Hundreds — or even thousands — of Kenyans and Ugandans may have been told that they are infected with HIV when they are not, thanks to faulty rapid, 15-minute tests administered at VCT centres.

Many others may have wrongly been declared negative, clearing them for unprotected sex, when they actually are HIV-positive.

That is the worrying conclusion of a study involving 6,255 people carried out in Uganda and Kenya, which bluntly says that the misuse of rapid tests at most VCT centres makes them fraught with error and that they cannot by themselves alone determine whether one is HIV-positive or not.

The three HIV rapid tests which were evaluated in the study, the findings of which appear in a recent issue of the East African Medical Journal, were Determine by Abbott Laboratories, Uni-Gold by Trinity Biotech of Ireland and Capillus, also by Trinity Biotech.

The tests are widely used in poor countries because they are cheap. Each HIV screening with the tests costs about one dollar, compared with about $40 using the much more reliable PCR test, which is considered the gold standard.

The East African Medical Journal is published by the Kenya Medical Association.

In Kenya, according to Dr Peter Cherutich, Assistant Director of Medical Services at the Ministry of Health and head of HIV prevention at NASCOP, the National Aids/STI Control Council, the three most used rapid tests are Determine, Bioline from SD Bioline of South Korea and Uni-Gold. The first two are usually used as first-line tests while Uni-Gold is confirmatory.

The risk of HIV misdiagnosis using the rapid tests rises substantially when they are used once, without the benefit of a confirmatory test. This practice is thought to be rampant, especially at VCT sites outside medical facilities, which are usually manned by non-medical staff.

In Kenya, according to Dr Cherutich, the requirement for one to work at a VCT site is a secondary education and three weeks’ training in the administration of the tests.

Results from the Kenya-Ugandan study, which involved men aged between 18 and 60 seeking VCT services from a rural village in Masaka and the Kakira sugar plantation, both in Uganda, as well as a coastal village in Kilifi, and the Nairobi slum of Kangemi, confirm that there are important issues of quality assurance and dependability of results obtained using the rapid tests.

When used as a single test in Masaka, for example, Determine was able to correctly identify only 45.70 per cent of those infected with HIV as carrying the virus.

This means that out of every 100 HIV-positive people, 54 could actually have walked out of the VCT centre thinking they were HIV-negative, when in actual fact they were not, if they were not advised to take a confirmatory test.

In Kakira, the same test only captured 65.71 per cent of the infections, while the Capillus test captured 86.62 per cent of infections at the same location.

When two different tests were administered on all 6,255 subjects, a total of 131 people had “discrepant results” — where one was positive and the other negative.

Using a third confirmatory test, 27 of the subjects were finally confirmed to be carrying the virus, meaning that without such controls and using just one test, 104 people could have been told they were HIV-positive when they were not.

While the tests were found to be able to establish HIV-negative persons with a fairly high degree of correctness, some results were worrying.

In Kangemi, Nairobi, for example, the Uni-Gold test was found to have an accuracy of 99.13, meaning that out of every 10,000 HIV-positive people tested, 87 would wrongly be declared HIV-negative.

Say the researchers involved in the study: “With VCT uptake at government sites (rising), use of a single test could misdiagnose thousands of HIV-infected individuals. No rapid test is performed well-enough to recommend its use as a single diagnostic measurement.”

The researchers were led by respected Kenyan Aids scientist Prof Omu Anzala and included specialists from Uganda’s Virus Research Institute, Kampala’s Joint Clinical Research Centre, the Kenya Aids Vaccine Initiative (Kavi) and the International Aids Vaccine Initiative (IAVI).

According to NASCOP’s Dr Cherutich, about 2 to 2.5 million Kenyans visit VCT sites every year. Kenya’s Ministry of Health, he said, has instituted guidelines that at least all service providers use two rapid tests before informing their clients of their status.

Kenya has an estimated 1,000 independent VCT sites, which in total screen about 800,000 people annually. A similar number is tested at government hospitals voluntarily, and another 700,000 after seeking treatment, making up the total.

While acknowledging that “a significant” number of false positives are probably being reported with the rapid tests at various sites, Dr Cherutich was last week however categorical that the use of the methods was the most advisable for a country like Kenya.

“The rapid tests are cheap, easy to use and do not require special conditions such as refrigeration,” Dr Cherutich told The EastAfrican. “In our environment, where HIV/Aids is a national emergency, we should ask ourselves what is the most cost-effective screening method that can be rolled out nationally most effectively and efficiently.”

Because of their relative unreliability, however, rapid tests are no longer used in the Western world.

According to Dr Cherutich, the fact that the rapid tests are more likely to give false positives that declare an uninfected person HIV-positive is not an insurmountable obstacle.

“A false positive will probably be eliminated with subsequent tests,” Dr Cherutich said. “The rapid tests have given the fight against Aids a boost that we wouldn’t have otherwise had. We simply can’t afford PCR for everybody.”

Prof Anzala and his colleagues agree, but with an important rider — that the rapid tests be always administered with a confirmatory standard before a definitive HIV status is declared.

According to Kenya’s National Aids Control Council, there are an estimated 1.3 million Kenyans who are HIV-positive. Across East Africa, about five million people are thought to carry the Aids virus. Many would have been so declared at VCT sites.