METHODS

We reviewed three surveillance systems, held at the Communicable Disease Surveillance Centre (CDSC), that together provide an informative picture of HIV/AIDS among black Caribbean adults (⩾15 years) in England, Wales, and Northern Ireland between 1997 and 2001 inclusive. Ethnicity is generally self assigned, and not always directly collected; country of birth is used as a surrogate marker in some surveillance systems.

New HIV diagnoses

Voluntary confidential reports of new HIV diagnoses are received from virologists and clinicians, the latter also reporting new AIDS diagnoses.¹⁸ Ethnicity data are collected on laboratory and AIDS reports, and ethnic categorisations (based on the Office for National Statistics categories) have not changed since 1993. Clinician reporting of new HIV diagnoses was introduced in 2000, collecting information on country of birth, and date of arrival in the United Kingdom, as well as ethnicity. Data on probable route of infection are collected on all reports, through follow up when the information reported is initially incomplete. Probable country of infection is collected on all reports, and for heterosexuals, is followed up.

Misallocation of black Caribbean ethnicity probably occurs at all stages of data handling, typically resulting in reports mistakenly coded black African, and vice versa. Records were “cleaned” before analysis, using country of probable infection, country of birth, and follow up information to identify “true” black Caribbeans. In total, 34 relevant errors in ethnicity coding were corrected; 26 records of black Caribbeans were reassessed as records of black Africans and four as white or ethnicity not reported, and four black Africans reassessed as black Caribbean.

Accessing services for HIV related treatment and care

The Survey of Prevalent HIV Infections Diagnosed (SOPHID) gives a measure of the number of individuals living with diagnosed HIV infection in England, Wales, and Northern Ireland.¹⁹ This annual survey aims to collect information for each individual seen for HIV related treatment and care within the previous calendar year, including the level of antiretroviral therapy and CD4 count when last seen for care, and area of residence. Ethnicity data have been collected since the survey began in 1995.

Undiagnosed HIV

The Unlinked Anonymous Prevalence Monitoring Programme (UAPMP) measures the level of undiagnosed (that is, unrecognised) infection in different population subgroups.²⁰ Most of the surveys test for HIV in blood samples left over after completion of routine clinical tests. All specimens have patient identifying details permanently removed before testing. The UAPMP genitourinary medicine (GUM) survey uses residual blood taken for syphilis serology, and provides information on HIV prevalence among men who have sex with men (MSM) and heterosexuals attending 15 GUM clinics (seven in London, eight elsewhere in England, Wales, and Northern Ireland), as well as information on the uptake of testing and co-infection with other STIs. Ethnicity data are not collected, but country of birth is, and this is used as a surrogate for ethnicity. A UK GUM clinic is synonymous with an STD clinic in other countries.

The unlinked anonymous survey of neonatal dried blood spots provides information on HIV prevalence among pregnant women by area of residence, using the blood taken from newborn infants for routine metabolic screening to test for maternal antibodies to HIV. Electronic linkage to birth registration records in three participating regions allows the collection of additional demographic data, including mother’s country of birth. Data linkage was established in 1997 for the North Thames Region, and 2000 for the South East Thames and the North West regions. A small number of mothers born in central America are included with those born in the Caribbean.

Statistical methods

Descriptive epidemiology has been supplemented by statistical tests where appropriate. χ² Tests for trend were calculated using Epi-Info 6 (v.6.04d), and median age in years including the interquartile range (IQR) in Stata 8.

Ethics

Reports of new diagnoses and of those diagnosed and living with HIV are voluntary and confidential. To maintain patient confidentiality no names are held on the database, and surname Soundex codes are used instead.²¹ The reporting system has approval under the section 60 regulations of the Health and Social Care Act (Statutory Instrument 1438–June 2002). The ethical and legal basis for unlinked anonymous testing has been described elsewhere²² and the programme complies with recent guidelines published by the Medical Research Council.²³ All data are stored on restricted and secure databases at CDSC, with strict adherence to the Data Protection Act and Caldicott guidelines.²⁴

RESULTS

New diagnoses

Between 1997 and 2001, 528 black Caribbean adults were newly diagnosed with HIV, representing 3.3% of all reported infections in adults over the same period. A 2.8-fold increase in the numbers of newly diagnosed black Caribbeans occurred between 1997 and 2001, the largest proportional increase within an ethnic group after black Africans (table 1). In 1997 there were 62 new HIV diagnoses in black Caribbeans (2.5% of the total reported), in 2001 there were 176 new HIV diagnoses (3.7% of the total reported).

