Reviews

Published on January 2017 | Categories: Documents | Downloads: 60 | Comments: 0 | Views: 706
of 16
Download PDF   Embed   Report

Comments

Content

A Nicoll, ON Gill

Summary: Patterns of HIV infection and disease are changing. HIV will soon enter the top five causes of death worldwide and is now believed to cause more deaths than malaria. Commonwealth countries account for around 60% of prevalent HIV infections worldwide. Around half of all global HIV transmissions are to people under 25 years of age. HIV is lowering life expectancy and reversing gains in child survival in east and central Africa. The incidence and prevalence of HIV infection have increased enormously in southern Africa recently. A more generalised pattern of heterosexual HIV transmission is emerging in parts of South and South East Asia. Injecting drug use and related HIV transmission is increasing in resource-poor countries. Countries where HIV prevalence is low but rising and there are high levels of other STIs offer particular opportunities for early intervention, as it is easier to intervene against HIV when it is entering a country than when it has become established. In countries with weak family and social networks there is inadequate care for the increasing numbers of parentless children. HIV is prejudicing tuberculosis control programme in most African countries and will do the same in Asia. AIDS mortality has fallen in industrialised countries but the prevalence of HIV infection and treatment costs are increasing. Countries that have used multisectoral approaches at an early stage and have had political support for HIV prevention early in their epidemics have been able to limit transmission. Others, intervening later, have been able to reduce transmission. Surveillance methods need to adapt to changing patterns of infection and disease.
Commun Dis Public Health 1999; 2: 85-95.

Key words: acquired immunodeficiency syndrome disease transmission epidemiology HIV infections international cooperation population surveillance preventive health services

Introduction and data sources
In the two decades since the first cases of AIDS appeared in the United States (US) and Africa over 47 million people are estimated to have been infected with HIV, nearly 14 million of whom have already died (table 1). The spread of HIV has been uneven and the epidemiology of HIV has changed substantially in the past three years. These points must be understood in order to determine priorities, and to target, evaluate, and adjust interventions. The epidemiology of HIV infection is incompletely understood, but it is better known than the epidemiology of many other infections. This paper is based on data and analyses
A Nicoll, ON Gill HIV and STD Division PHLS Communicable Disease Surveillance Centre Address for correspondence: Dr Angus Nicoll HIV and STD Divison PHLS Communicable Disease Surveillance Centre 61 Colindale Avenue London NW9 5EQ tel: 0181 200 6868, ext 4695 fax: 0181 200 7868 email: [email protected]

from many sources, but draws particularly on those presented at regional and global Monitoring of the AIDS Pandemic (MAP) seminars in Abidjan, Lima, Manila, and Veyrier, at the 12th international conference on AIDS, and in global reports from the joint United Nations Programme on AIDS (UNAIDS)1,2. It updates an earlier paper 3, placing emphasis on recent changes in the pandemic and pays particular attention to the developing world, and countries of special relevance to the United Kingdom (UK). Unless stated otherwise all prevalence and incidence statements refer to best estimates of the global position at the end of 1998 4 with country specific estimates for the end of 1997 2. Data from serological surveys 5,6 have been combined with natural history studies and modelling techniques to estimate the incidence of infection, disease (AIDS), and deaths due to HIV infection7, 8. Field surveys in developing countries have confirmed the validity of this approach, showing that the incidence of infection and premature mortality are comparable with predictions 9,10. Such approaches provide more reliable data in resource-poor countries than clinical AIDS case reporting, which is subject to incompleteness of presentation, recognition, and

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

reviews
85

The global impact of HIV infection and disease

reviews

BOX Definitions Adults – here defined as people aged 15-49 years Children – <15 years of age Core Groups – groups of people experiencing or responsible for more HIV transmission than their numbers would suggest. Pandemic – an epidemic that affects the whole world. Prevalence of HIV – the level of HIV infection - comprises pre-existing infections, including those recently acquired, but not those among people who have died. Prevalence includes people whose infection has advanced to AIDS and represents the current burden of infection on the community and the current and future burdens of disease and death it will experience. It can be expressed both as the numbers of people living with HIV (prevalent infections) and the percentage of the adult population currently living with infection. For example, Uganda has an estimated 930 000 prevalent HIV infections among adults and children, and the prevalence is 9.5% among adults. Prevention programmes – programmes intended to reduce HIV transmission, or to prevent transmission rates from rising, often through modifying behaviours associated with transmission. Sentinel surveillance – measuring levels of infection in particular groups (such as pregnant women) so as to inform changes in levels of infection in the wider population. Transmission or incidence of HIV – the number or rate of new HIV infections in the community. Young people – adolescents and young adults (adults < 25 years of age).

reporting (though AIDS reporting is a useful indicator of important trends11). A number of countries remain, however – notably in North Africa, the Middle and Far East – where knowledge of the prevalence of HIV infection is particularly sketchy (figure 1).

Global epidemiology of infection and transmission
HIV infection continued to spread in the late 1990s so that the entire world was affected by the end of 1998, with about 33 million people living with HIV infection or AIDS, and about 16 000 new infections each day (table 1). Around 90% of new infections arise in poorly resourced countries, which together account for only

10% of global gross national products. Commonwealth countries are disproportionately affected and contain around 60% of prevalent HIV infections worldwide 2. Probably about nine out of ten people living with HIV are unaware of their infection. Transmission has recently increased in countries where levels of infection were low previously (see Definitions Box). As epidemics of HIV infection fuelled largely by heterosexual transmission have developed in resource-poor countries the age at which transmission occurs has fallen: half of all global transmission is now believed to be to people under the age of 25 years12. The predominant mode of adult transmission continues

TABLE 1 Global estimates of the HIV/ AIDS epidemic – end of 1998 People newly infected with HIV in 1998 Total Adults Women Children < 15 years Total Adults Women Children < 15 years Total Adults Women Children < 15 years Total Adults Women Children < 15 years Total Adults Women Children < 15 years 5.8 million 5.2 million 2.1 million 590 000 33.4 million 32.2 million 13.8 million 1.2 million 2.5 million 2.0 million 900 000 510 000 13.9 million 10.7 million 4.7 million 3.2 million 47.3 million 42.9 million 18.5 million 4.4 million 8.2 million (end 1997)

Number of people living with HIV/AIDS

AIDS deaths in 1998

Total number of AIDS deaths since the pandemic began

Total number of HIV infections since the pandemic began

Total number of AIDS orphans* since the pandemic began
* Defined as children whose mother or both parents died as a result of AIDS when they were under 15 years of age Source: References 2 and 3

86

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

TABLE 2 Recent change in prevalence of HIV infection in adults aged 15 to 49 years in countries receiving aid from the UK Prevalent infections among adults (%) Country Bangladesh Bolivia Botswana Brazil Cambodia China Ghana India Kazakstan Kenya Malawi Nepal Nigeria Pakistan Peru Russia South Africa Tanzania Uganda Zambia Zimbabwe at end 199432 15 000 2000 125 000 550 000 90 000 10 000 172 000 1 750 000 500 1 000 000 650 000 5000 1 050 000 40 000 30 000 3 000 650 000 840 000 1 300 000 700 000 900 000 (0.03) (0.06) (18) (0.65) (1.9) (0.0015) (2.3) (0.38) (0.006) (8.3) (13.6) (0.051) (2.2)* (0.063) (0.25)* (0.004) (3.2) (6.4) (14) (17) (17) at end 1997 2 21 000 2600 190 000 570 000 120 000 400 000 200 000 4 100 000 2 500 1 600 000 670 000 25 000 2 200 000 62 000 71 000 40 000 2 800 000 1 400 000 870 000 730 000 1 400 000 (0.03) (0.07) (25) (0.63) (2.4) (0.06) (2.4) (0.82) (0.03) 11) (15) (2.4) (4.1) (0.09) (0.56) (0.05) (13) (9.4) (9.5) (19) (26) Percentage increase in prevalent infections 28 30 60 4 33 >100 16 >100 >100 60 3 >100 <100* 55 <100* >100 >100 43 -30† 4 55

