It is not impossible to provide mobile, point-of-care viral load tests in resource-poor settings, the 2014 Treatment as Prevention workshop in Vancouver heard last week. By using a number of techniques including sample pooling, high throughput, and training lay testers, it should be possible to provide a viral load for not much more than the cost of a point-of-care CD4 count.
Sharonann Lynch of Médecins sans Frontières (MSF) told the workshop that an increasing number of resource-poor countries were including viral load testing within their guidelines and that in future, if global guidelines moved, as widely predicted, towards a policy of universal treatment on diagnosis, then viral load testing would become more important than CD4 count in terms of monitoring treatment failure and treatment adherence.
Can it be afforded, though? Traditionally, viral load tests have been both more expensive and also not capable of being done in the field, as they rely on DNA amplification, a technique that has hitherto required bulky laboratory-based equipment and also a source of mains electricity.
Lynch told the workshop that several different point-of-care viral load testing machines, some battery-powered, were due to come onto the market over the next few years that would make it easy to perform viral load tests in the field. These would contain an entire automated viral-load testing kit within a machine: all that is needed is a drop of blood. Results are available within minutes.
The four machines MSF studied are the Alere Q HIV test, the DRW SAMBA test, the Wave-80 EOSCAPE-HIV test, and the Lumora “BART” test.
At present, to MSF’s knowledge, 22 low- or middle-income countries incorporate viral load testing for treatment failure in their guidelines (6 saying it must be done, 16 as an option) and 12 incorporate viral load testing for routine monitoring (5 as an expectation, 7 as an option). MSF calculates that 19 out of 22 countries recommending it for treatment failure and 9 out of 12 recommending it for routine monitoring are actually unable to afford viral load testing to the level their guidelines would indicate.
Will point-of-care (POC) testing be cheaper? Not necessarily, Lynch said, and to start off with, the per-test cost could be more.
One reason is that 63% of the cost of a viral load test is the disposable chemical reagents and other disposables such as filters that form part of any viral load test. These cannot easily be made cheaper. Other costs, however, could be lower with point-of-care testing such as the 11.7% proportion that is due to human resources such as technicians, and the 1.92% that is lab charges. In addition, the 9.6% that is due to quality control might be lower in a self-contained machine.
It was estimated that the manufacturers’ costs per current, laboratory-based viral load test varied from $1.61 to $6.77 (average $4.36), with the more recent ones for which royalties had to be paid costing more. The price paid by wholesalers ranged from $11 to $25 per test and the final price from $18.62 to $36.38.
There was a wide range of costs actually paid by African countries, however. Two, Kenya and Uganda, had knocked the price per test down to $10.50 while neighbouring Tanzania paid the region’s highest price at $55.
The new point-of-care tests were more expensive because they had features such as having reagents and vessels contained within special units that were more expensive to make. As a result the manufacturers’ costs ranged from $4.84 per test to $9.33 (average $7.28 – nearly $3 more than a lab test).
Point-of-care tests became cheaper as they were used more often. The final cost of a point-of-care test machine would be about $42 per test if the machine was used to 25% of its capacity and $33 if it was run at 75% capacity.
This, however, compares with an average cost of $7.33 for a point-of-care CD4 test.
There were other ways of reducing costs. One is to pool samples and then only re-test samples that produce a positive signal – a particularly economical method if used in patient groups that are on treatment and mainly expected to have an undetectable viral load. Sensitivity can initially be set at a figure of, say 1000 or 5000 copies/ml, so only people likely to be infectious are picked out and not isolated ‘blips’. Pooling samples could reduce costs by 28% if the limit of detectability was set at 1000 copies/ml and by 51% if it was set at 5000 copies/ml.
With all other economies, this could potentially reduce the minimum per-test cost for point-of-care tests from $24 to about $12 – still more than a CD4 count, but not multiply so. Even more costs could be saved if a decision was taken to stop using CD4 counts for monitoring and only take one initial one.
Without viral load, Lynch said, it is impossible for sites in resource-poor settings to know how well antiretroviral therapy is working. The good news is that it was working at least as well as in richer settings. In one area in Kenya, for instance, 85% of patients on ART and an estimated 40% of the entire HIV-positive population, including the undiagnosed, had a viral load below 1000 copies/ml (often used as the threshold for infectiousness) while in another in Malawi 91% on ART and 62% of everyone with HIV aged 15-59 (thus excluding children) has below 1000 copies/ml. Two sites in KwaZulu Natal, South Africa, had achieved viral suppression in 94.7% of diagnosed people within six months of starting ART.
Lynch said MSF was campaigning for an ambitious target of 86% viral undetectability for everyone living with HIV, diagnosed and otherwise, and was recommending that international agencies adopt something similar for their goals for 2020.
Lynch S Viral load testing – are cost and complexity out-dated arguments? 2014 Treatment as Prevention Workshop, Vancouver, abstract 5082, 2014. See programme here.