Sex between men and women was the predominant (63%) probable route of infection among black Caribbeans (table 2), compared to 50% (8089/16 170) overall. The proportion of heterosexually acquired new diagnoses among black Caribbeans increased by 18% between 1997 and 2001, an equivalent rise occurring for all ethnicities (38% (945/2482) to 58% (2750/4726)) over the same period. The proportion of black Caribbeans newly diagnosed, who had probably acquired their infection through sex between men, significantly decreased from 44% in 1997 to 29% in 2001 compared to heterosexuals (χ² = 5.56, p = 0.02), and overall accounted for 32% of newly diagnosed infections within this ethnic group.

Sex between men and women

Of the 335 black Caribbeans with heterosexually acquired HIV diagnosed between 1997 and 2001, 161 (48%) were male and 174 female. Median age at diagnosis was 35.1 (IQR: 30.1–43.9) and 32.8 (IQR: 26.6–39.9) years respectively. Where probable country of infection was reported, 97 (63%) black Caribbean male heterosexuals were probably infected in the Caribbean, including 58 in Jamaica and 37 (24%) in the United Kingdom (table 3). Of those with a known country of birth, 44 (75%) were born in the Caribbean and 14 (24%) in the United Kingdom. By comparison, 71 (44%) black Caribbean heterosexually infected women were probably infected in the Caribbean, including 51 in Jamaica, and 66 (41%) in the United Kingdom (table 3). Of those with a known country of birth, 47 (62%) were born in the Caribbean and 24 (32%) in the United Kingdom.

Men who have sex with men

The median age at HIV diagnosis of the 171 black Caribbean men probably infected through sex between men was 32.3 years (IQR: 27.4–38.0). Where probable country of infection was reported, 48 (62%) black Caribbean MSM were probably infected in the United Kingdom, 27 (35%) in the Caribbean, including 17 in Jamaica (table 3). Of those MSM for whom country of birth was known, 35 (53%) were born in the Caribbean and 27 (41%) in the United Kingdom.

HIV infected individuals accessing treatment and care

A 2.4-fold increase was observed in the numbers of diagnosed black Caribbeans accessing treatment and care services in England, Wales, and Northern Ireland between 1997 and 2001, equivalent to the increase seen in new diagnoses. In 1997, 294 (2.0%) of the 14 755 adults seen for HIV related treatment and care services were black Caribbeans, rising to 691 (2.8%) of 24 341 in 2001 (table 1). This was the greatest increase observed within an ethnic group, other than the black Africans, over the 4 year period.

From 1999 more black Caribbeans seen for treatment and care had acquired HIV heterosexually rather than through sex between men (table 2). In 2001, 372 (53%) black Caribbeans patients accessing care had heterosexual acquisition of infection reported, and 256 (36%) through sex between men (χ² = 21.26, p<0.01), representing proportional increases of 84% and 52% respectively since 1999.

Uptake of treatment

In the 2001 SOPHID survey, CD4 count and information on the level of antiretroviral therapy at last clinic attendance was reported for 496 (72%) black Caribbeans compared to 78% (18 960/24 350) in all adults. Seventy three (75%) of 97 black Caribbeans with a CD4 cell count <200 cells ×10⁶/l were receiving triple therapy or more, compared to 78% of black Africans (1022/1309) and of white people (1244/1599) (χ² = 0.42, p = 0.81). Uptake of treatment at last clinic attendance was equivalent among heterosexuals and MSM, and between black Caribbeans, black Africans, and white people (data not shown).

Area of residence

In 2001, 528 (76%) black Caribbeans diagnosed and living with HIV were London residents. South east London (boroughs of Lambeth, Lewisham, and Southwark) had the largest resident numbers (185). Outside London, the West Midlands (40) and the south east (31) had the highest numbers of resident black Caribbeans living with diagnosed HIV. Rates per 1000 black Caribbean population were highest in London (1.54 per 1000), south west, and south east (1.14 per 1000), with the West Midlands having a rate of 0.49 per 1000 population (fig 1).

