randomly sampled from their prior distributions, and each parameter set sampled was then used to
simulate a HIV epidemic which was run for a randomly sampled duration between 10 to 45 years, consistent with
the range of estimates of the durations of injection drug epidemics in different settings worldwide.288
To establish
the baseline in the absence of the improvements in intervention that we are considering here, we simulated status
quo base case epidemics.
To compare model predictions with observed data, we applied a Bayesian melding approach, incorporating a
likelihood function and using all available prior information regarding model inputs to generate posterior
distributions of model parameters. We used prior distributions defined from epidemiological and clinical studies
(Tables S1-S2) for each input model parameter. The likelihood distribution was defined as
prison HIV incidence [Community PWID prev range] [Prison PWID prev range]( ) 1 1Lik Tr which is the product of a triangular
distribution (with mode: 0.1%, and ranges detailed above) for observed HIV incidence data and uniform
distributions for the observed HIV prevalence data among PWIDs in communities and HIV prevalence among
incarcerated PWIDs. The function 1[a,b]
(x) =1,if xÎ [a,b] and 0 otherwise.
To describe the Bayesian melding algorithm, we denote the simulation model (Eqs. 1.1-1.9) by M, the
epidemiological parameters (Tables S1-S2) by and the model-predicted output by ( )M . We denote the
prior distribution for each model parameter by ( )q . We denote the data by W and the associated likelihood of the
model outputs by ( ) Pr( )L W . The posterior distribution of inputs is then proportional to ( ) ( )q L . We
implemented a sample-importance-resample algorithm [32] to generate the posterior distribution. Specifically, we
first generated a set of input parameters, ( )i , by randomly sampling from the respective prior distributions i times.
We then evaluated the model using that set of parameters, ( ) ( ( ))i M i , for each run i. Next, we calculated the
corresponding likelihood for the model run. For each sample ( )j , with non-zero corresponding likelihood, the
sampling weight was
1
( ) ( ( )) ( ( ))nT
k
j L j L k
. To ensure a sufficient sample from the posterior
distributions, we set i=100,000.
37
The number of non-zero likelihood samples, nT , was 2796 for the high PWID HIV prevalence communities and
1341 for the moderate prevalence communities. To generate the posterior distribution for the inputs, we repeated
this sampling procedure 10000 times with replacement, using a probability of selection proportional to the sampling
weights. Output from the simulation resampled most frequently (i.e., the simulation that most closely fit the
empirical prevalence data) represented the estimated mode for the output parameters of interest. The 2.5th
and 97.5th
percentiles of the inputs (and the resulting model outputs) correspond to 95% credible limits.
The fitted model was used to evaluate the contribution of incarceration to HIV infection among PWIDs and the
potential impact of a range of different program interventions involving incarceration and post-incarceration harm
reduction as well as lower incarceration rates. Simulated epidemics cover a wide range of scenarios from endemic
state to slowly increasing early phase epidemic scenarios with low prevalence that are collectively applicable to
different settings (Figure S2). Interventions were assumed to be initiated in 2016, which was the latest year in model
fitting (Figure S2).
Figure S2 Sample of model simulations of HIV epidemics parameterized to correspond to a range of HIV
prevalences. This sample of 250 simulated epidemics are shown for illustration.
Given that our focus was estimation of the contribution of incarceration to injection related HIV transmission among
injection drug users and the possible impact of interventions on the HIV epidemic among PWIDs, we did not
address sexual, tattoo-related or mother-to-child transmission of HIV. As for any modeling study, we made simplify
assumptions that capture the fundamental dynamics pertinent to the objectives of our analyses without including
extraneous details. For simplicity, our model did not incorporate the dynamics of initiation and cessation of injection
drug use, such that the proportion of PWID in the population was assumed to be constant. To assess the robustness
of our model projections to this assumption, we conducted uncertainty analysis to consider the uncertainty and
variability surrounding incarceration and post-incarceration injection drug behaviors and harm reduction
interventions.
38
Table S1: HIV natural history parameters used in our HIV transmission dynamic model.
