Wednesday, May 27, 2015

DG XXII: The Return of Doctor Akbaraly

Recap:
Last week Mig Greengard was misdescribing a journal article (DG XXI: Doctor Akbaraly) which we’d previously seen him cite in support of a claim that there was "strong evidence of cognitively stimulating evidence" helping with dementia (DG XX: Homework).

Today, we move on to the study’s ...
Akbaraly et al (2009), Leisure Activities and the Risk of Dementia in the Elderly: Results from the Three-City Study, Neurology vol 73 no 11, 854-861
... conclusions. If you’re suspecting that they might not be entirely as advertised by Garry Kasparov’s (Doctor Garry is In) bag carrier (DG XVIII: Mig), you would be correct.

That’s for later, though. Let's start with,
Stimulating leisure activities were found to be significantly associated with a reduced risk of dementia … and Alzheimer (sic) disease
Akbaraly p. 854
That you can see in the Abstract that is available online. Read the article itself (not on the internet and therefore harder to find, but it’s available at the British Library) and you find more details about the results that the researchers obtained.
Stimulating leisure activities were associated with a 50% reduction in risk of dementia in participants with high or moderate levels after controlling for potential confounders. Similar associations were observed for mixed/vascular dementia.
p. 856
Only for "stimulating" activities, though.
For passive and physical leisure activities, no significant association was found between levels of participation and risk of any type of dementia.
p. 856

Ball in the back of the ’strong evidence’ net for Mig, then? Actually, no.

It might seem like an open and shut case that Akbaraly’s article support’s Mig’s claim, but in fact we'll see that it doesn’t. Well see it doesn’t because there’s a difference between 'association' and causation. Well see it doesn’t because of the nature of Alzheimer’s Disease and other dementias and a known research problem.

Aside from anything else, though, we’ll see that the Akbaraly study doesn’t provide "strong evidence" that cognitively stimulating activities help with dementia because Akbaraly and his fellow authors actually say themselves that it doesn't.




Association and causation. I must admit, I’ve never found it particularly taxing to distinguish these concepts. It does seem to trip up an awful lot of people, however, so let’s do a speedy review:-

  • When two things come together in some way - that’s an association.
  • When one thing makes another happen - that’s causation.

That’s pretty much it.

An association may also be a causal relationship but it ain't necessarily so.  If A and B come together it may be that A makes B happen. Or maybe not.  Maybe it’s that B leads to A. Or maybe it’s a bit of both with A and B interacting with each other. Or maybe it's something else entirely that causes A and B to occur.

So if you find an association it may be that what you’re observing is one thing causing another, but until you’ve ruled out all the other possibilities you can’t be sure. That’s why researchers are very cautious about what they say.

And what Akbaraly and his colleagues say is "association". There’s a reason for that.

" ... the pathologies underlying dementia likely develop insidiously for decades before clinical onset ... Thus it may be that a lower number of total activities is a preclinical marker of dementia."


For dementia research, just like many other areas, the difference between association and causation is not just an abstract concept. It’s a fundamental problem to resolve. There are many reasons for this, but arguably the principle one is the nature of dementia itself.

Alzheimer’s Disease (like many other forms of dementia) is a progressive condition. Symptoms develop over time. It is entirely possible, likely even, that people may have Alzheimer’s long before it is recognised and acknowledged.

Something else that we know is that people who live with Alzheimer’s also very often become increasingly socially isolated and withdrawn. This presents us with a research challenge.

When we take a group of people and find that those who are actively engaged in cognitively stimulating leisure activities are less likely to develop dementia that than those who aren’t, is this because,

(a) engaging in cognitively stimulating activities protects against developing dementia; 
or 
(b) withdrawal from leisure activities is an early sign of a dementia that hasn’t yet been recognised or formally diagnosed?

That is, do leisure activities have an impact on risk of dementia or does dementia affect leisure activities? The answer - as frequently recognised by the authors of your homework reading articles - is that it could be either one.

"Both explanations have important implications for the cognitive health of older adults and should equally considered when interpreting the results of studies with relatively short follow-up periods (i.e., less than 10 years)."

One of the known potentially problematic issues in dementia research is a study’s length of follow-up. That is, how long do you keep watching after your 'baseline measurements’ of activity levels are first taken? For Akbaraly it was 4 years. That’s not so different from a lot of other studies, but nevertheless, according to Hughes and his fellow writers, it’s nowhere near long enough to come to a conclusion either way, regardless of what you find.

Needless to say, the Akbaraly authors were aware of all of this. That’s why they acknowledge that their short follow-up period is a "limitation" of their research.

…. it is plausible that apathy and loss of initiative might precede the cognitive impairment by more than 4 years, and thus our results could reflect early symptomatic behavioural changes of dementia.
Akbaraly et al (2009), page 860

and go on to state explicitly that,

... the design of our study ... does not permit us to conclude that there is a causal link between mentally stimulating leisure activities and the onset of dementia and AD ....Akbaraly et al (2009), page 860


" ... [there is a] growing body of evidence that engagement in activities in late life that require some degree of cognitive effort may offer protection against dementia.

Controlled intervention trials of carefully selected activities may be the next step in elucidating these relationships and demonstrating their efficacy in maintaining cognitive health and preventing or delaying dementia."


So, yes, Akbaraly and his fellow writers found an association between taking part in stimulating leisure activities and reduced risk of dementia. It’s also true that they concluded, "Our findings support the hypothesis that cognitively stimulating leisure activities may delay the onset of dementia in community-dwelling elders." (see the Abstract).

The key phrase there being "support the hypothesis". The key word being "may".

