Post
by **Matt** » Fri Nov 13, 2015 7:38 pm

I see your problem. Obviously words like, bogus, scam, quack, con etc are likely to raise hackles and lead to a defensive response. Enquiring about how accurate a diagnostic process is, might work better. Or even how useful a diagnostic process is.

I'm reminded that the background incidence of a condition is a relevant factor in interpreting results.

So consider a diagnostic process that is 99% accurate in terms of both sensitivity and specificity.

99% sensitivity means that out of 100 patients with the condition 99 will be identified, 1 will be mistakenly given the all clear.

99% specificity means that out of 100 patients without the condition 99 will will be correctly given the all clear, however there will be one false positive.

So if you're screened for a condition and receive a positive diagnosis from a 99% accurate test does that mean there is a 99% chance that you genuinely have the condition diagnosed?

Not really. In the special case that 50% of the population being screened do have the condition then this is true. Out of our 200 patients considered above 100 received a positive diagnosis and in 99/100 cases the diagnosis was accurate.

However if the background rate falls to one in a thousand then receiving a positive result may be far less conclusive.

Screen a million people, of them 1000 people have the condition, 990 of those will receive an positive diagnosis.

Of the other 999,000 people without the condition 1% will get a false positive that’s 9,990 people.

So in that scenario a positive diagnosis that is "99% accurate" really means a roughly 1 in 11 chance of genuinely having the condition.

Since no real world diagnostic technique is 100% accurate positive diagnoses should be confirmed through independent methods.

Now I find that fascinating (which may be a little geeky of me) but perhaps it's something that could be discussed with or in the presence of your friends.

Then at some later date you could refer back to it in asking what the background incidence of food allergies is amongst domesticated animals, or what confirmation they get when they get a hit from bioresonance therapy.