Variable Meaning (units) Prior distributions refs
Posterior distribution
Median (95% credible interval)
0
idu HIV transmission per use of shared
syringe (if last person to use was
infected)
Uniform(0.0029,0.024) 309,310
M: 0.0032 (95%CrI: 0.003 – 0.013)
H: 0.0092 (95%CrI: 0.003 – 0.013)
acute Relative infectiousness of individuals
in acute phase compared to latent
infection
Uniform(4.5,27) 311,312
M: 18.39 (95%CrI: 4.86 – 26.02)
H: 18.79 (95%CrI: 4.85 – 26.45)
latent Relative infectiousness of individuals
in latent phase compared to latent
infection
1 NA
pre aids Relative infectiousness of individuals
in pre-AIDS phase compared to latent
infection
Triangular(5,7,10) 311
M: 7.16 (95%CrI: 5.52 – 8.11)
H: 6.47 (95%CrI: 5.45 – 8.12)
aids Relative infectiousness of individuals
in AIDS phase compared to latent
infection
0 311
NA
art Relative infectiousness of virally-
suppressed individuals on ART
compared to latent infection
0.08 313
NA
21 Mean duration of acute phase
infection (months)
Triangular(1.23,2.9,6) 311
M: 5.91 (95%CrI: 1.30 – 5.79)
H: 4.88 (95%CrI: 1.35 – 5.8 )
31 Mean latent period with CD4 count
higher or equal to 350 (years)
Triangular(5.3,6.5,7.7) 314-316
M: 7.27 (95% CrI: 5.37 – 7.62)
H: 6.64 (95%CrI: 5.37 – 7.65)
41 Mean latent period with CD4 count
lower than 350 (years)
Triangular(1.7,2.5,3.3) 314-316
M: 2.83 (95%CrI: 1.75 – 3.28)
H: 2.99 (95%CrI: 1.76 – 3.27)
51 Mean duration with elevated viral load
during pre-AIDS phase (months)
Triangular(4.81,9,18) 311,315,317
M: 16.19 (95% CrI: 5.09 – 17.29)
H: 13.19 (95%CrI: 5.18 – 17.6)
61 Mean duration of AIDS before death
(months)
Triangular(6.97,10,28) 311,315
M: 10.59 (95% CrI: 7.52 – 27.37)
H: 25.17 (95%CrI: 7.56 – 27.52)
71 Mean duration of viral suppression for
individuals on ART with high CD4
(years)
Triangular(8,20,40) 288,318
M: 8.35 (95% CrI: 8.20 – 19.64)
H: 11.57 (95%CrI: 8.27 – 19.64)
81 Mean duration of viral suppression for
individuals on ART with low CD4
(years)
Triangular(5,12,20) 288,317-319
M: 16.18 (95% CrI: 5.23 – 19.27)
H: 7.9 (95%CrI: 5.26 – 19.49)
91 Mean duration of viral suppression for
individuals on ART with AIDS
(years)
Triangular(2,4.8,9.7) 318
M: 5.57 (95% CrI: 2.13 – 9.48)
H: 9.13 (95%CrI: 2.21 – 9.46)
rta Annual ART initiation rates for low
CD4 count and AIDS
Calibrated to have ART
coverage between 5 - 60% 320,321
M: 0.44 (95% CrI: 0.03 – 0.66)
H: 0.53 (95%CrI: 0.03 – 0.6)
39
Table S2: IDU and incarceration parameters used in our HIV transmission dynamic model.
