First African study offering PrEP to the general population finds that about a quarter of people at high HIV risk start it

By far the biggest predictor of PrEP use and high adherence is self-perceived risk
Image: Avatar_023/

An interim survey of PrEP uptake and adherence in the SEARCH study in Kenya and Uganda, from which recently reported qualitative findings, has found that just over a quarter of people who were either assessed, or assessed themselves, as being at high risk of HIV, started PrEP.

Of those only 56% consistently engaged with the PrEP programme, and of those, 66% had adherence sufficient to prevent HIV infection – defined as evidence of four or more PrEP pills taken in a week – though this varied significantly between subgroups.

This implies that roughly 10% of those who tested HIV negative in the SEARCH study and were at elevated risk of HIV were using enough PrEP for it to be consistently effective in preventing HIV.


multivariate analysis

An extension of multivariable analysis that is used to model two or more outcomes at the same time.


Qualitative research is used to explore and understand people’s beliefs, experiences, attitudes or behaviours. It asks questions about how and why. Qualitative research might ask questions about why people find it hard to use HIV prevention methods. It wouldn’t ask how many people use them or collect data in the form of numbers. Qualitative research methods include interviews, focus groups and participant observation.


A serodiscordant couple is one in which one partner has HIV and the other has not. Many people dislike this word as it implies disagreement or conflict. Alternative terms include mixed status, magnetic or serodifferent.

voluntary male medical circumcision (VMMC)

The surgical removal of the foreskin of the penis (the retractable fold of tissue that covers the head of the penis) to reduce the risk of HIV infection in men.


Breakdown product of creatine phosphate in muscle, usually produced at a fairly constant rate by the body (depending on muscle mass). As a blood test, it is an important indicator of the health of the kidneys because it is an easily measured by-product of muscle metabolism that is excreted unchanged by the kidneys.

The programme found that uptake and adherence rates were, as has been observed from other studies, significantly lower in younger people and especially in young women, and in frequently mobile people – exactly the groups at highest risk of HIV infection.

On the other hand, people who were polygamous and/or divorced, separated and widowed, and people with HIV-positive partners, especially female partners of HIV-positive men, were more likely to take PrEP.

But by far the biggest influence on whether someone had engagement with the PrEP programme, and high adherence, was that they considered themselves to be at current HIV risk. People in this category were over 12 times more likely to be on PrEP and adherent than people who did not see themselves as being at risk – even if their opinion differed from the assessment of an algorithm that calculated risk.

The SEARCH Study

The researchers, headed by Dr Catherine Koss from the University of San Francisco, reported the first interim analysis of population data from the PrEP extension of the Sustainable East Africa Research in Community Health (SEARCH) study in The Lancet HIV. This is the second phase of a study whose first phase produced some of the highest proportions of people on treatment and virally undetectable in community-randomised strategies of its type.

The PrEP phase of SEARCH started in 2017, but this is the first detailed analysis of the characteristics of people who decided to start – and those who did not. It was accompanied by an interesting qualitative study of young people’s attitudes to PrEP, which reported on last month.

Data for both studies, qualitative and quantitative, were collected in the year between June 2016 and June 2017. At this time, PrEP was an unfamiliar HIV prevention intervention. Since then, Kenya has adopted a national rollout of PrEP and currently, with an estimated 56,000 people on PrEP in a population of 50 million, or more than one in every 900 Kenyans on PrEP, has the highest percentage of the general population on PrEP of any country.

Who was offered PrEP 

Of 85,047 adult residents in the SEARCH study area, 70,379 (83%) were tested for HIV, and of these 1258 people (1.8%) tested HIV positive. Of the 69,121 people known not to have HIV, 81% received specific post-test counselling which included information on PrEP, but more targeted, individualised PrEP counselling was given to 12,935 people who were either identified, or identified themselves, as being at high risk of HIV (18.7%).