* Despite these published figures it is now considered that the 1994 estimates were an underestimate and therefore that the true increase in prevalence over the period was probably less than 100% for these countries † The reduction in Uganda is likely to be real, reflecting both reduced transmission and the effects of HIV related mortality (see Impact page 93 onwards)

to be unprotected, penetrative heterosexual intercourse (that is, without effective use of a barrier contraceptive). The presence of other sexually transmitted infections (STIs) – especially those causing ulcers, which are common in most developing countries – facilitates heterosexual transmission13. In countries where infection is concentrated among certain vulnerable groups, however, transmission through unprotected anal intercourse between men and sharing of injecting equipment between injecting

drug users are of greater importance (figure 1). Most HIV infected children acquire infection from their mothers, and in resource-poor countries substantial numbers of adults and children continue to acquire infection through infected blood products, mainly transfused blood that has not been screened for antibodies to HIV 2,14. These generalisations hide substantial complexities both within and between countries. National prevalences

FIGURE 1 Current patterns of HIV transmission by country, 1997/8

Concentrated
Transmission mostly confined to behaviourally or economically vulnerable groups (homosexual men, injecting drug users, commercial sex workers, and heterosexual transmission in some deprived groups)

Generalised
Transmission widespread in the heterosexual population as well as vulnerable groups

Mixed
Within the same country both patterns existing in different regions

Uncertain
Low prevalence at present, unclear if moving to generalised transmission

Source: PHLS AIDS and STD Centre, Communicable Disease Surveillance Centre

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

reviews
87

reviews

TABLE 3 Impact of HIV – by country (more than 500 000 HIV infections, an adult (15 to 49 years) HIV prevalence of more than 1%, or where HIV prevalence more than doubled between 1994 and 1997) 2 Prevalence of HIV infection Countries with more than 500 000 infections or adult (15 - 49 years) prevalence of more than 1% 1 (k = thousand, m = million) sub-Saharan Africa West Africa Benin (2%), Burkina Faso (7.1%), Cote d’Ivoire (670k, 10.1%), Gambia (2.2%), Ghana (2.4%), Guinea (2.1%), Guinea-Bissau (2.25%), Liberia (3.6%), Mali (1.7%), Niger (1.5%), Nigeria (2.2m, 4.1%), Senegal (1.8%), Sierra Leone (3.2%), Togo (8.5%) Burundi (8.3%), Eritrea (3.1%), Ethiopia (2.5m, 9.3%), Kenya (1.6m, 11.6%), Malawi (670k, 14.9%), Mozambique (1.2m, 14%), Rwanda (12.8%), Tanzania (1.4m, 9.5%), Uganda (930k, 9.5%), Zambia (730k, 19%), Zimbabwe (1.4m, 25%) Angola (2.1%), Cameroon (4.9%), CAR (10.1%), Chad (2.8%), Congo (7.8%), Democratic Republic Congo (990k, 4.3%), Gabon (4.2%) Botswana (25%), Lesotho (8.3%), Namibia (20%), S Africa (2.9m, 13%) Incidence Prevalence doubled between 1994 and 1997

no countries

East Africa

no countries Angola South Africa

Central Africa Southern Africa South and south east Asia South Asia South east Asia Rest of Asia and the Pacific Central Asia East Asia The Pacific Australasia The Americas North America The Caribbean Latin America Europe and central Asia Northern, southern, and western Europe Eastern Europe and central Asia North Africa and the Middle East (western Asia)

India (4.1m, 0.82%) Cambodia (2.4%), Mynamar (1.8%), Thailand (2.2%)

India, Nepal Cambodia, Mynamar, Vietnam

no countries no countries no countries no countries

Kazakstan China, Indonesia, Philippines no countries no countries

USA (810k, 0.8%) Dominican Republic (1.9%), Haiti (5.2%), Jamaica (1%) Trinidad (1%) Brazil (580k, 0.63%), Honduras (1.5%)

no countries Dominican Republic Honduras, Venezuela

no countries no countries

no countries Belarus, Estonia, Russia, Ukraine Turkey

no countries

* excludes countries with less than a million population

of HIV infection (as a percentage of the adult population) vary many hundred-fold and even within sub-Saharan African countries prevalences differ by a factor of 20 (table 2)2. Countries that are heavily affected, where either more than 500 000 infected people live or where the prevalence exceeds 1%, or is rising swiftly, are listed in table 3. Another important difference between countries is the occurrence of ‘concentrated’ and ‘generalised’ patterns of transmission(figure 1). Transmission is said

to be concentrated where the infection is endemic but occurs mostly among vulnerable groups such as men who have sex with men (through unprotected anal intercourse) and injecting drug users (IDUs; through sharing injecting equipment). Examples would be the countries of western Europe, North America, Australasia, and most parts of Latin America. In a few of these counties, deprived and/or marginalised populations (homosexual men, IDUs, commercial sex

88

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

Proportional increase over 100% (27) 10% to 100% (45) 0.01% to 10% (33) no increase (31) no 1997 data (36)

Source: Adapted from UN AIDS/WHO report on the global HIV/AIDS epidemic - June 1998

workers) are experiencing substantial amounts of heterosexual transmission - for example, some black populations in the US1. The prevalence may be high (over 5%) or very high (over 50%) in affected groups but because they form a minority in the population the overall adult prevalence is generally under 1% 2 . Transmission is said to be in generalised in countries or areas within countries where the infection is spread broadly through the adult population and heterosexual transmission predominates - such as in countries of east and central Africa, where HIV infection has been endemic for over 15 years and where the prevalence in adults is well over 1%2. Countries with generalised
FIGURE 3 HIV and injecting drug use, 1997

transmission have often passed through a phase of concentrated transmission, usually attributed to heterosexual transmission among core groups (see Definitions), such as commercial sex workers, businessmen, and truck drivers. High levels of blood transfusions in developing countries also make a potent contribution. It is believed that nearly a quarter of the estimated 2.5 million blood transfusions given in Africa in 1995 had not been screened for HIV antibodies. Many studies have shown that a large proportion of the transfusions were unnecessary14. Generalised transmission is now reaching broader populations in parts of India and Brazil; both

IDU with reported HIV (96 countries) IDU without reported HIV (20) no 1997 data (36)

Source: WHO Programme on Substance Abuse (published in UN AIDS/WHO report on the global HIV/AIDS epidemic - June 1998)

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

reviews
89

FIGURE 2 Proportional increase in country HIV prevalence rates between 1994 and 1997 .

reviews

countries have areas of concentrated and of generalised transmission (figure 1). Countries such as Russia and the Newly Independent States (NIS) of the former USSR may experience the same transition because of their high incidence of other STIs15. Finally there are more than 30 countries with low levels of infection and where it is unclear whether transmission will become concentrated or generalised (figure 1), including the large populations in China, Pakistan, and Bangladesh. Countries where the prevalence of HIV infection is rising The prevalence of HIV infection has remained low for several years in some countries, while in others transmission has begun recently to rise or reach high levels, (particularly those in the third column of table 3, and shown in dark in figure 2)2. Highly populated countries with a low prevalence but rapidly increasing HIV transmission are China, India, and Russia and the NIS. Drug injecting and HIV HIV transmission caused by sharing injecting equipment and spread of HIV to and among sexual partners of IDUs have been major risk factors for infection in many industrialised countries. Injecting drug use is also of growing importance in resource-poor countries (figure 3)2. Early intervention in such situations to reduce needle sharing among IDUs may prevent explosive epidemics. In European cities the prevalences of HIV infection among IDUs have usually remained under 20% 1 . In North American cities, however, once prevalence among IDUs has risen above 10%, it has almost invariably risen to levels of 40% or more1,2,16. Children and HIV transmission Between one and two thousand infants are infected daily through mother to child transmission, equivalent to around 590 000 infections a year (table 1). These occur largely where adult transmission is generalised, so that 90% are in resource-poor countries. Infections can occur before birth but most take place during birth or as a result of breastfeeding. Several industrialised countries, notably the US and France, have reduced mother to child transmission substantially through
FIGURE 4 HIV prevalence among pregnant women: selected provinces of South Africa, 1990-1997
30 25 KwaZulu Natal Free State Gauteng Eastern Cape

routine antenatal HIV testing followed by a short course of antiretroviral therapy for mother and infant, delivery by caesarean section, and the avoidance of breastfeeding 17. Antenatal HIV screening has not become routine in the UK, even in high prevalence areas, and efforts are underway to increase testing18. Trials in Thailand have shown that shorter courses of antiretroviral therapy reduce the risk of vertical transmission19, but the practicalities, cost, and possible adverse effects of short antiretroviral regimens pose special difficulties for resource-poor countries 20.

