Dementia affects over 57 million people worldwide – and this number is only projected to grow. By 2030, 78 million people are estimated to have dementia. By 2050, it's projected that number will reach 139 million people.

Despite this, a surprising new study has suggested that dementia risk has actually declined with each generation. However, there are good reasons to be sceptical of this finding.

The researchers analysed data from 62,437 people aged 70 and over. Data was collected from three longitudinal cohort studies on ageing, including one conducted in the US, one in Europe and one from England.

To conduct their analysis, the researchers compared probable dementia diagnoses from people born in eight different generation cohorts. The people in the first cohort were born in 1890-1913, while those in the most recent cohort were born in 1944-48.

Related content: Dementia in Younger People Is Often Overlooked. Here Are 5 Key Reasons.

The researchers employed an algorithm that suggested probable dementia diagnosis. This was based on participants' demographic characteristics, as well as their cognitive performance and everyday functioning skills (including how well they were able to perform daily functions, such as washing and feeding, and how well they could remember things).

These are standard assessment tools used in clinical practice to diagnose dementia.

To then validate the algorithm's projections on probable dementia cases, the predictions were compared against a sub-sample of participants from the US Aging, Demographics and Memory study who had a clinical diagnosis of dementia.

The participants in this study had undergone a rigorous three- to four-hour cognitive assessment. The algorithm used to create dementia projections showed an over 85% agreement with clinical diagnosis data from that sub-sample data.

Once dementia status was calculated, the authors computed two models to ascertain the relationship between age, cohort and dementia onset. They also included gross domestic product (GDP) in their analyses, as there's a correlation between GDP and health – with research showing that people in higher-income nations tend to be healthier than those living in lower-income nations.

Their findings revealed that dementia cases decreased for each subsequent generation. For instance, in the US, the algorithm indicated that 25% of people born between 1890-1912 developed dementia, while only 15% of those born in the most recent cohort (1939-1943) developed dementia.

In England, almost 16% of people born between 1924-28 were indicated to have developed dementia, compared with around 15% in those born between 1934-38. This effect was also apparently more pronounced for women than for men.

It remains unclear why dementia cases fell across the generations, with more recent cohorts having reduced dementia cases.

Projected diagnoses

What does this mean? And how does this compare against existing projections?

While the authors used a large sample from three established ageing research cohorts, the findings are based on data from high-income countries only.

It's well known that dementia can be better diagnosed and cared for in high-income countries, where there are more and better healthcare services and professionals. Dementia is hugely stigmatised in low- and middle-income countries as well. As a result, many people are not as aware of dementia as people living in high-income societies.

This means people in lower-income countries may know less about the associated risk factors for dementia and are less likely to receive a diagnosis and support. This is particularly relevant given the fact that most people with dementia reside in low- and middle-income countries.

With a lack of cohort data on older adults and dementia cases in low- and middle-income countries, the findings from this study do not provide representative projections on dementia diagnoses globally.

It's also important to consider the methods the authors used in their study. The authors used a prediction model. Although this model had high agreement with clinical diagnosis, there are still cases of dementia that will have been missed out as a result.

Similarly, the authors did not distinguish between dementia subtypes in their modelling. Dementia is just an umbrella term. About 60-70% of dementia cases are actually Alzheimer's disease.

But there are also many rarer subtypes – such as Lewy Body dementia or semantic dementia. Each subtype brings with it different symptoms. A generic model is unlikely to pick up each subtype dementia case correctly.

All these factors may possibly explain how the study came to their conclusions.

Dementia cases worldwide are actually predicted to increase. As such, the findings from this study should be considered with caution. It may not be the case that dementia prevalence continues to fall for more recently born generations.

Part of the reason for these projections is due to the fact that people are living longer and growing older. Dementia primarily affects people aged over 65, so with more people living to be over 65 this means that more people will be at risk of developing the disorder.

The world population is growing, too. So naturally we're going to see more people living with dementia – particularly in low- and middle-income countries, where people may have less knowledge of dementia symptoms and may be less able to address the modifiable risk factors linked to greater risk, due to poor healthcare infrastructure.

We know that overall, people from more socioeconomically disadvantaged backgrounds experience greater health inequalities – and these health inequalities may contribute to increased risk of dementia.

But as this factor was not taken into account in the study, it's difficult to know whether there really will be any differences in the projected number of dementia cases in younger generations.

Clarissa Giebel, Senior Research Fellow in the Institute of Population Health, NIHR Applied Research Collaboration North West Coast, University of Liverpool

This article is republished from The Conversation under a Creative Commons license. Read the original article.