Rose's treatment failure has been detected late, once it's already in full swing - and this reduces the likelihood that the second-line treatment will work.

Rose is 30 years old, and has been receiving treatment in an MSF clinic in Malawi for three years. Rose has shown good results since starting treatment and seems to have been adhering well to treatment, but now her health has deteriorated: she's losing weight again, feeling tired and has picked up a case of oral candidiasis, a common opportunistic infection related to HIV/AIDS.

Concerned by her decline, Rose's doctor sends her for tests. The results confirm that Rose's CD4 count has fallen to 140 although it was at 240 eight months ago when it was last measured. Her doctor determines that Rose is in "treatment failure" - that is, the ARV drugs she is taking are starting to no longer work and the virus is replicating again. If left unchecked, Rose will likely die in the coming months.

Her doctor decides to change from the first-line drugs she was taking to more powerful second-line drugs which should be able to better combat the virus.

But Rose's treatment failure has been detected late, once it's already in full swing - and this reduces the likelihood that the second-line treatment will work. Since she is changing her treatment this late, her chances of the new combination working have dropped from 80% to 50%.

The problem is that detecting treatment failure at the proper time depends on sophisticated laboratory equipment often not available or affordable in poor countries like Malawi.

There are two main types of tests to monitor patients under treatment. The first type of test counts CD4-plus lymphocytes, a key component of the immune system which HIV attacks - the lower the CD4 count, the weaker the immune system. The second type measures the viral load, the amount of virus present in the blood - if ARV treatment is working, viral load will be low or undetectable. Viral load monitoring detects treatment failure earlier than CD4 count.

But the tests and equipment used to record these tests are expensive. Even simplified CD4 machines cost around US$16,000 plus reagents costs for each test run. Viral load machines are upwards of US$130,000 and also require reagents.

MSF in Malawi is using a CD4 machine which performs up to 50 tests per day. This allows for the routine testing of all ARV patients once a year and suspected treatment failure cases on demand. But MSF presently has no access to viral load tests and equipment: there is only one in Malawi, at the other end of the country.

There are two strategies which can be used simultaneously to improve access to CD4 and viral load tests. First, simple rapid versions of these tests need to be developed that can be used at typical health posts by personnel with minimal training. Second, because it is not going to be feasible to put CD4 and viral load tests in every local clinic, techniques need to be developed to allow samples to be safely transported to laboratories in district or national hospitals and then have results communicated back.

[This example is a composite of several patients' stories.]