Sub-Saharan Africa21 Six in every ten HIV infected men, eight in ten infected women and nine in ten infected children in the world live in sub-Saharan Africa, which still accounts for nearly half of all the new infections worldwide, mostly associated with generalised transmission (figure 1). Heterosexual transmission predominates, but transfusion-related transmission is a significant and readily preventable source of infection. Africa is not universally affected by HIV and a mosaic of epidemics is progressing at varying rates in different areas. In southern Africa, transmission in Botswana, South Africa, and Zimbabwe has intensified considerably in the mid to late 1990s, where the prevalence of the infection among pregnant women in some areas has reached the highest levels ever recorded: 43% in Francistown, Botswana in 1997; 59% in Beit Bridge, Zimbabwe in 1996, and 27% in Kwa Zulu Natal, South Africa in 1997 (figure 4) 2. These countries have been especially vulnerable because of the phenomenon of oscillatory migration: men leave their families for work for prolonged periods, during which they live in hostels, often having sex with commercial sex workers. Similarly, the employment of women as sex workers in neighbouring or distant countries has been a potent mechanism for spreading HIV between countries 1. In contrast, prevalences in west Africa countries have been lower and stable in the past three years (figure 2). The highest prevalence, 10%, is in Cote d’Ivoire and the lowest, 1.8%, is in Senegal, where HIV prevention has been particularly successful
FIGURE 5 HIV prevalence among pregnant women: Dakar, Senegal, 1989-1996
10 HIV 1 HIV 2

Regional summaries

8

HIV prevalence (%)

20 15 10 5 0 1990

HIV prevalence (%)

6

4

2

91

92

93

94

95

96

97

0 1989

90

91

92

93

94

95

96

Year
Source: Department of Health, South Africa (published in UN AIDS/WHO report on the global HIV/AIDS epidemic - June 1998)

Year
Source: National AIDS Programme, Senegal (published in UN AIDS/WHO report on the global HIV/AIDS epidemic - June 1998)

90

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

(figure 5)2. The situation in Nigeria, the most populous country in Africa, remains unclear. Turning to east African countries, the prevalence of HIV infection in women under 25 years of age attending urban antenatal centres in Uganda has been falling since 1994 11,22, indicating that transmission has declined substantially from far higher rates in the 1980s. In some other countries – for example, Zambia – some signs of lessening impact are being seen in new generations of young women 23. The declines in urban Uganda have followed changes in behaviour; the deferment of first sexual intercourse and greater use of condoms in casual relationships24. These trends remain exceptional, however, and elsewhere in east and central Africa the picture is less optimistic. The overall prevalence increased in Tanzania and Kenya between 1994 and 1997 (table 2), while in Malawi the prevalence of infection in rural areas has risen, to approach the high levels seen in urban areas5. Even where prevention has had some effect, situations where 6% of girls are already HIV infected by the age of 16 years (in Zambia 23) or 10% by the age of 20 (in urban Uganda 22) indicate appreciable transmission. South, south east, and east Asia, and the Pacific25,26 This part of the world illustrates the need for extensive cooperation between countries to prevent HIV transmission 2. The extensive network of social and economic links across the larger ‘golden triangle’ (from Eastern Mynamar (Burma) to Yunnan (China) and including Manipur (India) and the Mekong Delta (Cambodia and Vietnam) makes it more meaningful to consider the area as a whole than to look at transmission country by country. Another set of foci is the large metropolitan areas in western and southern India (Mumbai (the former Bombay) and Chinnai). The pattern of infection in this region has been of concentrated transmission, but some countries are now also experiencing generalised transmission. Large epidemics of infection associated with sharing injecting equipment sharing are continuing in Vietnam, Malaysia, Myanmar, and parts of India and China. Extensive transmission through heterosexual sex, either on its own or in association with drug injecting has taken place in Thailand and Cambodia, parts of India, and perhaps also China. There are indications that heterosexual transmission in Thailand and Cambodia is beginning to predominate over transmission associated with injecting drug use25-27. Use of commercial sex workers by young and older men has been a tradition in South East Asia. Sex workers and their regular clients have acted as potent core groups, with high levels of STIs, which has greatly amplified HIV transmission. This has been the case especially in Cambodia and Thailand where ‘high intensity’ sex workers (women who have many partners each day) have compounded the problem25. The prevalence of HIV infection has remained lower in Indonesia and the Philippines, even among female sex workers, possibly because their pace of work is less intense. Variation and change is also seen in HIV transmission associated with drug

injecting. For example, southern areas of Vietnam are more highly affected than the north and there is evidence of a recent resurgence of infection through injecting in Thailand 25,26. In India infection levels and transmission are known to have risen sharply in a few states recently. One third (10/31) of the states accounted for 96% of the AIDS cases and the major impact has been on Maharashtra in the west, Tamil Nadu and Pondichery in the south, and Manipur in the north east 2,3. In Manipur most transmission is associated with sharing drug injecting equipment, but elsewhere heterosexual sex accounts for most transmission. HIV was concentrated previously among sex workers and people attending clinics for STIs but in some areas infection is now moving towards a generalised pattern. In 1996 and 1997 HIV began appearing at increasing levels in pregnant women: in Maharashtra, Chinnai, and Manipur (where very few women inject drugs) a prevalence of 1% to 2% was observed2,3. Trends in HIV transmission remain unclear in China, although extensive public health surveillance has begun1. Transmission associated with injecting drug use is common in Yunnan and, although there are also unconfirmed reports of more generalised transmission (Liao SS and Detels R, personal communication June 1998), reported levels of risky sexual behaviour are lower than those reported elsewhere28,29. The economic downturn of 1997/8 in South East Asia may exacerbate the situation, perhaps by driving more young women into the sex industry 1. Latin America and the Caribbean 30 Sex between men is a far more important route of infection in Latin America than in Africa or Asia. HIV transmission has stayed more concentrated than generalised, and most new infections are occurring among people who live on the social and economic margins of society. The prevalence of HIV infection in pregnant women is low overall, but there have been some recent isolated increases - for example, in Honduras (to 1%) and Porto Alegre, Brazil (to 3%-4%)1,2,5. Brazil is something of an exception as generalised spread is taking place in some areas and prevalences in pregnant women in sum urban areas have risen to over 1%2. In 1996 nearly 60% of AIDS cases were attributed to heterosexual transmission of HIV, which seems to be moving into younger age groups and into small towns and rural areas. Only 708 Brazilian municipalities reported AIDS cases in 1987, compared with 2585 in 1997. In the Andean sub-region (Bolivia, Columbia, Peru, and Venezuela) and the ‘southern cone’ (Argentina and Chile) sex between men remains the main mode of transmission1. Heterosexual transmission is common throughout the Caribbean and there are signs that some islands could be approaching generalised transmission. This is already the case in Haiti, where the prevalence is remarkably high (8% in pregnant women)5. In Jamaica and Trinidad the overall prevalence in adults is approaching 1%2.

