Key points are not available for this paper at this time.
Introduction HIV prevalence in some fishing communities in low and middle-income countries is known to be high relative to national average seroprevalence rates 1,2. Most of the studies supporting this claim refer to the men involved in fish-catching operations (fishermen). However, they acknowledge that the men and women who work in associated occupations such as fish trading and processing are also vulnerable, in part because they are often within the fishermen's sexual networks 3–5. This vulnerability stems from the nature and dynamics of the fish trade and fishing lifestyle in which a number of known or hypothesized ‘risk factors’ converge. Most people involved in fishing as an occupation are within the age groups (15 to 35 years) most vulnerable to sexually transmitted infections (STIs). Many fishing people are mobile or migratory, so the social structures that constrain sexual behaviour in home communities may not apply in the context of fishing camps or ports. Fishing is a high-risk occupation which can contribute to a culture of risk denial or risk confrontation, extending to displays of bravado and risk-taking in the social and sexual arena. Fishing people are often socially marginalized and have low status, which can cause, among men, exaggerated or ‘oppositional’ forms of masculinity that challenge norms of behaviour adopted by those in ‘mainstream’ society. Masculinity in this context often includes the expectation of multiple sexual partners. Alcohol use is widespread among fisherfolk in many parts of the world, to help cope with the dangers or stresses of their occupation. This further compounds vulnerability to HIV. In addition, fisherfolk are vulnerable to HIV and AIDS due to inadequate prevention, treatment and mitigation measures and limited access to sexual health services more generally. These points are summarized from a more detailed explanation and bibliography pertaining to HIV-related risk factors among fisherfolk, given in two of our earlier publications 2,6. The occurrence of these risk factors should not be taken as characteristic of all fishing communities but it appears, through numerous government, donor agency and non-governmental organization reports as well as recent academic studies, to be widespread enough to cause concern that fishermen, female fish traders and processors and their sexual partners are at significant risk of HIV infection. This is particularly so in the Asian, African and Latin American countries with established or emerging epidemics, where over 95% of the world's 200 million fisherfolk (men and women fish catchers, processors and traders) live 3. Thus, there is mounting evidence that fisherfolk and people living in fishing communities comprise a significant sub-population at risk. Despite this, the policy response to their vulnerability has, to date, been limited and fragmented. There is, for example, no clear targeting of fishing communities in prevention, care and mitigation guidelines issued by WHO and UNAIDS. Key national and international fisheries policies do not mention HIV and AIDS 3,7. It was only in March 2005 that the International Labour Organisation's Governing Body was asked to consider a draft FAO/ILO/IMO ‘Fishing Vessel Safety Code and Voluntary Guidelines’ recommending that ‘those who are HIV positive, but not disabled, are not excluded from opportunities to work’ 8. There is evidence, however, that the previous neglect of HIV in the fisheries sector is changing. Recent initiatives, like those supported by the Food and Agriculture Organisation of the United Nations (FAO) and the WorldFish Center, have started to raise awareness of HIV in the sector 3. Urgent action is still required to increase awareness and improve the response. A first step needed was to review epidemiological evidence of the threat posed by the epidemic to fisherfolk and fishery-based livelihoods. Estimates of seroprevalence and numbers likely to be infected are critical to informing policy and planning responses to the AIDS epidemic 9. Population-size estimates alert policy-makers to the existence and magnitude of a sub-population that may be at risk to HIV. They can guide the level and types of funding allocation towards appropriate prevention, treatment and mitigation measures, and monitoring of the effectiveness of interventions designed to reduce infection 10. Sex workers, injecting drug users (IDUs), men who have sex with men (MSM), and people in prison are known to be at higher risk of infection, due to non-use of condoms, participation in or coercion into sex acts that have high viral transmission probabilities, and the re-use of injecting equipment. Others at higher risk are individuals in social and occupational groups in which economic and socio-cultural factors combine to foster behaviours that carry an enhanced risk of HIV infection and transmission. Such factors include frequent visits to sex-workers, impaired decision-making due to alcohol or drug abuse, masculine norms of behaviour that include expectation of multiple partners, non-use of condoms with at-risk partners, non-treatment of sexually transmitted infections and needle sharing among workmates 11. With the exception of sex workers, the literature on HIV/AIDS usually identifies occupation-based groups at higher risk as being male-dominated, including long-distance truck drivers, military personnel and miners or, in