Undiagnosed HIV and uptake of testing

Between 1997 and 2001, prevalence among Caribbean born male heterosexuals attending GUM clinics with previously undiagnosed HIV was 0.7% (37/5682) (χ² = 0.06 p = 0.81), and for Caribbean born female heterosexuals, 0.6% (26/4604) (χ² = 4.71 p = 0.03) (table 4). By comparison, prevalence among UK born male and female heterosexuals was 0.16% (180/112125) and 0.12% (142/120497) respectively, and for those born in sub-Saharan Africa, 2.9% (229/7832) and 4.7% (376/7920) respectively. Prevalence among Caribbean born MSM previously undiagnosed at time of clinic visit was 10.4% (20/192) (χ² = 0.29 p = 0.59). The equivalent prevalence for UK born MSM was 3.8% (793/20 906).

Between 1997 and 2001, of previously undiagnosed HIV infected Caribbean born heterosexuals, 73% of males and 50% of females left the clinic undiagnosed (table 4), compared to 61% (110/180) and 61% (87/142) of previously undiagnosed HIV infected UK born male and female heterosexuals and 47% (107/229) and 46% (173/376) of previously undiagnosed HIV infected sub-Saharan Africa born male and female heterosexuals. For Caribbean born MSM the proportion was 65%, compared to 60% (477/793) for UK born MSM.

Acceptance of voluntary confidential testing (VCT) for HIV increased in Caribbean born heterosexuals from 15% (296/1921) in 1997 to 40% (1032/2594) in 2001 (χ² = 333 p = <0.00001) and in Caribbean born MSM from 28% (11/40) in 1997 to 41% (19/46) in 2001 (χ² = 0.91 p = 0.34). Similarly, acceptance of VCT for HIV increased in UK born and sub-Saharan Africa born heterosexuals (24% (11 991/50 669) and 28% (929/3272) in 1997 to 35% (16 977/48 329) and 47% (1810/3837) in 2001 respectively), and UK born MSM (44% (1988/4485) in 1997 to 55% (2434/4451) in 2001).

Between 1997 and 2001, 41% (28/69) of Caribbean born HIV infected heterosexuals and 42% (16/38) of Caribbean born MSM were co-infected with an acute STI at the time of UA HIV testing. By comparison, 29% (121/417) of UK born HIV infected heterosexuals, 21% (182/854) of sub-Saharan Africa born HIV infected heterosexuals and 36% (558/1535) of UK born HIV infected MSM were co-infected with an acute STI at the time of testing.

HIV prevalence among Caribbean born women giving birth

HIV-1 prevalence among women born in central America/Caribbean and giving birth in the North Thames Region between 1997 and 2001 was 0.31% (13/4189), with no significant trend over the years. Ten (77%) of the HIV infected women were born in Jamaica, giving a prevalence of 0.44% (10/2254) among Jamaican born women giving birth between 1997 and 2001 in this region. HIV-1 prevalence among women born in the United Kingdom and giving birth in the North Thames Region between 1997 and 2001 was 0.03% (96/298 485), and for women born in sub-Saharan Africa, 1.51% (606/40 184).

For 2000 and 2001, HIV prevalence among central American/Caribbean born mothers was 0.27% (3/1125) in the South East Thames Region, and 0.0% (0/125) in the North West Region. For comparison, HIV prevalence among UK born women giving birth was 0.02% (15/67 372) in the South East Thames Region, and 0.01% (8/67 125) in the North West Region over the same period. Equivalent figures for mothers born in sub-Saharan Africa were 2.04% (122/5982) and 1.01% (11/1084) respectively.