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

reviews
91

Russia, the Newly Independent States, central Asia, and eastern Europe Intensive screening of populations at low and high risk in these countries found very small numbers of HIV infections until 199531. Since 1995, however, HIV has spread swiftly among IDUs in cities in the Ukraine, Belarus, Moldova, and the Russian Federation. HIV is now emerging in the Caucasus, the Baltic States, and Kazakstan. Estimates of numbers of HIV infections in Russia and the NIS have risen from less than 3000 in 1994 to over 190 000 in 19972,31,32. There is considerable potential for further increases as there are large groups of impoverished IDUs – for example, in the Moscow area – to which HIV has yet to spread, and job-related migration may make such spread inevitable (V Pokrovsky, personal communication, June 1998). Drug use practices are extremely risky. Equipment used to manufacture opiates is often contaminated with HIV or other bloodborne viruses (hepatitis B and C), human blood is added to drug solutions as a cleansing agent, and ready-made drugs are sold in used syringes. Transmission through sex between men is also occurring but is underreported because of official intolerance of homosexuality. Little is known about the numbers of sex workers and the levels of HIV infection among them, but the numbers of women involved may have increased due to economic insecurity (V Pokrovsky, personal communication, June 1998)1. Women from this region have migrated as sex workers to western Europe and the Middle East, and even to China and India1. There is concern that concomitant epidemics of syphilis among populations at risk will lead to extensive heterosexual transmission15. So far there is no clear evidence that transmission is becoming generalised, as the prevalence in blood donors and pregnant women is low, and any increases seen can be explained by infections in IDUs of both sexes. HIV transmission appeared among IDUs in Poland and Yugoslavia in the late 1980s. Unlike the NIS there has been little evidence in these countries that transmission has intensified in the 1990s and the incidence of AIDS remains low. The prevalence of HIV infection in IDUs has stayed low in Slovenia, Slovakia, and the Czech Republic, but rates of around 10% have been reported in Poland6. There is only limited evidence of westward spread of the Russian syphilis epidemic15, but cases of syphilis associated with transmission in Russia have presented in the UK33. North America, western Europe, and Australasia Transmission is almost exclusively a concentrated phenomenon in these regions: most HIV infections are acquired through sex between men or in association with injecting drug use (figure 1). In a few countries (for example, the UK and Belgium) many infections are among immigrant populations, mostly acquired abroad in countries where the prevalence is higher. Between 400 000 and 650 000 people in the US are HIV infected, of whom probably two thirds are aware of the fact1. Similar proportions have been suggested for Canada and western European countries 1,34 . In the US great

reviews

differences exist between racial groups, with prevalences ten or 20 times higher among blacks than whites in some urban areas. In the same poor inner city areas HIV transmission has become generalised, with prevalence rising above 1% among pregnant women1. HIV transmission in Canada recently increased from 2500 to 3000 new infections per year from 1989 to 1994 to around 4200 in 1996, an annual rate of 26 per 100 000 adults (compared with 8/100 000 in England and Wales in the late 1990s)34,35. The rise in Canada has been attributed to epidemics among IDUs34 . In Europe transmission through sex between men has on the whole predominated in northern countries while in south western Europe (Portugal, Spain, and Italy) most infections have been associated with drug injecting1,35a. The incidence of AIDS and HIV related mortality has fallen in North America, western Europe, and Australasia since the introduction of highly active antiretroviral therapy (HAART)4. The improvement in survival is increasing the prevalence of HIV infection, especially that of people receiving treatment. HAART is expensive, so the total costs of HIV care in these countries have increased enormously and will probably continue to do so34,36,37.

New variants of HIV
Three new groups of recombinant HIV-1 viruses have emerged recently in parts of Russia, China, and Nigeria38. These are apparently sudden reassortments of previously recognised subtypes, rather than a gradual evolution of HIV-1. The biological and public health significance of these recombinants is yet to be assessed, but their emergence and spread requires close monitoring.

Effective HIV prevention and prevention opportunities
Several countries have been successful in either averting substantial HIV transmission (such as Senegal, the UK, the Netherlands, and the Nordic countries) or of reducing previously substantial HIV transmission (such as Switzerland, Thailand, and to some extent Uganda). The common factor linking these countries is centrally supported, multisectoral HIV prevention programmes, with political support39,40. Per capita spending on HIV prevention has been lower in the US and Canada than in the UK and Switzerland, which are having greater success at containing HIV 41. Following sustained publicity and much public awareness of HIV, however, the US has recently seen improvements in sexual behaviour in its youth42. As the costs of treatment rise the cost-benefit ratio of primary prevention in well-resourced countries becomes more attractive, making effective sexual health promotion and needle exchanges an even better investment than before. It is considered far easier and more cost-effective to intervene against HIV epidemics at their start before transmission becomes intense or generalised 43 . Evidence for this has come from STI intervention trials

92

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

Average life-expectancy at birth, in years

65 60 55 50 45 40 35 0 1955 Botswana Zimbabwe Zambia Uganda Malawi

Thailand and Uganda, although successful in reducing transmission, are left with a massive legacy of prevalent infections that will burden their care services for many years11,27,47.

60

65

70

75

80

85

90

95

2000

Year
Source: United Nations Population Division, 1996 (published in UN AIDS/WHO report on the global HIV/AIDS epidemic - June 1998)

in east Africa, which have been effective where the prevalence of HIV infection is rising, but less so where infection is already widespread in the population 44,45. Theoretical work has also suggested that early interventions should be more successful at preventing HIV transmission than if the same resource is applied later to reduce an already high HIV incidence 43 . Several countries where the prevalence of HIV infection is currently low but rapidly rising and where the incidence of other STIs is high – including Bangladesh, China, India, Nepal, Pakistan, and Russia – therefore present particular opportunities for prevention 46. There are other countries where the prevalence of HIV is already moderate or high but where infection is still spreading into new populations and where preventive measures would be a particularly good investment. These include Botswana, Brazil, Cambodia, Kenya, South Africa, and Zimbabwe (table 3). HIV prevention targeted at new cohorts of young people remains a good investment in any country where HIV transmission is substantial. Even in countries that have reduced HIV transmission still face substantial problems caring for existing patients.
FIGURE 7 Increase in mortality in men aged 15-60 years, based on household reports (sibling histories): selected African countries, 1986 to 1997 Probability of dying from all causes (%)
70 60
1994 1995/96

50 40 30

1993 1991

Mortality Nearly 14 million people are thought to have died from HIV infection since the pandemic began in the late 1970s and it is estimated that in 1998 about 2.5 million people died because of HIV infection (table 1). In 1997 HIV was one of the top ten causes of deaths worldwide and it is likely soon to become one of the top five. In 1998, for the first time, deaths from HIV exceeded those from malaria2. The extensive early spread of HIV in east and central Africa means that the experience there can be taken to show what will happen in other resource-poor countries if infection reaches high levels (5% to 10% or more). Population based surveys have shown that death rates among adults aged 15 to 60 years have doubled in Malawi, Tanzania, Uganda, Zambia, and Zimbabwe (figure 6)10. This will progressively reduce life expectancy in subSaharan Africa (figure 7). HIV is also increasing child mortality, undoing improvements in child survival seen in the past 20 years. One study in Uganda found that HIV was reducing average life expectancy by 16 years 48. These effects will next be seen in the next 10 years in southern Africa. The effect on national and regional development will be especially potent because age specific mortality rates associated with HIV are highest among adults at ages when they are at their most productive and when they are responsible for the greatest numbers of elderly and child dependants1,2. South, South East, and East Asia; the Pacific; eastern Europe; and Central Asia are only starting to feel the impact of HIV disease and mortality. Even before the economic downturn of 1997, secondary care facilities in Thailand where extensive spread had arrived earliest were beginning to find it difficult to cope with HIV associated morbidity 49. Cambodia presents a more typical Asian picture, where in late 1997 there were an estimated 120 000 prevalent HIV infections but a cumulative AIDS total of less than 10 000. Irrespective of current and future HIV prevention this picture will change dramatically in the next decade as HIV associated mortality and morbidity inevitably increase50. Orphans The greatest impact of HIV related mortality is among young adults, often soon after they have become parents. By December 1997 around 8 million children, more than 90% of them in sub-Saharan Africa, had already lost their mothers prematurely because of HIV (table 1)1. The ability of societies to support orphans varies. In the advanced epidemics of east Africa many communities retain family networks that can sustain many children. Such support is less the common in

Current and future impact

1995 1986 1990 1988

1989

1990/91

20 10 0

Zimbabwe

Tanzania

Malawi

Uganda

Zambia

Source: Timaeus I, London School of Hygiene and Tropical Medicine, from Demographic and Health Survey data (published in UN AIDS/WHO report on the global HIV/AIDS epidemic - June 1998)