Variable Meaning (units) Prior distribution refs
Posterior distribution
Median (95% credible interval)
incP Proportion of incarcerated
population
Uniform(0.0015, 0.012) 322,323
M: 0.009 (95% CrI: 0.002 – 0.0117)
H: 0.007 (95%CrI: 0.002 – 0.0117)
1 Rate of exiting the model for
non-PWID
0.02* NA
2 Average non-HIV mortality rate
for PWID
Uniform(0.009,0.07) 324
M: 0.06 (95% CrI: 0.011 – 0.06)
H: 0.01 (95%CrI: 0.011 – 0.07)
iduJ Proportion of population who
are PWID
Uniform(0.002,0.052) 325
M: 0.009 (9%% CrI: 0.0027 – 0.05)
H: 0.007 (95%CrI: 0.0035 – 0.05)
idu
incJ Proportion of incarcerated who
are PWID
Uniform(0.1,0.75) 287
M: 0.18 (95% CrI: 0.12 – 0.73)
H: 0.50 (95%CrI: 0.12 – 0.73)
idu Proportion of PWID who share
syringes out of prison
Uniform(0.1,0.75)288-290,326,327
M: 0.28 (95% CrI: 0.11 – 0.40)
H: 0.72 (95%CrI: 0.19 – 0.72)
idu
inc Proportion of PWID who share
syringes in prison Max(
idu
,Uniform(0.6,0.95)) 287
M: 0.89 (95% CrI: 0.60 – 0.94)
H: 0.73 (95%CrI: 0.61 – 0.93)
iduf Number of injections per non-
incarcerated PWID annually
Uniform(107,260) 327-329
M: 120 (95% CrI: 112 – 256)
H: 190 (95%CrI:111 – 256)
idu
incf Number of injections per PWID
in prison annually
Uniform(6,100) 330,331
M: 6 (95% CrI: 7 - 79)
H: 64 (95%CrI: 11 - 97)
ART Proportion of recently released
HIV infected individuals who do
not discontinue ART during the
post-release transition
Uniform(0,0.6) 285,286,292,293
M: 0.28 (95% CrI: 0.02 – 0.58 )
H: 0.20 (95%CrI: 0.02 – 0.58)
idu Proportion of injections of
syringe-sharers that are shared
(non-incarcerated)
Uniform(0.49,0.72) 327
M: 0.51 (95% CrI: 0.5 – 0.71)
H: 0.52 (95%CrI: 0.5 – 0.71)
idu
inc Ratio of injections of syringe-
sharers that are shared (in prison
compared to non-incarcerated)
Uniform(1,2)Assumed±
M: 1.02 (95% CrI: 1.01 – 1.88)
H: 1.68 (95%CrI: 1.02 – 1.87)
iduI Proportion of PWID who has
never being incarcerated
Uniform(0.2,0.7) 299,332
M: 0.62 (95% CrI: 0.21 – 0.68)
H: 0.29 (95%CrI: 0.21 – 0.68)
11 discr Average duration of
incarceration for non-PWID
(years)
Uniform(2,10) 333,334
M: 2.52 (95% CrI: 2.15 – 9.76)
H: 4.85 (95%CrI: 2.19 – 9.85)
2 31 ,1disc discr r Average duration of
incarceration for PWID (years)
Uniform(1,4) 333,335
M: 2.28 (95% CrI: 1.06 – 3.90)
H: 2.76 (95%CrI: 1.10 – 3.92)
31 elpr Average duration of post-release
transitory period (months)
Uniform(1,6) 283,284,336,337
M: 5.92 (95% CrI: 1.18 – 5.85)
H: 3.53 (95%CrI: 1.1 – 5.85)
1
rincr Rate of re-incarceration for non-
PWID (years-1
)
Uniform(0.1,0.5) 338
M: 0.42 (95% CrI: 0.1.02 – 0.44)
H: 0.28 (95%CrI: 0.102 – 0.45)
2 3,rinc rincr r Rate of re-incarceration for
PWID (years-1
)
Uniform(0.1,3) 297,338
M: 0.23 (95% CrI: 0.104 – 1.46)
H: 0.22 (95%CrI: 0.103 – 1.48)
40
Variable Meaning (units) Prior distribution refs
Posterior distribution
Median (95% credible interval)
2,1P Fraction of PWID who are
currently in prison 2 2
2
2 2
inc
inc
rA
r
M: 0.16 (95%CrI: 0.06 – 0.39 )
H: 0.21 (95% CrI: 0.07 – 0.49)
1
incr Rate of initial incarceration for
non-PWID 1 1,1 1 1,1 11
( ) ( )P A P A ¥
2 3,inc incr r Rate of initial incarceration for
PWID 2(1 )idu iduI I €
idu Percentage of temporary syringe
sharers who share during the
post-release transitory period
Varied from 0 to 100%
¥ 1,1 (1 )idu
incP P J . € 2,4
iduP I . * We model adult population age 20 to 70 year olds, tracking incarceration
histories of individuals over 50 years. ± Relative percentage of injections by syringe-sharers that are shared (in
prison compared to non-incarcerated) was conditioned upon the percentage of shared injections in prison being less
or equal than 100%. M = communities with moderate PWID HIV prevalence, and H = communities with high
PWID HIV prevalence.
