More Statistics.

While we are talking about statistics, I will raise a point that I touched on in an earlier post.


As soon as you start researching TB you will, no doubt, trip across the following two widely quoted statistics.


- 30% of the world population is infected with latent TB.
- The lifetime risk of latent TB becoming active is 10%.


So, (ignoring geographic distribution of TB) the above statistics when taken together would indicate that of a group of 1000 people, 300 would be infected with latent TB and 30 of them would progress to active TB during their lives. 


But, let's just look at that a little closer. How were those figures arrived at?


We can probably assume that the total number of active TB cases in any year can be reasonably correctly arrived at. However, how do we know that 30% of the world is infected with latent TB? There is no gold standard for diagnosing latent infection. Until the arrival of QuantiFERON-TB the only diagnostic available for latent TB was the TST. You may remember that the TST has an exceptionally poor specificity (ie it reports many false positives). This is largely due to it's inability to differentiate between past infection and BCG vaccination vs true current latent infection. It is not unreasonable to presume, therefore, that the percentage of people infected with latent TB is a lot less than 30%.


What happens if the correct figure is 15%?


That would mean that of our group of 1000 people, 150 would be infected with latent TB and 30 of those would contract active TB during their life.


That changes our second statistic to read: The lifetime risk of a latent TB infection becoming active is 20%. It's getting scarier, isn't it?


What if the worldwide rate of TB infection is even lower? 15%? 10%? 5%?


Suddenly the whole picture looks a lot different, doesn't it? Spelling it out, the higher the true conversion rate (from latent to active), the more useful, cost effective and imperative it becomes to test for and treat latent TB. 


Recent trials, studies and findings (eg Halder et al) are indicating that the scenario that I have painted above is true. It is becoming clear that the "predictive value" of QFT is several times that of TST. i.e. If you are diagnosed with latent TB by QFT your chance of progressing to active TB is much higher than if you are diagnosed with latent TB by TST. This is not because QFT is finding "a different sort of TB", it is purely due to the fact that TST is reporting as positive a whole raft of people who did not actually have TB in the first place  whereas QFT is reporting true infection.


The conclusion of all of the above is obvious. Not only can we rely on QFT to more correctly identify latent TB infection than with TST but the whole diagnostic imperative changes with a switch to QFT testing. It now becomes a worthwhile  (and necessary) exercise to test more of the population. 


The medical system is driven by many forces - one of them is effective use of limited resources, including money. Given that, it is not hard to see that if a diagnostic has a low predictive value then it becomes less likely that mass screening with that diagnostic is a resource effective solution to a problem. Remember the saying in the TB fraternity - "A decision to test is a decision to treat". Treating a whole lot of people that do not actually have TB is both a waste of resources and, from the patients point of view, an unnecessary evil. If, on the other hand, you have a diagnostic with a much higher predictive value then the dynamics of TB control change quite dramatically. Suddenly, screening of population groups becomes a viable control mechanism.


Once the above becomes accepted as fact (and the evidence is appearing right now) then it is quite likely that we will see a re-shaping of TB control policies in the developed world to include more population groups being screened.