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

reviews
93

FIGURE 6 Projected changes in life expectancy: selected African countries with high HIV prevalence, 1955-2000

reviews

urban societies or societies with traditions of migrant labour. It is predicted that where nuclear families are the norm (for example, in Thailand), or where prolonged wars have prejudiced family networks (for example, in Cambodia and Angola), societies will be less able to care for their orphans1. HIV and tuberculosis Fifty per cent of young adults are already infected with Mycobacterium tuberculosis and an estimated 15.3 million people were infected both with HIV and M. tuberculosis by the end of 1997 (nearly 80% of them in sub-Saharan Africa) 1. Most HIV negative people infected with M. tuberculosis will not develop disease. HIV infection is the strongest risk factor for progression to active tuberculosis: in areas with a high prevalence of HIV infection, the incidence of tuberculosis is also rising 1. People dually infected also have tuberculosis recurrence rates after treatment that are far higher than for people without HIV infection. When the prevalence of HIV infection rises above 10% in the adult population numbers of tuberculosis cases double over what was previously expected. Tuberculosis control programmes in some African countries are close to being overwhelmed. The clinical presentation of tuberculosis is less straightforward in people with HIV infection, challenging tuberculosis control programmes with difficulties in diagnosis as well as by increasing case numbers 51. In 1996 an estimated 7.4 million people developed tuberculosis, more than 80% of whom were in 22 countries, with India and China together contributing 43%. Clearly if HIV transmission becomes generalised in Asia there will be a major impact on tuberculosis incidence.

behaviours, and to monitor key associated diseases (notably tuberculosis) and mortality among young adults 52 . WHO and UNAIDS have to update previously published surveillance guides53,54. The objectives of HIV surveillance need clarifying: there is as yet no systematic global surveillance for HIV subtypes. Much development work remains to be done at international, regional, and national levels.

References
1. Monitoring the AIDS Pandemic (MAP). The status and trends of the HIV/AIDS epidemics in the world. Geneva: MAP, June 1998. 2. Joint United Nations Programme on AIDS (UNAIDS) and the World Health Organization (WHO). Report on the global HIV/AIDS epidemic June 1998. Geneva: UNAIDS & WHO, 1998. 3. HIV and STD Division, PHLS Communicable Disease Surveillance Centre. Health and population occasional paper: sexual health and health Care: HIV, AIDS and sexually transmitted infections, global epidemiology, impact and prevention. London: Overseas Development Administration, 1996. 4. UNAIDS and WHO. Report on the global HIV/AIDS epidemic – update December 1998. Geneva: UNAIDS & WHO, 1998. 5. United States Bureau of the Census. HIV/AIDS surveillance database, July 1998 release. Washington DC: US Bureau of the Census, 1998. 6. European Centre for the Epidemiological Monitoring of AIDS. European HIV prevalence database (December 1997 update). St Maurice: CESES, December 1998. 7. Mulder DW, Nunn AJ, Kamali A, Nakiyingi, Wagner H-V, Kengeya-Keyondo J. Two year HIV-1 associated mortality in a Uganda rural population. Lancet 1994; 343: 1021-3. 8. Chin J, Lwanga SK. Estimation and projection of adult AIDS cases: a simple epidemiological model. Bull World Health Organ 1991; 69: 399-406. 9. Wawer MJ, Serwadda D, Gray RH, Sewankambo N, Li N, Nalugada F, et al. Trends in HIV-1 prevalence may not reflect trends in incidence in mature epidemics: data from the Rakai population-based cohort, Uganda. AIDS 1997; 11: 1023-30. 10. Timaeus IM. Impact of the HIV epidemic on mortality in subSaharan Africa: evidence from national surveys and censuses. In: Carael M, Schwartlander B (editors.) Demographic impact of AIDS. AIDS 1997; 12 (suppl 1): S15-27. 11. Stoneburner R, Low-Beer D, Tembo G, Mertens T, AsiimusOkiror G. Human immunodeficiency virus dynamics in East Africa deduced from surveillance data. Am J Epidemiol 1996; 144: 682-95. 12. UNAIDS and WHO. Report on the global HIV/AIDS epidemic. Geneva: UNAIDS & WHO, December 1997. 13. Cohen MS. Sexually transmitted diseases enhance HIV transmission: no longer a hypothesis. Lancet 1998; 351 (suppl III): S5-7. 14. World Health Organization. Preventing HIV transmission in health facilities. Geneva: WHO, 1995. 15. Tichonova L, Borisenko K, Ward H, Meheus A, Gromyko A, Renton A. Epidemics of syphilis in the Russian Federation: trends, origins and priorities for control. Lancet 1997; 350: 210-3. 16. Des Jarlais DC, Friedman SR. HIV epidemiology among injecting drug users. Int J STD AIDS 1997; 7 (suppl 2): 57-61. 17. Madelbrot L, Le Chenadec J, Berrebi A, Bongain A, Benifla J, Del Fraissy K-L, et al. Perinatal HIV-1 transmission: interaction between zlodovudine prophylaxsis and made of delivery in the French perinatal cohort. JAMA 1998; 280: 55-60. 18. Nicoll A. Antenatal screening for HIV in the UK: what is to be done? J Med Screen 1998; 5: 170-1. 19. Mojenson L. Short-course zidovudine for prevention of perinatal infection. Lancet 1999; 353: 766-7. 20. WHO. Recommendations on the safe and effective use of

Future surveillance (‘second generation HIV surveillance’)
Numbers of AIDS cases in many resource-poor countries are swamping reporting systems while the value of AIDS case reporting in industrialised countries is diminishing because of the effects of the use of potent antiretrovirals. In response a number of counties are developing or enhancing methods for HIV surveillance – collecting data on the distribution of HIV infection in the population, describing the characteristics of those who probably became infected recently, and studying the behaviour and sociobiological factors associated with current HIV transmission. This requires substantial improvement in existing national and international surveillance systems. Industrialised countries are adopting the reporting of diagnosed HIV infections, if this is not already in place, and relying less on AIDS case reporting alone. For developing countries UNAIDS and WHO are taking a lead in developing ‘second generation HIV surveillance’. They are being encouraged to develop or strengthen unlinked sentinel surveillance of HIV infection, to adopt methods for surveillance of risky