There’s nothing wrong at all with the Akbaraly study. The issue is one of how it’s been interpreted. Yes, the study supports a hypothesis that cognitively engaging activities may help with dementia. They may not, however, and besides the evidence that Akbaraly found supports another hypothesis too, that people might be reducing their level activity because they are in the early stages of dementia.

There’s nothing in Akbaraly’s work that allows us to choose between these two possibilities. We need more research, more of a different kind of research - controlled intervention rather than observational studies - before we can say for sure what the findings of "Leisure Activities and the risk of Dementia in the Elderly" and other similar articles actually mean.

Which is, when all is said and done, pretty much exactly what Verghese concluded (DG VI: Doctor Susan) in 2003 (DG II: Dogs That Don’t Bark). Titles not being the only thing that the papers produced by our old friend Joe and Akbaraly have in common.

The results of both studies are certainly encouraging and should be pursued further. Furthermore, older people - and everybody else for that matter - should be encouraged to remain active anyway (Remember DG VII: Doctor Coyle?).

All this is a very long way from Mig Greengard’s claim of "strong evidence", though. In trying to use Akbaraly’s study to support that assertion he is simply wrong. Again.



Chess and Dementia Index





NOTES:
For dementia research, just like many other areas, the difference between association and causation is not just an abstract concept. It’s a fundamental problem to resolve. There are many reasons for this ....

e.g.
as the Akbaraly study itself says (page 854):-
"… as leisure activities may reflect socioeconomic advantages gained earlier in life, it is also possible that the association between late-life activities and the risk of dementia may be explained by education level or occupational attainment."

Similarly, Dartigues et al (2013), page 5-6:-
"... the risk of dementia could be related to the fact that board game players had better cognitive performances and were less depressed at baseline screening than non-players."



Alzheimer’s Disease (like many other forms of dementia) is a progressive condition.
See: The Alzheimer’s Society: The progression of Alzheimer’s disease and other dementias



Something else that we know is that people who live with Alzheimer’s also very often become increasingly socially isolated and withdrawn.
Again, see: The Alzheimer’s Society: The progression of Alzheimer’s disease and other dementias



The answer - as frequently recognised by the authors of your homework reading articles - is that it could be either one.
e.g.
Verghese et al (2003),
"The presence of preclinical dementia might reduce participation in leisure activities ...."

Hughes et al (2010), page 432
"... a negative association between late-life leisure activity and dementia risk could indicate that either a higher level of engagement provides protection against dementia or a lower level of engagement is an early effect of dementia."

Wang et al (2012), page 489
" ... decreased leisure activities may be a consequence of a symptom of preclinical dementia ... the possibility of reverse causation remains a major limitation for the observational studies of cognition and dementia."

Dartigues et al (2013), page 6
"Less board game playing might be an early marker or an early consequence of dementia that precedes the decline ...."

Sorman et al (2013), page 2
"... the significant relationship demonstrated between activity level and dementia possibly signal a causal influence reversed to that suggested before such that lowered activity represent early, preclinical symptoms of the disease."



"For Akbaraly it was 4 years. That’s not so different from a lot of other studies ...."
e.g.
Sorman et al (2013), page 7
"It is warranted to note that a considerable proportion of the evidence that late-life mental and social activities may protect against dementia was from studies involving a relatively short follow-up period."



Akbaraly and his fellow writers found an association between taking part in stimulating leisure activities and reduced risk of dementia
That’s the common finding in such studies, e.g.,
Dartigues et al (2013), page 6
" ... [this study] ... suggests that playing board games has a protective effect on cognitive decline, depression and dementia."

although not all of them. E.g.,

Sorman et al (2013), page 7
"A significant relationship [between leisure activities and risk of dementia] was observed for the "Total Activity" and "Social Activity" indexes but not for "Mental Activity ...."

However, the Sorman study considered "Read books, Read magazines, Movies/Concerts/Theater, Play musical instrument, Needlework, Hunting/Fishing" (page 4) as their mental activities and they say,
"... we cannot rule out the possibility that more demanding activities, such as board games ... or those targeted by intensive training programs might have more substantial effects."



We need more research, more of a different kind of research ....
e.g.
Verghese et al (2003)
"Clinical trials are needed to define the causal role of participation in leisure activities."

Wang et al (2012), page 490
speak of the need for "Well designed studies with large sample size and long follow-up time confirming the potential protective effect of other types of leisure activities on cognitive function and dementia ...."

Dartigues et al (2013), page 6
"Only controlled studies could truly establish whether playing board games is beneficial and could rule out reverse causation."



The results of both studies are certainly encouraging and should be pursued further.

Akbaraly’s final paragraph reads,
"The observed risk reduction of 50% over 4 years for all dementias in participants engaging at least twice weekly in stimulating leisure activities suggests that the promotion of these activities in the elderly community-dwelling populations may constitute a positive public health initiative without negative side effects which could delay dementia onset."

2 comments:

Niall said...

The problem with these kinds of claims is that 90.37% of people (I read that figure in a study :)) will just read the headline and a bit of the article, and that’s it. A small percentage of people will actually mentally question the claims, and an even tinier percentage will read up on the subject and contact the person who has made the claim in the article or tweet. By this time, the person originally making the claim will have moved on to something else, and will either not answer the person who’s questioning the claims or fob them off. In olden times you might get a tiny correction buried away at the bottom of page 12.

Either way, it doesn’t matter, because the person making the claim has got a nice bit of exposure in the media for whatever claim they’re pushing, and for themselves. Why worry about whether the claim is actually true? Truth is for nerds. And as for those annoying people going around asking questions and looking for evidence, why worry about them? It’s not as if anyone’s really listening to them.

Jonathan B said...

I don’t think you need the "really" in that last sentence Niall.

Still, if the blog worried about whether anybody was listening we’d have wandered off and done something else years ago.