A Simple Gene Test Helps Curb Tuberculosis in Prisons

Diagnosis may be the weakest link in treating and preventing tuberculosis, a contagious bacterial disease that infects some nine million people and kills about 1.5 million each year, according to the World Health Organization (WHO). In a study published today in PLoS Medicine, Stanford researchers and their collaborators show that a DNA-screening device called GeneXpert is the most cost-effective option for diagnosing the bacterial disease in a prison environment and reducing its spread compared to the methods currently recommended by the WHO.

In the prisons of the former Soviet Union, TB rates are among the highest in the world. These prisoners are 10 times more likely than the general population to have the disease and extremely likely to have a drug-resistant form of TB. This situation is a problem not only for the prisoners but also for the people that interact with them once they are freed. The prisons act as a reservoir for the disease. Chest X-rays can diagnose TB, but they can’t distinguish between the TB strains that should be treated with a typical course of drugs or drug-resistant versions that need different and more costly medications. GeneXpert can quickly diagnose the type of strain from a small mucous sample.

But this genetic screening test is more expensive than other diagnostic methods. So Stanford researchers set out to determine whether a larger investment in the beginning of patient treatment was cost-effective over the long term. The researchers used computer models to predict that GeneXpert can reduce TB in inmates by nearly 20 percent within four years as long as patients are given the appropriate drug treatment. According to a press release on the study, the authors say these findings could help shape government and medical decisions in the fight against TB.

It appears the idea is already catching: The WHO endorsed the genetic analysis technique late in 2010 and is monitoring its global roll-out.  

MIT Technology Review will explore this issue in depth in a feature story next week.