94

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

short-course ZDV for prevention of mother-to-child transmission of HIV. Wkly Epidemiol Rec 1998; 73: 313-20. 21. MAP. The status and trends of the HIV/AIDS epidemics in subSaharan Africa. Abidjan, Cote d’Ivoire: MAP Network Workshop, 3-4 December 1997. 22. Tarantola D, Schwartlander B. HIV/AIDS epidemics in subSaharan Africa: dynamism, diversity and discrete declines? AIDS 1997; 11 (suppl B): S5-21. 23. Fylkesnes K, Musonda RM, Kasumba K, Ndhlovu Z, Mluanda F, Kaetano L, et al. The HIV epidemic in Zambia: socio-demographic prevalence patters and indications of trends among childbearing women. AIDS 1997; 11: 339-45. 24. Musinguzi J, Asiimwe-Okiror G, Opio AA. Sexual behaviour change due to HIV/AIDS results of population based KABP surveys conducted in five districts in Uganda. Working paper for UNAIDS Best Practice Workshop, Nairobi, February 1997. 25. Dore GJ, Brown T, Tarantola D, Kaldor JM. HIV and AIDS in the Asia-Pacific region: an epidemiological overview. AIDS 1998; 12 (suppl B): S1-10. 26. Mills S, Ungchusak K, Srinivasan V, Utomo B, Bennett A. Assessing trends in HIV risk behaviours in Asia. AIDS 1998; 12 (suppl B): S79-86. 27. Siriwasin W, Shaffer N, Roongpisuthipong A, Bhiraleus P, Chinayon P, Wasi C, et al. HIV prevalence, risk and partner serodiscordance among pregnant women in Bangkok. JAMA 1998; 280: 49-54. 28. Liu H, Xie J, Yu W, Song W, Gao Z, May Z, et al. A study of sexual behaviour among rural residents of China. J Acquir Immune Defic Syndr 1998; 19: 80-8. 29. Cleland J, Ferry B (editors). Sexual behaviour and AIDS in the developing world. London: WHO and Taylor & Francis, 1995. 30. MAP. The status and trends of the HIV/AID/STD epidemics in Latin America and the Caribbean. Rio de Janeiro: MAP Network, 1997. 31. Gromyko A. Epidemiological trends of HIV/AIDS and other sexually transmitted diseases in eastern Europe. Copenhagen: WHO Europe, 1997. 32. WHO.Working estimates of adult HIV seroprevalence as of end 1994. Wkly Epidemiol Rec 1995; 70: 356-7. 33. Ratcliffe L, Nicoll A, Carrington D, Wong H, Eggleston SI, Lightfoot NF, et al. Reference laboratory surveillance of syphilis in England and Wales 1994, to 1996. Commun Dis Public Health 1998; 1: 14-21. 34. Anon. Estimates of HIV prevalence and incidence in Canada. Ottawa: Bureau of HIV/AIDS and STD, Laboratory Centre for Disease Control, November 1998. 35. Report of a Working Group (Chairman: Professor NE Day) convened by the Director of the Public Health Laboratory Service on behalf of the Chief Medical Officers. The incidence and prevalence of AIDS and the prevalence of other severe HIV disease in England and Wales for 1995 to 1999: projections using data to the end of 1994. Commun Dis Rep CDR Rev 1996: 6: R1-24. 35a.Hamers FF, Downs AM, Infuso A, Brunet JB. Diversity of the HIV/AIDS epidemic in Europe. AIDS 1998; 12 (suppl A): S63-70. 36. Anon. National Centre in HIV Epidemiology annual report. Darlinghurst, Australia: National Centre in HIV Epidemiology, 1998. 37. Department of Health. HIV/AIDS strategy (report of a conference). London: DH, 1998. 38. Goudschmidt J. The significance of virus sub-types for epidemiology and pathogenesis. 12th World AIDS conference; Geneva, June-July 1998 (presentation No. 486). 39. Malcolm A, Dowseth G (editors). Partners in prevention: international case studies of effective health promotion practice in HIV/AIDS. Geneva: UNAIDS, 1998. 40. Tarantola A, et al. Amounts, patterns and trends of national and international financing of the response to HIV/AIDS in developing countries. 12th world AIDS conference; Geneva, June - July 1998 (abs 44232). 41. Albert T, Williams G. The economic burden of HIV/AIDS in Canada (CPRN Study No. H02). Toronto: Canadian Policy Research Networks Inc, 1998. 42. CDC. Trends in sexual risk behaviours among high school students – United States, 1991-7. MMWR Morb Mortal Wkly Rep 1998; 47: 749-52. 43. Robinson NJ, Mulder DW, Auvert B, Hayes RJ, Grosskurth H. Modelling the impact of alternative HIV intervention strategies in rural Uganda. AIDS 1995; 9: 1263-70. 44. Grosskurth H, Mosha F, Todd J, Mwijarubi E, Klokke A, Senkoro K, et al. Impact of improved treatment of sexually transmitted diseases on infection in rural Tanzania: randomised controlled trial. Lancet 1995; 346: 530-6. 45. Wawer MJ, Sewankambo NK, Serwadda, Quinn TC, Paxton LA, Kiwanuka N, et al. Control of sexually transmitted diseases for AIDS prevention in Uganda: a randomised community trial. Lancet 1999; 252: 525-35. 46. Gerbase A, Rowley J. Heymann D. Global STD estimates. Sex Transm Dis 1998; 74 (suppl): S12-4. 47. Anon. The HIV/AIDS collaboration annual report 1997/8. Nonothaburi, Thailand: The HIV/AIDS Collaboration, Thailand, 1998. 48. Boerma JT, Nunn AJ, Whitworth JAG. Mortality impact of the AIDS epidemic: evidence from community studies in less developed countries. In Carael M, Schwartlander B (editors.) Demographic impact of AIDS. AIDS 1997; 12 (suppl 1): S3-14. 49. Surasiengsunk S, Kiranandana S, Wongboonsin K. Demographic impact of the HIV epidemic in Thailand. AIDS 1998; 12: 775-84. 50. WHO. STD/HIV/AIDS surveillance report. No 10. Manila: Western Pacific Region, October 1997. 51. Chintu C, Muringa A. An African perspective on the threat of tuberculosis and HIV/AIDS – can despair be turned to hope? Lancet 1999; 353: 997. 52. Schwartlander B. Broadening the tools for public health surveillance 12th world AIDS conference; Geneva, June July 1998 (presentation No. 193). 53. World Health Organization Global Programme on AIDS. Unlinked anonymous screening for the public health surveillance of HIV infection. Proposed international guidelines. Geneva: WHO, 1989. 54. Chin J. Public health surveillance of AIDS and HIV infections. Bull World Health Organ 1990; 68: 529-36.

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

reviews
95

reviews

Weekly Returns Service of the Royal College of General Practitioners
DM Fleming
Summary: General practitioners in 69 practices in England and Wales monitor the spread of epidemic diseases in the community through the Weekly Returns Service (WRS) of the Royal College of General Practitioners, which has existed for over 30 years. Participating general practitioners summarise diagnoses and consultation/ episode type (new episodes/ongoing consultations) for a defined population (currently about 570 000) and data are extracted to provide the ‘weekly return’, which includes age specific weekly incidence of new episodes of selected illnesses. The service has been used extensively to measure the burden of influenza and total acute respiratory illness in the community and the impact of enteric infections. It also provides information about illnesses for which there are no other major data sources – for example, chickenpox, scabies, and (historically) mumps. The entire network is electronically linked. Direct links with microbiological laboratories are being forged in order to integrate clinical and microbiological data in defined populations.
Commun Dis Public Health 1999; 2: 96-100.

Key words: communicable diseases episode of care family practice morbidity sentinel surveillance

Introduction
Common infectious illnesses usually have a limited impact on individuals but because many people contract them their impact on the population is considerable. On average less than 1% of the population consult their general practitioners for flu-like illnesses in winter influenza epidemic periods1, but seroconversion rates in populations suggest that over 10% of the population may be infected2. Almost everyone suffers influenza at some time but people do not equate the experience with a risk of death, even though about 25 000 excess deaths in England and Wales in the winters of 1989/903 and 1996/97 were attributable to influenza4. About 25 deaths are attributed to chickenpox each year5, yet very few children escape infection before the age of 10 years6. This report describes the Weekly Returns Service (WRS) of the Royal College of General Practitioners (RCGP) and discusses its contribution (actual and potential) to monitoring common illnesses. The service was set up to serve a sentinel function and to warn particularly of the emergence of epidemic illnesses. The report shows how recording methods have improved
DM Fleming, Birmingham Research Unit Royal College of General Practitioners Address for correspondence: Douglas Fleming Royal College of General Practitioners Birmingham Research Unit Lordswood House 54 Lordswood Road Harborne Birmingham B17 9DB tel: 0121 426 1125

over the years and describes how the data are used and where they can be found.

History of the Weekly Returns Service
The WRS evolved from the Epidemic Observation Unit of the College of General Practitioners that was set up in 1953 7 . This unit provided a forum for general practitioners interested in the epidemiology of common infectious diseases to share experiences about the spread and impact of conditions diagnosed and treated in their practices. The Records Unit of the college (now the Birmingham Research Unit) set up consistent methods of recording and of data capture in order to monitor and study disease7. At that time, the data were collected using diagnostic indexes8 and relevant details were extracted each week to provide the ‘weekly return’. About 20 diagnoses (classified in the College Classification of Disease, a truncated version of the International Classification of Disease (7th Revision)) were covered. The data were sent by post and aggregated at the college’s Birmingham unit. In 1976 additional diseases and health problems were included in response to specific requests – for example, asthma, duodenal ulcer, and acute myocardial infarction were added either because of their seasonality or because associations with infection were considered likely. The incidence of new episodes of duodenal ulcer has declined (figure 1) since Helicobacter pylori has been identified as a causal factor and has been treated specifically. Since the mid 1980s the Birmingham unit responsible for the WRS has published an annual report of weekly incidence data. Previously, reports had been made for the Council of the RCGP and for the Department of Health but these had not included summaries of the data.