Ten per cent of this ‘elevated HIV risk’ group (1.9% of all who tested HIV negative) consisted of people who were in a relationship with an HIV-positive person. Although the viral loads of the positive partner in linked pairs were measured in SEARCH, this information was not available in time for the risk assessment; figures obtained later showed that 80% of HIV-positive partners at any one time were in fact virally suppressed. Given that 70% of serodiscordant partnerships lasted more than a year, this knowledge may have influenced uptake of PrEP by HIV-negative partners, which was unusually low compared with some other studies of serodiscordant couples.

"Women who said they were married and monogamous were more likely to start PrEP than unmarried women."

Fifty-four per cent of people given individualised PrEP counselling were people who were predicted to be at HIV risk based on an algorithm that took into account age, sex, marital status, polygamy, education, occupation, alcohol consumption and, in the case of men, circumcision (it did not include questions about condom use, transactional sex, other concurrent sex, or same-sex encounters). The researchers emphasise that this ‘machine learning-based risk score’, which was based on seroconversion data in the first two years of SEARCH, “was not used to exclude individuals from PrEP eligibility, but to foster conversations about it.”

The other 36% were people who self-identified as being at high HIV risk and came forward for PrEP counselling, even if the algorithm did not predict it. There were overlaps: 61% of people with HIV-positive partners also assessed themselves as being at high risk, as did 32% of all people identified by risk score did.

Who started PrEP

These 12,935 people at ‘elevated risk’ were offered PrEP. However, only 3489 people – 27% of the ‘elevated risk’ group, or 5% of all those who tested HIV negative – started PrEP within three months of being tested and counselled (82% of these did so immediately).

Gender, educational attainment, alcohol use and, in men, circumcision status did not have much or any influence on whether people started PrEP. In multivariate analysis, occupation also became statistically non-significant. HIV-negative people who were in a serodiscordant relationship were more than two times more likely to start PrEP.

The factors that reduced the likelihood of people starting PrEP were:

  • Youth. Compared with people aged 35-44, people aged 25-34 were 39% less likely to start PrEP, and 15-25s were 45% less likely. 
  • Unmarried people. Compared with them, people who were married and polygamous, or divorced/separated/widowed, were respectively 54% and 59% more likely to start PrEP.
  • Mobility: people who had changed address or moved into the district in the last 12 months were 39% less likely to start PrEP.

The age disparity among PrEP starters was particularly marked in women – while men aged 25-34 were 28% less likely to start PrEP than men aged over 35, women in this age group were 48% less likely to start PrEP than older women, and women aged under 25 were 57% less likely. In the qualitative study it was revealed that young women, especially, often felt obliged to ask partners’ permission before starting PrEP.

One other factor where women were less likely to start PrEP than men was occupation. Women who worked as fisherwomen, in transport, or in bars were 32% less likely to start PrEP than farmers, even though women in this occupation sector were nearly five times more likely than farmers – the most numerous and least at-risk group – to be at elevated HIV risk.

Women who drank alcohol were less likely to start PrEP, whereas men were no less likely.

On the other hand, women who said they were married and monogamous were more likely to start PrEP than unmarried women, or married and monogamous men – probably because it was their partners they did not trust to be monogamous, rather than themselves. Similarly, women with HIV-positive partners were 140% more likely to start PrEP, whereas men with positive partners were only 54% more likely.

Who stayed on PrEP

The researchers measured who was still actively engaged in the PrEP programme at four and 12 weeks after they started, and then every 12 weeks up to week 72 (about 15 months). 

The vast majority of people who dropped out of the PrEP programme did so within the first month: 36% of people who initially started PrEP did not turn up for their second appointment. After this, however, retention varied little, with 60% attending their third appointment and a constant 54-56% attending all subsequent quarterly appointments till week 72.

"Self-perception of HIV risk was, by a long way, the factor most strongly associated with staying on PrEP."

Self-reported adherence, defined as taking at least one dose of the last three (though in practice most people who reported taking one dose reported taking all three doses), was 40% of participants at four weeks, 35% at week 12, and 25-27% thereafter.

These adherence levels look fairly dismal; it means that only 27% of the group who were assessed as being at high risk of HIV (7.3%) was taking PrEP at adequate levels to prevent it, and only 1.4% of the entire group of HIV-negative tested people.