the case of women and men, broader groups such as traders and migrant workers 12. However there is no comparative analysis of the number of people with HIV in each of these groups. The objective of this paper is to present estimates of the prevalence levels and numbers of people infected with HIV among fisherfolk and compare these estimates to those for other sub-populations considered at higher risk of HIV infection, to provide a basis for assessing the case for including fisherfolk among the groups meriting targeted interventions. We reviewed the literature on HIV in fishermen (men involved in fish catching operations), fisherfolk (any persons involved in fishing and fish trading and processing) and people living in fishing communities (any person resident in a port, village or fish landing station where fishing is a prominent occupation). We also reviewed the literature on HIV prevalence in other occupational groups commonly stated as being at higher risk: sex workers, long-distance truck drivers, military personnel, miners and seafarers. Data for miners and seafarers were scarce and confined to one or two countries, so results are not reported here. The category of ‘migrant worker’ was too generic to be useful. For comparison we also attempted to locate, for countries where we had estimates for fisherfolk, the prevalence and group size for the (non-occupational) sub-populations known to be at highest risk; namely IDUs, MSM and people in prison. We also included estimates of seroprevalence in the general sexually active adult (age 15 to 49 years) population as a baseline for comparison. Methods We searched electronic library databases, including Medline and Embase. Search terms used for fishing community literature included HIV, Human Immunodeficiency Virus, AIDS, Acquired Immunodeficiency Syndrome, seroprevalance, fishing, fishery, fisheries, fisherman, fishermen, fisherfolk and fishers, with additional searches in French and Spanish. Relevant references cited in these articles were checked. The same procedure was carried out for the other groups at risk, substituting fishing-related search words with the appropriate key words. The websites of relevant organizations such as UNAIDS, WHO, FAO, UN Office on Drugs and Crime, and the International Centre for Prison Studies were used, as were keyword searches in Google and Scholar Google search engines. The CD-ROM “HIV/AIDS Surveillance Data Base”, produced by the US Census Bureau was used to obtain data on prevalence and population size for prisoners, IDUs and military personnel. Care was taken to find the most up-to-date and reliable information on the population sizes of the different groups. The number of HIV-infected people in each sub-population was calculated by taking the number of people in the group and multiplying by an HIV prevalence estimate based on the most up-to-date seroprevalence data available. Given the paucity of data, inclusion criteria for the review could not be limited to studies with random sampling and high response rates so all published or reported surveys were included. Where available, the upper and lower HIV prevalence estimates were used to establish an average prevalence estimate. Confidence intervals on sub-population group size estimates were seldom available, so the values are taken as point estimates. Seroprevalence surveys as well as surveillance data reported in UN, WHO or government documents, journals or international conference presentations were included. HIV prevalence estimates from nationally-representative surveillance (hereafter referred to as ‘national estimates’) were preferentially used over surveillance estimates limited to a single location or region (‘regional estimates’), unless the latter were more than 3 years more up-to-date. If regional data were available for more than one region within a country, we provided an estimate using the lower and upper prevalence measures from the different regions. The final selection of 10 countries reported here was based on the availability of seroprevalence data on fishermen, fisherfolk or fishing communities. In all the Asian country studies, the prevalence data were collected from fishermen, rather than fisherfolk more generally. In Africa and Honduras, the data are from surveys in fishing communities and are therefore assumed to apply to all adults living in communities where fishing is a major occupation. Data on HIV prevalence and group size for the other sub-populations at risk (truck drivers, IDUs, etc) were available for a wider set of countries. However as the focus of this paper is on fisherfolk, we did not extend the analysis to calculate number of HIV infections among these groups for all countries for which prevalence rates were available. Although defining the boundaries of ‘risk group’ membership is always problematic, it is particularly so for some of the groups that are most frequently mentioned as at highest risk for HIV. For this reason we did not attempt to estimate the size of sex worker populations nor the number infected, because estimates of this group's size are often based on the numbers of women involved in brothel-based sex work which may be a minority, particularly where brothels are illegal. For non-brothel-based sex work, it can be difficult to define the boundaries between a sex worker and someone who engages in ‘transactional sex’ – the exchange of sexual services for goods or opportunities instead of or in addition