96

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

From 1976 to 1996 selected weekly data were published routinely in the OPCS Monitor, and the data have been used in reviews of specific diseases published by the PHLS. Weekly data are not currently published routinely, but data that relate to influenza and the interpretation of epidemic conditions are published by the PHLS Communicable Disease Surveillance Centre. Since 1994 a quarterly newsletter has been sent to participating practices, which helps them to avoid recording problems, draws attention to topical issues, and clarifies the use of diagnostic rubrics.

5 Male 4 Female

Incidence

3

2

1

How the WRS works
Reports from the WRS are based on new episodes of illness, including ‘first ever diagnoses’. A new episode occurs when a patient presents with what is clearly a new diagnosis or for advice and treatment for an exacerbation of pre-existing illness, but not when a patient consults simply for management to be reviewed or to obtain a repeat prescription. The identification of individual episodes is particularly important for conditions such as asthma, which are influenced by seasonal and environmental factors. Computerisation in general practice in the late 1980s enabled data to be collected about a wider range of illnesses. The discipline of data entry increased as practices began to record assessment diagnoses and the types of consultation/episode from every consultation. In 1991, several of the WRS practices contributed to the Fourth National Morbidity Survey in General Practice 9, but not all had appropriate computer software. Some practices recruited for the survey joined the WRS. The network currently includes 69 practices that care for 570 000 people, and some growth is expected as more practices become able to communicate electronically. Data extracts are taken from the practice each Tuesday night/Wednesday morning and procedures exist to follow up missing data. Practice computerised information systems provide tabular data on numbers of patients consulting with diagnostic Read codes and information about the practice population, by age groups and sex. Read codes are mapped to the International Classification of Disease (9th revision) for analysis. A preliminary analysis of data covering the seven days ending on the previous Sunday for a population of
FIGURE 2
45 40 35 30

0 1976

78

80

82

84

86

88

90

92

94

96

98

Year

450 000 patients is usually available by noon on Wednesday. Data are thus reported in the week after the patient first consulted, and allocated to diagnoses based on initial clinical impressions. The intention is to provide information at the earliest possible opportunity. Notifications from other sources are often delayed until diagnoses are confirmed by laboratories and by clerical procedures in clinical settings and departments of public health, postal delays, and time taken to check and analyse the data.

The need for information from primary care
Many illnesses, including several of importance to public health, are treated almost exclusively in primary care for example, common childhood infections, enteric infections, asthma, and epidemic respiratory diseases. Information about disease incidence in the community is therefore vital. Microbiological investigation is undertaken rarely for most disorders; thus diagnosis in primary care is essentially clinical. Statutory notifications of some clinical diagnoses places the onus to notify on the attending physician. Rates of conditions treated in primary care ascertained through the WRS are generally higher than those ascertained through statutory notifications. The difference is particularly noticeable for enteric infections, for which the estimated incidence based on the WRS population greatly exceeds that based on microbiological reports, which in turn exceeds notifications of food poisoning10, defined as ‘any

Mumps – mean weekly incidence (per 100 000, all ages) in four week periods: 1967 to 1997

MMR vaccine introduced

Incidence

25 20 15 10 5 0 1967 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97

Year

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

reviews
97

FIGURE 1 Duodenal ulcer – mean weekly incidence (per 100 000, all ages) in males and females: 1976 to 1997

reviews

FIGURE 3
35 30 25

Chickenpox – mean weekly incidence (per 100 000, all ages) in four week periods: 1970 to 1997

Incidence

20 15 10 5 0 1967 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97

Year

disease of an infectious or toxic nature caused by or thought to be caused by consumption of food or water’11. Notifications based exclusively on microbiological investigation underestimate the problem because organisms are not necessarily recoverable from stool specimens examined and because many patients consult and provide specimens relatively late in the course of the illness, or indeed provide no specimen, or do not consult. For the purpose of statutory notification doctors are required to notify on suspicion of diagnosis12, but notification is often delayed until a microbiological diagnosis has been made. The timeliness of accurate information is particularly important in the surveillance of influenza. Influenza activity increases every winter and clinical surveillance aims to identify its onset and monitor its impact. Virological surveillance complements clinical surveillance by identifying and characterising the causative organism, and in this we have collaborated with the PHLS Central Public Health Laboratory 13. Isolates identified in this collaborative project are often the first to be typed and characterised for antigenic profile. During the winter season of 1996/97, the first viruses were identified by sentinel surveillance schemes in seven out of the eight European countries in which they operate, as opposed to other sources14. Sentinel practice data can be used to evaluate the need for a vaccine and to monitor its effects. Analysis of WRS data for mumps in 1976 revealed a three yearly

cycle of mumps activity and patient records were examined to evaluate complications15. The database has since been used to monitor the impact of vaccination16 (figure 2). Each bar of the histogram represents the mean weekly incidence (all ages) of mumps reported in a four week period. The combined measles, mumps, and rubella vaccine was introduced in 1988 at a trough in the natural cycle of incidence of mumps, yet the impact of the vaccination programme was evident within six months. For chickenpox the WRS also provides important data17 on which to judge both the need for a vaccination programme and to monitor its impact should it be introduced. Incidence varies year by year, usually peaking in late spring or early summer (figure 3). About a fifth of cases are aged 15 years and over. Monitoring illness in the community identifies unrecognised or potentially important health problems that have not been adequately explained. The incidence of acute bronchitis in children aged 0 to 4 years during the four winters 1993/94 to 1996/97 varied in magnitude but peaked consistently in week 51 (figure 4), matching the pattern of isolations of respiratory syncytial virus reported to the PHLS. Acute bronchitis is also a problem in elderly people, however, in whom the incidence peaks consistently around the first week of the New Year (figure 5), when the death rate (from all causes) commonly, though not invariably, reaches its highest level regardless of ambient temperature18. The causes of acute bronchitis

FIGURE 4 Acute bronchitis – incidence (per 100 000 children aged 0 to 4 years) by week and year compared with reports of respiratory syncytial virus infection: weeks 45-08 1993 to 1997
1600 1400 1200 1000 800 600 400 200 0 45 48 51 2 5 8 45 48 51 2 5 8 45 48 51 2 5 8 45 48 51 2 5 8

1993/94
Isolates Incidence

1994/95

1995/96

1996/97

Week

98

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

Incidence 900 800 700 600 500 400 300 200 100 0

1993/94
Deaths Incidence

1994/95

1995/96

1996/97

Deaths 25000

20000

15000

10000

5000

0 45 48 51 2 5 8 45 48 51 2 5 8 45 48 51 2 5 8 45 48 51 2 5 8

Week

in elderly people at this time of year are not known. Finally there are illnesses for which there are unlikely to be any other sources of data – for example, hand, foot and mouth disease19 and scabies20.

