¿ìè¶ÌÊÓƵ

WITHIN a year, we should know a good deal more than we do now about
sexual behaviour. About 60 surveys are now under way around the world, as
scientists and health educators attempt to gather data that will help them
to predict the spread of HIV and target their prevention campaigns.

These studies have sometimes thrown epidemiology into the political
spotlight. In Britain, a study was temporarily halted when the government
reversed a decision to fund it, and the interest of the popular press was
aroused. ¿ìè¶ÌÊÓƵs involved in these surveys can be very wary of talking
publicly about their work.

Many of the surveys are in Africa, Asia and Latin America as well as
in the West. The latest country to set up a study is France. At a conference
in Chamonix at the end of April, epidemiologists working for the French
medical research council, INSERM, announced details of a plan to survey
20,000 people about their sexual practices and attitudes (This Week, 5 May).
The team hopes to have initial results next year.

Most epidemiologists agree that the need for information on sexual behaviour
– and heterosexual behaviour in particular – has never been more urgent.
The World Health Organization estimates that 6.5 million people worldwide
are infected with HIV, most of them through heterosexual sex.

And, while the number of heterosexuals infected with HIV in the West
is still relatively low, it is now rising sharply: in Europe, the virus
is spreading more rapidly through sex between heterosexuals than among gay
men or intravenous drug users, according to the WHO.

To predict the course of the epidemic, researchers need certain basic
information that they lack at present. At the very least, they need to know
what the average number of new partners per year would be for a particular
group – for argument’s sake, heterosexual women younger than 20.

Once epidemiologists know this figure and two other variables – how
long a person remains infectious, on average, and how likely that person
is to infect another on any one sexual encounter – they can, in theory,
calculate the rate at which HIV infection would reproduce itself in that
population (see ‘Plotting the spread of AIDS’, ¿ìè¶ÌÊÓƵ, 26 March 1987).

Roy Anderson, an epidemiologist at Imperial College, London, has produced
many mathematical models of the epidemic, with radically different outcomes.
These models serve mainly to show us how little we know about human behaviour.
How frequently do teenagers change partners? Do people who use condoms use
them every time they have sex? And what is the average age of men and of
women when they first have sex? The answers to these and other questions
are long overdue, according to Anne Johnson, an epidemiologist at the University
College and Middlesex Medical School in London. It is surprising, she says,
how little we know about an activity as universal as sex.

Of the existing information, the most widely quoted is that of Alfred
Kinsey, who gathered data on the sexual behaviour of Americans in the late
1940s. Although the conclusions Kinsey reached have become almost a part
of Western folklore, the study was not based on modern sampling methods,
says Johnson.

‘These were mostly white middle-class males, and they were volunteers,
which is clearly not representative,’ she says. Other, smaller studies were
carried out in Britain during the 1960s and 1970s, but none has looked at
the population in general.

We do not know, for example, how such behaviour has altered in the wake
of half a century of social changes – including the advent of international
air travel, mass migration to cities, wars on a grand scale and contraception.

‘Epidemics tend to occur largely because of social and behavioural change,’
says Johnson. She is one of the team conducting Britain’s National Study
of Sexual Attitudes and Lifestyles, a survey of 20,000 British people, whose
first results should be available within about two years.

The study attracted widespread attention late last year when its funding
was refused by the Department of Health, reputedly because the Prime Minister,
Margaret Thatcher, vetoed it (This Week, 23 September 1989). Johnson and
her colleagues – at St Mary’s Hospital, Imperial College and Social and
Community Planning Research, a survey agency – were rapidly offered funding
instead by the Wellcome Trust, and the research has continued.

Apart from the British and French studies, Norway, Sweden and the Netherlands
are completing surveys. Other data on the behaviour of subgroups of the
population, such as teenagers, have been collected in Australia and Canada.

In the US, researchers led by Wendy Baldwin at the National Institute
of Allergy and Infectious Diseases had planned a major national study, but
Congress voted last year to withdraw the funds allocated to it. The plans
are now under review, but pilot studies for a further national survey have
now been carried out.

Despite political opposition in Britain and the US, most governments
approached so far have been willing to authorise such surveys, says Amir
Mehryar. He is consultant to the Social and Behavioural Research Unit of
the WHO’s Global Programme on AIDS, which is coordinating and funding research
by local scientists in various developing countries. Some of the results
are already in, but they have yet to be analysed and compared, Mehryar says.
The preliminary data should be available ‘in a matter of weeks or months’.

In Africa, about a dozen countries already have surveys under way, he
says. They include Uganda, Tanzania, Mauritius, Lesotho, Nigeria, the Central
African Republic and the Ivory Coast. Another batch of studies is in progress
in South America.

In Asia, one study in Singapore is completed, another is in progress
in Thailand, and some more general social surveys in Bangladesh and Sri
Lanka have provided limited information about sexual behaviour. Fewer Asian
countries have so far participated than elsewhere – partly, says Mehryar,
because the religious constraints of their cultures have made it difficult
for researchers to ask certain sensitive questions.

Not that sensitive questions are easy to ask anywhere, say the researchers.
In fact, well- designed studies are crucial if the data are to mean anything.
‘We are still learning the best way to ask questions,’ says Johnson.