However, this picture was transformed if one additional question was asked at quarterly visits – did the person consider themselves currently to be at high HIV risk?

The proportion answering ‘Yes’ to this question was 77% at week four and then varied little between 60% and 63% in all subsequent returns. People who did consider themselves at high risk had high levels of medication refill throughout – a consistent 89-93% for all visits – and much higher self-reported adherence, in the range of 70-76% for all visits.

It’s important to emphasise that this 77% of people who reported they were currently at high risk were not always the same people. Fifty-four per cent of people seen in follow-up appointments said they were not at risk in at least one visit, and there was a fairly steady rate of change, with roughly one in five people who said they were at high risk at one appointment saying they were not at risk at the next, while another 20% changed risk status the other way.

Many people who stopped PrEP later restarted it. Of all the people seen for a follow-up appointment (i.e. excluding the 36% who never attended again after their first visit), 82% stopped at least once but 50% of those later restarted, and 18% restarted more than once.

Factors that influenced joining the PrEP programme in the first place also influenced adherence. In multivariate analysis, i.e. singling out the influence of each factor by itself, young people aged under 25 were 41% less likely to report adequate adherence; divorced, widowed or separated people were two times more likely to report good adherence; and people with an HIV-positive partner were 64% more likely.

But again, by far the biggest positive influence on adherence was whether people considered themselves at current HIV risk: if people fell into this category at any one clinic visit, they were over 12 times more likely to report adequate adherence than if they did not.

Self-reported adherence varied from group to group. For instance, only 59-60% of young people of either gender aged 15-24 who said they were at current HIV risk reported adequate PrEP adherence, compared with 81% of women with HIV-positive partners. 

A subset of the people who said they were at current HIV risk also had adherence measured directly through drug levels in hair samples. This showed that self-reported PrEP adherence was reasonably accurate. At week four, 57% of all participants had drug levels in hair indicative of effective PrEP adherence (i.e. over four doses a week), and 44% seven days a week. However, fewer than 30% of young women and about 40% of young men appeared to be taking four or more doses a week, compared with 80% of men aged over 25 years (though only 50% of women over 25).

By week 24, somewhat encouragingly, the proportion of women with effective PrEP levels (four or more doses a week) had risen to 55% in women under 25 and nearly 70% in women aged over 25. Adherence among the group as a whole had risen a bit, from about 58% to about 67%.

However, as the researchers remark, these are still lower than ideal levels, especially for women, who for biological reasons may require higher levels of adherence than men.

Side effects related to PrEP were rare. Only one person, a man aged 71, had a creatinine elevation suggesting serious kidney damage, and his creatinine figures returned to normal when taken off PrEP. Although there was a total of 29 serious adverse events, only five were even possibly related to the study drug. There were seven deaths (mortality rate: 0.2% a year).


This was one of the first programmes anywhere where PrEP has been on offer to an entire population rather than to pre-defined risk groups (sex workers, men who have sex with men, and so on). The researchers emphasise that, although an attempt to objectively estimate people’s HIV risk was made, a third of people who started PrEP chose to do so despite not being assessed to be at elevated risk. Self-perception of HIV risk was, by a long way, the factor most strongly associated with staying on the PrEP programme and with high adherence.

"People who did consider themselves at high risk had high levels of medication refill throughout."

Despite this, a third of people who saw themselves as at high risk, and 50% or more of young people aged under 25, did not take protective levels of PrEP. As in so many studies, young women both had the highest risk of HIV and were the most likely to drop out of the programme and have poor adherence. In contrast, men, a group not previously prioritised unless they are men who have sex with men, showed higher-than-expected levels of interest and adherence.

The researchers comment that “rapid introduction of universal PrEP access during population-level HIV testing resulted in nearly 3500 PrEP starts within 90 days in rural Kenya and Uganda.

“We found that a substantial proportion of the population want PrEP and can take it, including many who would not otherwise have had access to PrEP without inclusive eligibility.”