Diagnostic validity
The validity of diagnosis is important in the interpretation of morbidity data from whatever source. Common illnesses are diagnosed in general practice within the confines of a consultation of five to ten minutes during an illness that may last only a week. Many diagnoses are made without supportive evidence from laboratory investigations because the additional benefits yielded by a diagnostic test in individual cases often cannot justify the inconvenience for a patient, the time involved in taking a specimen and reporting the result back to the patient, and the laboratory costs. Within a sentinel network, the concern is more for the validity of diagnosis at a population rather than an individual level. One approach to validation involves comparing data from differing sources. Figure 6 contrasts the variation in incidence of new episodes of asthma reported in the sentinel network in the past ten years in children aged 5 to 14 years with hospital admissions for the period 1990 to 1994. It shows synchronous timing in the peaks and troughs, indicating that the two data sources are recognising the same
FIGURE 6 Asthma in children aged 5 to 14 years – (percentage variation from average incidence) – new episodes reported by the WRS and hospital episodes for England: 1990 to 1994
120 100 80 wrs hes

phenomenon. Of additional interest are the actual numbers of peaks and troughs throughout the year and the marked difference in the relative sizes of the peaks in September (weeks 35-39). The validity of diagnosis in the surveillance population can also be demonstrated using data from supplementary investigations in a sample of the practices, as illustrated for influenza in figure 7. This comparison is particularly telling because of the difficulty of distinguishing influenza clinically from other common viral infections. The link between epidemics of flu-like illness and virologically confirmed influenza has been identified in several European surveillance networks21. Validation usually relies on the existence of a diagnostic gold standard and these rarely exist. In the early days of combined clinical and virological surveillance for influenza we were content to achieve a virus positivity rate of 10% to 15% in specimens submitted during epidemic periods. Positivity rates exceeding 30% are now common and when investigation by polymerase chain reaction is included even higher rates have been achieved. Improvements in the process of investigation (better specimens, sealed tubes, efficient delivery, improved laboratory techniques) bring as much to explaining the clinical syndromes as does greater care in clinical assessment. Isolating virus in swabs from
FIGURE 7 Influenza-like illness – incidence (per 100 000, all ages) and virus isolates of influenza in the WRS: winter of 1996/97

250 Flu A 200 Flu B Incidence

30 25

% difference

40 20 0 -20 -40 -60 1 4 8 12 16 20 24 28 32 36 40 44 48 52

150 15 100 10 50 5 0
38 40 42 44 46 48 50 52 2 4 6 8 10 12 14 16 18 20

0

Week

Week

COMMUNICABLE DISEASE AND PUBLIC HEALTH

VOL 2 NO 2 JUNE 1999

Virus isolates

60

20

Incidence

reviews
99

FIGURE 5 Acute bronchitis – incidence (per 100 000, age ≥ 65 years) by week and year compared with total deaths (all ages): weeks 45-08 1993 to 1997

reviews

elderly people is more difficult than in swabs from children, yet influenza vaccination policy is justified largely by the excess deaths among elderly people during influenza epidemics that are presumed to be due to influenza virus infection3,4. Delayed consultation by elderly people may account for the difficulty of isolating this virus but, whatever the explanation, it highlights the difficulty of defining a gold standard for diagnosis. Seroconversion might be considered a more secure standard but this implies exposure to an organism rather than illness caused by it. Some sentinel practice networks stipulate criteria for diagnosis, but the WRS has not adopted a rigid position. If criteria are to be useful they need to be specific, based on objective evidence rather than reported symptoms, and applicable to patients of all ages and at all stages of the disease. Some European surveillance networks use a raised temperature as a criterion for diagnosing influenza. At the very least this implies that a doctor cannot diagnose influenza without measuring the patient’s temperature, and that the diagnosis cannot be made in a patient once fever has abated. We have recently confirmed the diagnosis of pertussis in 49 out of 150 patients with protracted cough, on the basis of seroconversion (E Miller, personal communication). The clinical features usually associated with an attack of whooping cough, if rigidly applied, would not have enabled us to identify these cases. The children with pertussis tended to have a milder illness than was seen before immunisation was widespread. The elderly people diagnosed would not have been recognised at all. The diagnoses were made because we did not use strict criteria and because of improvements in the microbiological tests available.

results through the internet and through PHLS publications. WRS data already appear in quarterly summaries of infectious diseases published as CDR Supplements. Direct interaction with our practices to obtain additional information will help us to respond to specific requests for information.

Acknowledgements
Many thanks to the practices that participate in the WRS for their enormous contribution, and to Dr AM Ross for comments on this manuscript. The project is supported by the Department of Health, which has given permission to publish.

References
1. 2. Fleming DM, Zambon M, Bartelds A, de Jong JC. Duration and magnitude of influenza epidemics. Eur J Epidemiol (in press). Monto AS, Sullivan KM. Acute respiratory illness in the community. Frequency of illness and the agents involved. Epidemiol Infect 1993; 110: 145-60. Curwen M, Dunnell K, Ashley J. Hidden influenza deaths. BMJ 1990; 300: 896. Dedman DJ, Joseph CA, Zambon M, Fleming DM, Watson JM. Influenza surveillance in England and Wales: October 1996 to June 1997. Commun Dis Rep CDR Rev 1997; 7: R212-9. Mortality statistics: cause, Series DH2 (1979-1996). London: Office for National Statistics, 1979-1996. Preblud SR, Orenstein WA, Bart KJ. Varicella: clinical manifestations, epidemiology and impact in children. Pediatr Infect Dis 1984; 3: 505-9. College of General Practitioners Annual Report 1954. (available from the Royal College of General Practitioners, London) Research Unit of the Royal College of General Practitioners. Diagnostic Index. J Roy Coll Gen Pract 1971; 21: 609. McCormick A, Fleming DM, Charlton J (editors). Morbidity statistics from general practice; fourth national study 1991-92. Series MB5 no.3. London: HMSO 1995. McCormick A. The notification of infectious diseases in England and Wales. Commun Dis Rep CDR Rev 1993; 3: R19-25. Djuretic T, Ryan MJ, Fleming DM, Wall PG. Infectious intestinal disease in elderly people. Commun Dis Rep CDR Rev 1996; 6: R107-12. Chief Medical Officer. Definition of food poisoning. London: Department of Health, 1992 (PL/CMO(92)4). Fleming DM, Chakraverty P, Sadler C, Litton P. Combined clinical and virological surveillance of influenza in winters of 1992 and 1993-4. BMJ 1995; 311: 290-1. Zambon M. Sentinel surveillance of influenza in Europe, 1997/ 1998. Eurosurveillance 1998; 3: 29-31. Research Unit of the Royal College of General Practitioners. The incidence and complications of mumps. J Roy Coll Gen Pract 1974; 24: 545-51. Jones AGH, White JM, Begg NT. The impact of MMR vaccine on mumps infection in England and Wales. Commun Dis Rep CDR Rev 1991; 1: R93-6. Joseph CA, Noah ND. Epidemiology of chickenpox in England and Wales, 1967-85. BMJ 1988; 296: 673-6. Fleming DM, Cross KW, Crombie DL, Lancashire RJ. Respiratory illness and mortality in England and Wales. Eur J Epidemiol 1993; 9: 571-6. Bendig JWA, Fleming DM. Epidemiological, virological, and clinical features of an epidemic of hand, foot, and mouth disease in England and Wales. Commun Dis Rep CDR Rev 1996; 6: R81-6. Barrett NJ, Morse DL. The resurgence of scabies. Commun Dis Rep CDR Rev 1993; 3: R32-3. Fleming DM, Cohen J-M. Experience of European collaboration in influenza surveillance in the winter of 19931994. J Public Health Med 1996; 18: 133-42. Ross AM, Fleming DM. Incidence of allergic rhinitis in general practice 1982-92. BMJ 1994; 8: 161-76.

3. 4.

5. 6.

7. 8. 9.

10. 11.

Quality assurance
Some degree of quality assurance is provided by comparing information derived from differing sources, as was shown above for asthma and influenza. In 1992 a regional comparison of the incidence of allergic rhinitis during the pollen season (hay fever) demonstrated internal consistency 22. We have seen that the annual reporting rate for herpes zoster is remarkably consistent. Timely reporting is an essential ingredient of an effective sentinel network and this itself is a measure of quality. It is also relevant that even at critical times of the year (for example at Christmas and New Year) the weekly report has been based on a patient population of 500 000 despite staff leave and conflicting priorities at such times.

12. 13.

14. 15.

16.

17. 18.

Latest and future developments
During 1998, the data collection system in practices was completely automated and we began to gather information from all consultations and not just new episodes. This allows us to report on consultation frequencies and opens opportunities to report prevalence rates based on patients consulting in a year. Methods of linking routine microbiological data from practices with clinical incidence material are being developed in collaboration with Professor T Elliott of the University Hospital in Birmingham. We plan to disseminate the

19.

20. 21.

22.

100

VOL 2 NO 2 JUNE 1999

COMMUNICABLE DISEASE AND PUBLIC HEALTH

Sponsor Documents

Or use your account on DocShare.tips

Hide

Forgot your password?

Or register your new account on DocShare.tips

Hide

Lost your password? Please enter your email address. You will receive a link to create a new password.

Back to log-in

Close