Table S3: Posterior distribution of PWID HIV prevalence in and out of prison.
Transmission setting Temporary sharers HIV prevalence Posterior distribution
Median (95% credible interval)
Moderate PWIDs HIV
prevalence (5 – 20%)
0% sharing during transitory
period
In prison 14.7% (95%CI: 6.9 – 22.6%)
Out of prison 12.0% (95%CI: 5.7 – 19.3%)
50% sharing during
transitory period
In prison 17.5% (95%CI: 9.0 – 26.9%)
Out of prison 13.6% (95%CI: 6.9 – 19.0%)
100% sharing during
transitory period
In prison 19.7% (95%CI: 9.6 – 27.6%)
Out of prison 17.16% (95%CI: 6.8 – 19.4%)
High PWIDs HIV
prevalence (20 – 60%)
0% sharing during transitory
period
In prison 43.5% (95%CI: 25.0 – 68.1%)
Out of prison 35.0% (95%CI: 18.4 – 56.2%)
50% sharing during
transitory period
In prison 43.7% (95%CI: 24.9 – 68.7%)
Out of prison 34.2% (95%CI: 18.0 – 56.2%)
100% sharing during
transitory period
In prison 44.7% (95%CI: 25.8 – 68.2%)
Out of prison 34.3% (96%CI: 18.3 – 55.9%)
41
A14. Partial rank correlation coefficients (PRCCs) of model parameters.
Note: A parameter was considered to be important in affecting the impact of temporary shares on HIV transmission
if |PRCC|>0.4. The figure only shows parameters for which PRCC values were statistically significant (P-value <
0.05). Specifically, the proportion of PWID who share syringes out of prison, non-HIV PWID mortality rate,
proportion of PWID who has never being incarcerated, the rate of re-incarceration for PWID, proportion of PWID
who share syringes in prison and the HIV transmission per use of shared syringe were the most impactful
parameters. 0
idu , HIV transmission per use of shared syringe (if last person to use was infected);acute , Relative
infectiousness of individuals in acute phase compared to latent infection; 21 , Mean duration of acute phase
infection; 51 , Mean duration with elevated viral load during pre-AIDS phase; 81 , Mean duration of viral
suppression for individuals on ART with low CD4; 2 , Average non-HIV mortality rate for PWID; idu ,
Proportion of PWID who share syringes out of prison; idu
inc , Proportion of PWID who share syringes in prison;
iduf , Number of injections per non-incarcerated PWID annually; idu
incf , Number of injections per PWID in prison
annually; idu , Proportion of injections of syringe-sharers that are shared (non-incarcerated);
idu
inc , Ratio of
injections of syringe-sharers that are shared (in prison compared to non-incarcerated); 21 discr , Average duration of
incarceration for PWID; 31 elpr , Average duration of post-release transitory period; 2
rincr , Rate of re-incarceration
for PWID; iduI , Proportion of PWID who has never being incarcerated.
42
A15. Model Results: Effect of different incarceration-based interventions on HIV incidence after 5 years.
