Sorry, you need to enable JavaScript to visit this website.

Studying cognitive health in older adults: lessons from an ongoing all-India survey

Read time: 9 mins
Studying cognitive health in older adults: lessons from an ongoing all-India survey

“Old age is like learning a new profession. And not one of your own choosing.”
Jacques Barzun, Age of Reason

India’s older population is increasing. While our overall population growth is slowing down, the population of adults 60 years and older is expected to double from 9% to 19% by 2050. Meeting the needs of this population will be a growing challenge in the coming decades. Apart from the fact that the physical and financial requirements of older adults differ from those of younger adults, we also face a rising wave of age-related disabilities and diseases.

The LASI dataset

To address the growing need to understand India’s ageing population, the Ministry of Health & Family Welfare launched the Longitudinal Ageing Study of India (LASI) in 2016. ‘Longitudinal’ means that it will follow adults aged 45 and older for the next twenty-five years to understand how their health, social life and financial status change as they grow older. Seventy thousand people from every state and Union territory in India, a nationally representative sample, were enrolled in the study at its start. The mammoth LASI project is a joint undertaking of the International Institute for Population Sciences (IIPS) in Mumbai, the Harvard T.H. Chan School of Public Health, and the University of Southern California (USC), supported by the National Institute on Ageing (R01AG042778) in India.

A human being is influenced by thousands of different factors — everything from genetics and their mother’s health to what they eat for breakfast on any given day. Countless variations in these numerous factors make it very difficult to figure out what causes what. The advantage of a longitudinal study is that it compares a measurement — such as, say, weight — within the same person’s life at different time points. In this way, all the other factors that might cause a change to the measure are held constant, and the only variation is through the hypothetical cause that is being investigated. In the LASI dataset therefore, ageing is the main cause of variation in the population.

More than thirty papers have come out of the dataset, throwing light on older people’s mental, physical and social states and how these change with age. As one might expect, older people are more prone to chronic physical illnesses, such as hypertension and eyesight problems. They are also prone to declining cognitive health, affecting their ability to think, remember and form new memories. All of these abilities naturally decline with age, even as the likelihood of developing diseases like dementia and Alzheimer’s increases. Cognitive health problems are difficult, if not impossible, to cure. Prevention is the only safeguard against them, and some of those preventative measures may be literally a walk in the park.

Physical activity is linked to cognitive health

Physical activity seems to be an effective way to slow down the loss of cognitive ability through age. A recent study of the LASI dataset from the International Institute for Population Science, Mumbai, explored the effect of exercise on the cognitive health (‘cognitive functioning’) of people aged 60 years and older. The researchers drew on data from more than 30,000 individuals across India from various social and economic backgrounds. Cognitive functioning was a composite of five different measures: memory, orientation (the knowledge of one’s location in time and space), arithmetic ability, executive function (tasks such as folding paper or drawing shapes) and object naming. These measures contributed to a single overall score for cognitive functioning. The amount of physical exercise was simply coded as yes: frequent vigorous activity every day; or no: a few times a week or month or never.

The study was observational, meaning that the researchers collected data about what people did in their normal daily lives by asking them directly, instead of intervening and, say, putting them on an exercise program. Self-reported information is notoriously unreliable, especially concerning diet and exercise, which are fraught with social and moral baggage. People underestimate how much energy they take in through their food and are biased in recalling how much exercise they actually did. A lot of self-reported measures for exercise may not have been properly validated. However, the more intense the exercise, the more accurately it is recalled, meaning that the study’s use of a simple yes/no answer for frequent vigorous activity was more likely than other measures to capture what people really did.

To untangle the effects of physical activity on cognitive health from all the other possible factors that might affect it — age, gender, marital status, living arrangements, working status, tobacco and alcohol consumption and general physical health — the researchers used a statistical technique called propensity score matching. Essentially, this method places people into groups according to criteria such as their age, gender, health status and so on, and then calculates the probability that people exercise in each of these groups. Similar probability scores are then matched to create two groups that are similar except only for whether or not they exercise. In this way, any difference in cognitive functioning between the groups could be reasonably put down to whether or not the people in those groups exercised.

The results were clear across the board. For both men and women, people who enjoyed frequent physical exercise had higher cognitive functioning scores than their counterparts who did not exercise frequently.

As Mr Muhammad, the corresponding author of the paper, summarises, “What we show is the independent association between physical activity and cognitive function. Considering all the effects of these other factors - lifestyle factors, mental health, physical health – we are showing that this association is independent and statistically significant.”

