Any news about Alzheimer's attracts my attention these days since it will become the major health issue for Canada as the population age profile gets older and it will have many financial consequences for retired people.
It is thus that a press release from the Baycrest health institute affiliated with the University of Toronto announcing with breathless seriousness that "A Canadian science team has found more dramatic evidence that speaking two languages can help delay the onset of Alzheimer's symptoms by as much as five years." Wow! Here comes the government with programs for teaching French, English or other second languages (any two will do, apparently). As a bilingual person I would love to believe the conclusion but skeptical me wonders if these professionals have made a very fundamental mistake in confusing association or correlation with causation.
Do a Google search with the words spurious correlation and causation and read the dangers of confusing the two ideas. One of my favorites is the video Everything is Dangerous: A Controversy from the American Scientist in which many spurious claims in medical research are dissected by Stanley Young Director of Bioinformatics at the US National Institute of Statistical Sciences for their faulty science and application of statistics. The most basic smell test - does this sound too incredible to be true? - when as they say themselves "There are currently no drug treatments that show comparable effects for delaying Alzheimer's symptoms", should make everyone highly suspicious. The lack of any reference to a physical causal (chemical, biological etc) process to link language-speaking with slowing down the brain gunk (to use a term that shows the depth of my medical knowledge on the subject) present in Alzheimer's undermines my confidence in the announced results even more.
The researchers say they are "dazzled by the results" in the original 2007 study announcement from Baycrest. Too ironically true.
Wednesday 10 November 2010
Subscribe to:
Post Comments (Atom)
Stumble It!
2 comments:
They state it is a correlation, except they use the word "associated".
Confusing causation with correlation is bad. However, that doesn't mean all correlative studies are automatically bad.
If their study is well constructed, and they controlled for all variables other than bilingualism, then their statement is probably correct: Being a lifelong bilingual delays the onset of Alzheimer's symptoms.
And they do propose a mechanism: Cognitive Reserve.
There is a widely published joke about how nobody knows if parachutes work, because all we have is correlative information due to a lack of randomized trials necessary for causative proof.
Well, it makes sense: lots of studies show that people who use their brains more delay the onset of Alzheimer's, MCI, and other forms of dementia. Whether they're suduko players, bilingual, or just avid readers.
What's interesting though is that some of these same studies have found that while the onset can be delayed, once it does start the impairment proceeds more rapidly than in the controls, so that by a few years in there's no difference.
Post a Comment