Note: RPI = reduction in PWIDs incarceration rates. OST = opioid substitution treatment. For RPI, initial
incarceration and re-incarceration were instantaneously reduced at the same rate. OST was introduced during
incarceration and supplemented by follow-up after release throughout the transitory period of elevated risk. Bounds
for every scenario (A-F) reflect median and range of effect estimates for the range of HIV epidemics modelled
(whiskers are upper and lower 95% confidence interval). (A-C) effectiveness of interventions under the lower-bound
assumption that 0% of ‘temporary sharers’ share syringes post-release, (D-F) the upper-bound assumption that 100%
of ‘temporary sharers’ continue sharing syringes post-release. When OST is used in combination with ART
retention, we assumed the same coverage increase from baseline for both interventions
43
Web Appendix References:
1. Institute for Criminal Policy Research. World Prison Brief: Countries. 2015.
http://www.prisonstudies.org/world-prison-brief (accessed 02 Jan 2016).
2. UNODC. Rapid Assessment of HIV Situation In Prison Settings in Ethiopia: Assessment Report. Geneva,
2013.
3. BensonOtieno U, Chepkonga M, Kibosia J, et al. Increased capacity for integrated HIV/TB services in
Kenyan prisons. 6th IAS Conference on HIV Pathogenesis and Treatment: Abstract no CDD234. Rome;
2011.
4. Kimetu S. Sexual Health and HIV Knowledge, Practice, and Prevalence among Male Inmates in Kenya.
Liverpool: Liverpool VCT Care and Treatment, 2008.
5. Mwapasa V, Chipungu G, Masiye F, Mukaka M. Prevalence and risk factors for HIV, sexually-transmited
infections, and Tuberculosis in Malawian prisons: Malawi Prison Services; UNODC, 2011.
6. Chimphambano C, Komolafe I, Muula A. Prevalence of HIV, HepBsAg and Hep C antibodies among
inmates in Chichiri prison, Blantyre, Malawi. Malawi medical journal 2007; 19(3): 107-10.
7. Hanifa Y, Telisinghe L, Fielding KL, et al. The diagnostic accuracy of urine lipoarabinomannan test for
tuberculosis screening in a South African correctional facility. PLoS ONE 2015; 10(5).
8. Telisinghe L, Fielding KL, Malden JL, et al. High tuberculosis prevalence in a South African prison: The
need for routine tuberculosis screening. PLoS ONE 2014; 9(1).
9. Skiti V, Goodman E, Bribble P, Hausler H. Screening and testing for tuberculosis and HIV in correctional
facilities in the Western Cape, South Africa. 44th World Conference on Lung Health of the International
Union Against Tuberculosis and Lung Disease Paris, France; 2013. p. 188.
10. Dlamini PM, Dlamini P, Minisi MZ, Bertrand S, Hariga F. A situational assessment on TB, HIV, syphilis,
hepatitis C and hepatitis B infections and associated risk behaviour among prisoners and prison officers in
Swaziland. 19th International AIDS Conference: Abstract no THPE564. Washington; 2012.
11. Kinaalwa G, Nabukalu F. Integrated Legal Aid in HIV Programming in Prison Rehabilitation Servicesfor
Better Access and Quality HIV Care 18th International Conference on AIDS and STIs in Africa (ICASA).
Zimbabwe; 2015. p. 93.
12. Uganda Prisons Service, UNODC. A Rapid Situation Assessment of HIV/STI/TB and Drug Abuse among
Prisoners in Uganda Prisons Service: Final report. Kenya: UNODC ROEA; 2009.
13. Tanzania Commission for AIDS. HIV Prevalence and Related Risk Factors in Prison Settings in Mainland
Tanzania: Findings from a Rapid Situational Assessment. Dar es Salaam: TAC AIDS; UNODC; Ministry
of Home Affairs, 2013.
14. Angolwisye J, Kaymobo F, Nichombe F, et al. First survey on TB and HIV prevalence in the prisons of the
Mbeya Region in Tanzania. 42nd Wrold Conference on Lung Health of the International Union Against
Tuberculosis and Lung Disease. Lille, France; 2011. p. 157.
15. Harris JB, Siyambango M, Levitan EB, et al. Derivation of a tuberculosis screening rule for sub-Saharan
African prisons. The international journal of tuberculosis and lung disease 2014; 18(7): 774-80.