to a monetary exchange. Transactional sex is a strategy adopted by some poor and vulnerable groups, usually women, needing access to food, shelter or work and other opportunities and so is not a distinct occupational or sub-population group. This ambiguity leads to highly uncertain estimates of the number of persons involved and there is often a very large within-group difference in prevalence associated with different forms of sex work 12. We therefore confined the review to prevalence rate estimates for sex workers from HIV surveillance surveys reported by WHO and UNAIDS 13, and did not review primary literature nor did we attempt to estimate respective population sizes or numbers infected. Among the non-occupational-based sub-populations at risk we were able to obtain prevalence rates and estimated group sizes for IDUs and prisoners, but not for MSM for the countries included. In many African and Asian societies the illegality and strong social stigma attached to homosexuality makes these estimates unavailable. Although drug use is also illegal, the social stigma is often less and it proved easier to estimate the numbers and prevalence for IDUs in some of the included countries. Thus IDUs and prisoners represent our known non-occupational sub-populations commonly assumed to have the highest HIV prevalence rates. All of the above groups are likely to overlap. The sources of the data used in the estimates are presented in the following sub-sections. General population Prevalence estimates and risk population sizes are for adults (age 15–49 years) and based on UNAIDS data for 2003 14. Fishermen, fisherfolk and fishing communities The prevalence estimate is from a single regional estimate for Democratic Republic of Congo (DRC) 15, Honduras 16, Indonesia 17, Myanmar 18, Thailand 19 and Uganda 20. The prevalence estimate is from two or more regional rates for Brazil 21, Cambodia 18,22 and Kenya 23. The Malaysian rate is a national estimate based on a source stating that 7.8% of all persons with a known occupation living with HIV or AIDS are fishermen 24. Prevalence was then estimated by taking the number of fishermen with HIV as percentage of total numbers of fishermen. The above prevalence estimates are for fishermen only, with the exception of Brazil (fishermen and hunters), DRC partners of fishermen visiting antenatal clinics (ANC), Honduras (study carried out in a coastal community, with fishing as an important employment), Kenya (ANC clinic surveillance in fishing communities; and population-based serosurvey in fishing villages), and Uganda (population-based survey of lake-shore community, with high employment as fishermen). All data are from 1998 to the present. All fisherfolk population size estimates are for 2002 and are based on data from the UN Food and Agricultural Organization 25. These population size estimates usually include only those employed in the fish-catching sector (fishers, mostly fishermen) and so, in cases when seroprevalence data are for a wider group (fisherfolk or fishing communities) our estimates of numbers infected are for fishermen only, based on the assumption that HIV prevalence rates of fishermen are similar to those reported for the fishing community as a whole. Truck drivers The prevalence estimates are from a national rate for Thailand 26 and Uganda 27. Prevalence estimates are from a single regional rate for Brazil 28, Honduras 29 and Indonesia 17 and from two regional estimates for Myanmar 30 and Kenya 31,32. Data from Thailand date back to 1992, data from Myanmar are from 1993, Kenyan data are from 1993–1997, data from Brazil are from 1995/1996 and Honduran data are from 1998. Indonesian data are from 2000. The number of Kenyan truck drivers is derived from the number of trucks and lorries in Kenya in 2003 33. The Ugandan truck driver population size was extrapolated from the total Ugandan population, using the same proportion of Kenyan truck drivers to the total Kenyan population. For other countries (Brazil, Honduras, Indonesia, Myanmar, Thailand) the number of truck drivers was estimated by taking 5% of the total persons employed in the Major Occupational Division 7 transport, storage and communication (TSC) of the ‘International Standard Industrial Classification of all Economic Activities’, using LABORSTA Internet, an online database for labour statistics, maintained by the International Labour Organisation: www.ilo.org (accessed May 2005). The 5% figure is a conservative proportion, derived from data in Indonesia which was the only country we found to have figures for both the number of trucks and the number of persons employed in the TSC sector. Military The prevalence estimates are from national rates for Brazil 34, Honduras 35, Indonesia 36, Myanmar 37 and Thailand 34. The Ugandan prevalence estimate is from a single regional rate 38. The Cambodian prevalence estimate is from rates from 11 provinces 39. Prevalence for Brazil, Indonesia, Thailand and Uganda are based on rates for military personnel, and for Honduras and Myanmar rates are for military recruits. Data from Indonesia are from 1992–1993, Ugandan data from 1995–1996 and Cambodian and Honduran data are from 1996 or 1997. Data from other countries are more recent. All at-risk population estimates are from the International Institute of Strategic Studies 40. Prisoners The prevalence estimate is from a national rate for Brazil 41, Indonesia 42, Malaysia 43, Myanmar 44 and Uganda 45. The