DISCUSSION

Our data confirm that numbers of new diagnoses among black Caribbeans and the number of black Caribbeans HIV infected and accessing treatment and care services in England, Wales, and Northern Ireland, have increased rapidly in recent years. Both heterosexual and homosexual sex have contributed to the spread of HIV within this community, with injecting drug use not having a significant role. The increasing importance of heterosexually acquired HIV is reflected in the rising numbers of black Caribbean women newly diagnosed and the numbers accessing treatment and care services. Reasons for this increasing heterosexual contribution may include rising HIV prevalence in the Caribbean,⁶ the maturing of the HIV epidemic among black Caribbeans in England, Wales, and Northern Ireland leading to symptomatic individuals presenting with infection, and increased awareness about HIV among black Caribbeans at a local services level. The close inter-relation between HIV in the Caribbean and in England, Wales, and Northern Ireland has been clearly demonstrated. black Caribbeans newly diagnosed with HIV in England, Wales, and Northern Ireland are a mixture of UK and Caribbean born, acquiring HIV in either setting. Nevertheless, the prevalence of HIV among Caribbean born heterosexual GUM clinic attendees and mothers giving birth remains low. Prevalence among Caribbean born MSM attending GUM clinics is high, which may reflect differential patterns of inward migration to the United Kingdom, and is an area for further research.

The three surveillance systems reviewed provide an informative picture of HIV among black Caribbean communities in England, Wales, and Northern Ireland. However there are limitations. The introduction of clinician reporting of new diagnoses in 2000 will have increased the number of total reports received, and the number of black Caribbeans, with increased reports of black Caribbeans and the identification of individuals as black Caribbean who would have previously not had their ethnicity reported. However, data on the numbers of HIV infected individuals accessing treatment and care services show that HIV among black Caribbean communities is increasing, suggesting that reporting artefact has not significantly biased observations in new diagnoses among black Caribbeans. Heterosexual spread of HIV among Caribbeans within England, Wales, and Northern Ireland is also likely to be underestimated, as if exposure to HIV has occurred in more than one country, the country with highest prevalence will be assigned as the likely country of infection.

Information from new diagnoses shows that where country of birth was known a third of black Caribbeans were UK born. Such individuals will have been excluded from GUM and neonatal dried blood spot analyses (country of birth is collected not ethnicity), but may have similar risk behaviours as those born in the Caribbean, may maintain close links to the Caribbean, and be younger than those migrating. As country of birth has only been collected for new diagnoses since 2000, it is not possible to determine whether there has been a proportional increase of UK born black Caribbeans being newly diagnosed with HIV compared to Caribbean born. Prevalence estimates from the unlinked anonymous surveys may be biased, as only 15 GUM clinics are involved in the GUM survey and three regions within the neonatal dried blood spot survey have data linkage. However, of the 15 GUM clinics, seven are in London.

The epidemiology of HIV among black Caribbeans in England, Wales, and Northern Ireland has not been previously described. Concurrent infection with an STI facilitates HIV transmission, including non-ulcerative STIs such as gonorrhoea and chlamydia²⁵; however, the epidemiological impact of interactions is difficult to estimate.²⁶ The high prevalence of bacterial STIs within black Caribbean communities contrasts with the relatively low levels of HIV that exists among black Caribbean heterosexuals. These differences may be explained by sexual mixing patterns or, again, differential migration—areas for future research.

Issues that have not been addressed here include the role of travel back to Caribbean “home countries” in the acquisition of HIV among black Caribbean residents of England, Wales, and Northern Ireland and the characteristics of those self assigned as “black other,” typically including “British born black.” Numbers of new diagnoses among those of black other ethnicity remain low (table 1) and it is difficult to determine from routine reporting whether individuals are of Caribbean ancestry. Among British born second and third generation ethnic minorities a process of acculturation has been described during which there is a progressive shift of prescribed “cultural norms” of the ethnic minority group towards those of the ethnic majority.²⁷ This may result in differing HIV epidemiology, and health promotion and service requirements between those of black other ethnicity and black Caribbeans.

The national strategy for sexual health and HIV aims to reduce inequalities in sexual health, recognising that services must respond to the different needs of different populations.²⁸ As HIV spreads among the black Caribbean communities in England, Wales, and Northern Ireland, the emerging surveillance evidence must be used to ensure that prevention and services are targeted towards those who need them most, and that widespread transmission of HIV among this ethnic group is averted.

CONTRIBUTORS

SD, LP, KF, and BE conceived the idea for the paper; SD, LP, AB, LL analysed the new diagnoses, SOPHID, UAPMP GUM, and UAPMP DBS surveys respectively, with support from BE and OG; all authors were involved in interpretation of the results and drafting of the paper; SD undertook the main writing of the paper.