The problems are numerous. For example, how do you adjust for the fact
that people will tend to find it difficult to give accurate reports about
events in their past? Men also tend to exaggerate about how many partners
they have. ‘But the lies will be roughly the same from one question naire
to another,’ says Antoine Messiah, an epidemiologist from INSERM in Paris.

Literacy can also be a problem: in most of the questionnaires, there
are private sections that the person completes themselves rather than answering
an interviewer. But where most of the survey population will be illiterate,
the sensitive questions will have to be asked face to face, and this may
affect the findings.

Another difficulty is how to select the random sample in the first place.
This is no problem in countries where addresses are listed – in Britain,
for example, the 20,000 people were selected from 20,000 households in the
Post Office Address File, the usual source for social surveys. But in many
poorer countries, there is no such source. The WHO has worked out a method
for obtaining a random sample without an address list.

Johnson says the team has learnt more about how to ask the question,
what terminology to use, how to define ‘partner’ and so on. For example,
she says, people find it easier to tick boxes than to say out loud what
they do or think. Lettered cards can help, too, with one for each answer,
so that the person can simply say a letter instead of having to say embarrassing
words. Also, the researchers found that people prefer to use clinical words,
such as ‘penis’ and ‘vagina’, rather than the vernacular.

The London team began work in 1986, and spent two years refining the
design of the study, with two pilot projects. The WHO, meanwhile, has agreed
on a specific framework for the countries participating in surveys it is
funding, called ‘Partner relations and risk of HIV infection’. There are
also guidelines for smaller, specific studies. The aim is to collect data
that can be compared internationally.

The partner relations studies have two designs. The first, Phase 1,
was drawn up in 1988, a year before field research began, by researchers
from 17 countries representing East, Southern, West and Central Africa.
The teams met in Addis Ababa and Dakar to agree the design. Phase 1 is intended
mainly for African and Asian countries where relatively few people are literate,
says Mehryar. Phase 2, which followed shortly after, is intended for use
in South America, North America and Europe.

Some European countries are now using Phase 2, including France and
Greece. Britain’s national study is not using the WHO’s design, partly because
the London team had begun work some time before it was finalised. But Johnson
says the British questionnaire has many similarities with the WHO’s Phase
2.

Cost clearly limits the size of the surveys. The British study will
cost around Pounds sterling 900,000. The French team refused to say how
much its work is costing. Most of the WHO-funded surveys are much smaller
than these two, with sample sizes between 1000 and 3000. Western countries
that use the WHO protocol are expected to fund their studies themselves,
says Mehryar.

The French team says its pilot surveys revealed that the overwhelming
majority of people were willing to participate. ‘We were surprised to find
that people were asking for this,’ says Alfred Spira, an epidemiologist
from INSERM, who is coordinating the study. He says that those who refused
tended to come from higher socioeconomic groups.

Johnson and her colleagues, who have refined their methodology more
closely, found that the response rate – the proportion of those asked who
were willing to participate – was very similar to any achieved by other,
less sensitive, social surveys in Britain. The pilot surveys have also proved
representative of all ages and social classes.

The questionnaires are structured, with separate ones for men and women.
In most surveys, half of the interview is face-to-face with a trained researcher,
but the most sensitive questions, about specific sexual practices and relationship
patterns, can be answered by filling in a form.

The WHO questionnaires begin with basic information about age, gender,
social class, education and family. They then ask fairly general questions
about sexual behaviour, for example, how many partners, age at first regular
partnership, how many children.

Next comes a section on contraception, including a list of statements
about condoms. The respondent is invited to say whether these statements
are true or false. For example: ‘Condoms can prevent venereal disease if
used properly’, and ‘Condoms can climb up into the womb or stomach’.

The questionnaire then covers details about the person’s first sexual
experience: was this with the opposite sex, or the same sex; how old were
you; how long had you known the other person. It asks about sexual activity
during the past year, the past month, about experience of sexually transmitted
diseases and knowledge of AIDS. Finally, the WHO questionnaire has optional
sections on specific sexual practices, drinking, intravenous drug use and
injecting practice.

Mehryar says that in some countries there will be follow-up studies
on a much smaller scale, asking people questions in depth in a way that
allows for more open-ended answers than are possible in large scale surveys.
The aim is to check the validity of the data from the main survey, and obtain
information that might not emerge in a structured questionnaire.

The success of the surveys worldwide depends entirely on the local teams
of researchers coordinating them, says Mehryar. ‘We rely on their experience,’
he says. ‘They will be credited for any success we may have.’ Often the
questionnaire will have to be translated into the local language.

The coordinators will have to recruit interviewers, often teachers and
nurses, but sometimes people with no training in health or education. The
interviewers must complete a five-day course before they start.

Once the data have been collected, researchers analyse them locally
wherever possible, using whatever computer hardware, software and staff
are available. The WHO’s regional centres or consultants will provide help
where necessary and offer training.

The big question is what researchers will do with all the information.
As well as giving the modellers something to work with, the data should
point to areas in which health campaigns can improve, say the researchers.

For example, the studies might reveal common misunderstandings, about
contraception for instance, or about what practices are risky. Health educators
could then sharpen their messages to take account of these gaps in people’s
knowledge. Whatever else we learn, we should no longer have to rely on Kinsey.