prevalence estimate is from a single regional rate for Cambodia 46, Honduras 47 and Thailand 48. The prevalence data from Myanmar are from 1991 and Cambodian data are from 1993. All other studies are from or more recent. All risk population size estimates are from with the exception of Indonesia drug users The prevalence estimate is from a national rate for Brazil Indonesia 42, Kenya Myanmar and Malaysia and Thailand from Myanmar and Cambodia are from 1993. All other studies are from or more recent. population sizes are derived from studies in Brazil Indonesia 42, Thailand Malaysia Myanmar and Kenya The number of injecting drug users in Kenya is an estimate of all users so this is likely to be an Sex workers surveillance data were used to estimate prevalence rates 13, for Uganda where these data on were not available. Data from Uganda are for sex workers in based on the of the Uganda HIV/AIDS the low or middle-income countries in and Latin for which data were available, the estimated number of persons infected in each group and reported of seroprevalence for these groups are in In all cases Brazil, HIV prevalence rates in fishermen or fishing communities is to higher than the national average prevalence rate for adults 15 to 49 with the exception of Brazil for which includes in the as well as HIV prevalence rates among fisherfolk are higher than for both truck drivers and the military in all countries where comparative data are available and are higher than or to the seroprevalence rate for prisoners in of the countries for which data were available, and those for sex workers in two of the number of persons infected with HIV by group at risk in the or middle-income countries for which HIV prevalence data was available for the African countries of Kenya and Uganda the numbers of fishermen likely to be infected with HIV are higher than in of the other groups This is also the case in the Asian countries of Indonesia and Myanmar data reduce the of the Cambodia and in Thailand it was only by the number of HIV-infected In Honduras the number of people infected with HIV was also the highest among all the groups at risk; however, in Brazil, by the prevalence rates among fishermen with were so Brazil a large fishing population, the estimated number of infections among were lower than for other In the epidemiological data the many of and social factors that are to many fisherfolk at high risk to HIV HIV and numbers of infected people were in many cases estimated to be higher among fishermen or of fishing communities than in other sub-populations that are more usually considered to be at higher in the HIV, health and policy the exception of sex workers for numbers could not be estimates that prevalence survey data can be to the fishing which may be in terms of and Prevalence figures for fishermen and fishing communities presented in the are mostly based on regional estimates. were not or were and so could be or The estimates are a of those pertaining to fishermen only, and those to people in fishing communities all of may be that is, involved in fishing-related are also relevant as different countries are at different of their the estimates are a cause for The results presented here have for the in which the number of HIV infections is In countries with large but epidemics, including some of the most Asian and Latin American countries, national HIV infection to UNAIDS are by prevalence in sub-populations at risk, the sizes of these sub-populations we have in this and the estimates to obtain the number of infected persons The of estimated numbers infected with HIV on all relevant sub-populations at higher risk and including in the estimates for fisherfolk may to of the size of the epidemic in countries such as Indonesia and all of which have large fishing populations that have been to have high HIV Indonesia with and are the countries with the highest number of fisherfolk in the they for over of the world's million fishermen a similar proportion of the world's 200 million between The HIV prevalence among fishermen and in fishing communities in the latter countries are not fishermen in have been as being at high risk and fishing communities in the of have been as being at very high risk and have been targeted with HIV For and Latin it is that a epidemic be by targeting and treatment at known groups at higher risk such as sex workers and IDUs There is a that too on these groups sub-populations at high risk, such as Such mobile populations may be a for the of the HIV to the general population. We that a epidemic is to be in these then the of to be to include groups like fisherfolk Data action to HIV in the fisheries awareness of the of the is 3. Fishing communities in less countries are often mobile and socially marginalized and have poor access to and and low of available health services HIV awareness and have seldom been targeted at fishing in to other groups of policy neglect of HIV in the fisheries sector is likely to have health and Africa and towards the The sector the of over 200 million people and is an important to and and in many low and middle-income countries 3. responses include and appropriate of and treatment and on HIV transmission within fishing and between and other is also This was by the Center, Africa was supported by a to the which is by the for International and by the UN Food and Agriculture Organisation in with African countries.
Kissling et al. (Mon,) studied this question.