As one might expect, fewer people in urban areas could exercise frequently compared to those living in rural areas.

“A large proportion of the [rural] population is engaged in farming.” as Mr Muhammad explains. “Poorer people, less educated people, do not have a profession in their old age; but even though they are retired, it is common for them to continue farming. Thirty-four per cent of the population is engaged in such physical activity.”

In the Indian context as a whole, certain activities require a different kind of physical dexterity than in the social context of other countries. Among health factors such as depression and physical health, the study also looked at activities of daily living: normal self-care activities such as feeding and bathing oneself.

When I asked Mr Muhammad about the relevance of these activities to exercise, he took the example of eating: “... people in [other] countries use a spoon or other cutlery, but in the Indian context, you need grip strength and manual dexterity. Such skills are needed for many ‘daily-living-activities’ tasks.”

It is easy to fall into the trap of moralising where exercise is concerned. Indeed, it is true that exercise in adulthood can, in principle, be done under a vast range of circumstances and is under conscious control. Exercise may be linked to cognitive health, but this study shows that many factors that allow or encourage exercise are partly or mostly beyond a person’s deliberate choice, such as whether one lives in a traffic-choked city or village. Other factors that affect a person’s cognitive health are entirely outside of a person’s control, as a study of gender differences showed. 

Childhood education reduces the gender gap in cognitive function in old age

While men and women enjoyed a similar boost in cognitive functioning from exercise, further research into the LASI dataset revealed persistent overall gender differences. A study examining gender differences in cognitive health was done by researchers at the University of South Alabama, the University of Southern California, the University of Michigan and the Ann Arbor Center for Clinical Management Research in the USA, and the International Institute for Population Sciences, Mumbai. The researchers quantified cognitive function differently than the study examining exercise and cognitive function. The study participants completed 53 different cognitive tests, ranging from simply knowing the date to naming lists of items. The researchers put these together to make a single cognitive factor score. They considered the effect of gender, age, childhood socioeconomic conditions, education level and societal gender inequality in the region of India a participant was from.

An overall difference in cognitive function existed between men and women, with women scoring lower. About 75% of the gender difference could be explained by the effect of early-life socioeconomic conditions and education level. The role of education, in particular, was a complex and interesting one. On the whole, women who had completed higher levels of schooling when they were children had higher cognitive functioning than women with lower levels of schooling. How much schooling was needed to close the gender gap in cognitive functioning as older adults depended on something else: the level of gender inequality at the society level. Gender inequality was measured by factors such as reproductive health and inequalities in empowerment and the labour market. In regions of India with lower levels of inequality (roughly the Southern, Central and Western regions), women’s cognitive functioning equaled men’s when they had completed schooling up to at least the middle school level. Indeed women’s cognitive functioning actually surpassed men’s if the women had studied further than middle school. In regions of India with higher levels of inequality (roughly the Northern and Eastern regions), the cognitive function gender gap only closed when the women had completed high school. In other words, an average of nine additional years of schooling was needed to close the gender gap in cognitive function in Indian societies with higher levels of gender inequality.

As Dr Jain, the first author of the paper on gender differences in cognition, says, “Education is like a lag measure. The past is affecting the current cognitive score.”

Given these results, we might say that the cognitive health of older adults today, in a broad statistical sense, reflects the social conditions from many decades ago. Yet, Dr Jain is clear that we must not rush to draw conclusions about cause and effect.

“We do not claim that these effects are causal. We are very careful in our paper that none of these analyses is causal – they are purely descriptive. We are saying this is the gender gap, as evidenced by this particular measure of cognitive health, and once you start [analysing] in which groups that gap is big or small, this is what you see.”

When it comes to complex social issues like the cognitive health of older adults, or the gender differences therein, it can be difficult – sometimes nearly impossible - to disentangle cause and effect. Large amounts of data are the first step to understanding the patterns in how physical and social conditions affect health, even if they must be interpreted cautiously.

These studies from the LASI dataset show that older adults that exercise more have better cognitive functioning and that men have better cognitive functioning than women, but this gender gap closes with more years of education and better early-life socioeconomic conditions. The gender gap in cognitive function is larger in societies with higher levels of gender inequality. 

We may never find a cure for ageing, but social policies informed by good data can certainly make it a more pleasant experience for an entire society. Studies like these are a good place to start. 


This article has been run past the researchers, whose work is covered, to ensure accuracy.

Editor's note: The article was updated to correct a minor error. The error is regretted.