I came across this 2007 article on bilingualism and dementia through something I was reading on the New York Times website. A Canadian research team based at York University, University of Toronto and Rotman Research Institute at Baycrest have concluded that bilingualism may be protective against early onset of dementia. Bilingualism has benefits beyond getting that coveted government job in Canada? Interesting indeed.
As I patted myself on the back for having this supposed protective factor, I was interested in what the findings were, beyond what was reported in the article. To the original paper I went. One of the most important things I’ve learned in my current post is that study designs matter. Study design refers to how a research study was conducted – important features such as population sampled and number of measurements can greatly influence the conclusion we can draw with a degree of confidence. Without going into too much detail regarding study design (I may do this at a later time, but for now see this brief summary or this book) and delving into a philosophical and methodological digression about the merits of each type, this study was a clinical case series comprised of 228 consecutive patients (cases) attending a Memory clinic. The final sample comprised of 184 patients with a diagnosis of dementia AND determined to be either monolingual or bilingual (the other 44 were excluded from further analysis). Onset of dementia was found to be 4.1 years later for bilinguals than monolinguals. Along with being statistically significant, a delay of 4.1 years, the researchers’ suggest, is also clinically significant for reducing prevalence of dementia in the population.
Before we start enrolling ourselves and our children in language classes, bear in mind, the bilinguals in this sample were mostly immigrants while the monolinguals were not. However, the authors suggest that immigration did affect the direction of the findings (the association of bilingualism with later onset of dementia). The authors of both the article and journal paper acknowledge that long term studies are needed. A snapshot in time of disease onset based on the subjective accounts from 184 patients may not be representative of the entire population.
My real concern is over the reporting of this study. Not every reader will have an intimate understanding of study design, their relative advantages and shortcomings, and the strength of evidence that can drawn from each type. I have an inchoate knowledge of study designs myself. Although the author acknowledges that drawing a causal relationship between bilingualism and dementia would be premature at this time, the suggestive nature of strategic reporting leaves a different impression.
The researchers acknowledge a complex set of relationships linking biological conditions, cognitive processes and environment circumstances. Therefore, bilingualism should not be interpreted as a protective factor on its own. Bilingualism is associated with its own accumulative advantage in countries such as Canada – it can help with employment, improve access to health care, buffer against discrimination and so on.
Scientific reporting needs to take greater responsibility for what they write – they are often the only point of information for the general population, given the largely inaccessible database of journal articles (which is another point of debate). While causality may not be sexy, and sensationalist reporting (yes, even in the scientific realm) may capture more readers, leaving an impression that one course of action may be beneficial for health is irresponsible.
We need to know more before we consign ourselves (and any children) to hours of foreign language study. Does bilingualism confer advantage for other health conditions as well (mental or otherwise)? Does bilingualism offer the same advantage in different countries? Are these results replicable in a larger cohort of patients? Do particular language combinations affect the strength of this protective effect (speaking Mandarin and English as oppose to French and English)? How does this group of bilinguals compare to their counterparts in countries of origin? Currently there are more questions than answers, and this point needs to be made explicit and upfront. Any easy way to do this is write a sentence or two on study design and the confidence of conclusions we can draw from each type.
We are routinely inundated with information on how to improve our health – responsible journalism needs to report on the strength of studies so that readers, myself included, can decide how we want to balance our lives. The supposed pathways to health are numerous and all competing for our time and attention. Should we lift weights or run, eat vegetables or fruits, learn new languages or learn how to play an instrument, take up a hobby or join a club? Show me the evidence.