Dementia: it takes a village

Dementia is a growing issue worldwide, due to an ageing population – but the dementia burden is not distributed equally among countries. According to Alzheimer's Disease International, the majority of people with dementia live in low- and middle-income countries and these numbers are expected to grow, whereas in the UK, the actual numbers of people with dementia were less than the projected numbers in recent years. With few medical specialists and funds, and loss of income for both the person living with dementia and the carer, poorer countries are ill-equipped to deal with the high economic and human costs of dementia.

One such country is Indonesia. We have been doing research there since 2005, in dementia and its associated risk and protective factors. I had worked as a neuropsychologist and epidemiologist on research projects in the USA and at Oxford and Cambridge University, and had just moved to Loughborough University to take up a Chair as Biological Psychologist in the Department of Human Sciences. Loughborough is very pragmatic in its approach: rather than being satisfied with model building, many of its researchers do work that is useful for the people they study. And when I arrived for the first time, my collaborator and 'Indonesian mum', Professor Tri Budi Rahardjo, asked me what I would do for the Indonesian people, apart from take their data.

So, my research became focused on what we could study that would actually be useful to treat cognitive decline and prevent dependency in old age. Dementia is brought on by disease in the brain and is characterised by this decline in cognitive functions, such as memory, planning and visual-motor skills and personality changes. This results in increasing dependency on others to perform more complex activities of daily life, including shopping, cooking, travel and banking.

What could we actually do for the Indonesian people to reduce its risk?

A doubling of dementia?

In those days, dementia was not really on the map in Indonesia. More pressing issues such as malaria and HIV were at the forefront of people's minds. The average age at death was in the 60s, similar to India, and while no direct numbers existed, prevalence of dementia was thought to be around 5 per cent in older people (again, similar to that estimated in India). Those who survived communicable disease were apparently made of stern stuff. 

However, some alarming numbers of people with dementia have recently been reported in Indonesia, with prevalence numbers between 21 per cent up to 33 per cent in people over 60. These current estimates – if correct – would be three times higher than the dementia prevalence seen in most other (including neighbouring) countries, which is between 7-10 per cent (Subramaniam et al., 2015).

Our initial estimates – based on Oxford University developed algorithms, assessed and validated by the excellent and knowledgeable Indonesian psychiatrists, Dr Irawati Ismael and Dr Fidiyanshah (Hogervorst et al., 2011a, b) – suggested that dementia prevalence numbers varied between 3 per cent (in urban affluent areas) to 16 per cent in the over 60s in deprived rural areas with an average older and poorer population. However, these estimates were obtained from cohort studies carried out by us now almost 20 years ago. Could there be a more than doubling in the percentage of Indonesian people over 60 who developed dementia in just a few decades?

There are several reasons why these recent dementia prevalence estimates may be so high. Firstly, risk and protective factors for dementia vary between regions (Hogervorst et al., 2021) and risk for dementia may have actually increased over the past decades. The people over 60 in 2006 were different from those who are recently over 60 years of age. The people at risk for dementia then were born in a world without antibiotics for childhood disease and would have had a very strong immune system to survive the first years of life. We suspect that a good immune system probably also plays a crucial role in later life, in dementia, like it does in cancer and probably heart disease.

Risk factors for dementia, including the most common type of dementia – Alzheimer's disease – are the same as those for cardio- and cerebro-vascular disease (see Hogervorst et al., 2023). This includes having had a stroke and diabetes, both mentioned as risk factors in some of the recent Indonesian studies. Whereas in our 2009 study diabetes prevalence was low, this could have increased due to more Westernised lifestyles, including sedentary behaviour, less engagement in physical activity and diets rich in glucose and bad fat (triglycerides, trans-fats) resulting in obesity.

However, protective factors, such as engaging in psychosocial and community activities, including exercise; reporting good subjective health; and not being underweight; were reported in all (also in the more recent) Indonesian cohorts. Living rurally, an older age, and low education (closely linked to being female in Indonesia) also remained important contributors to dementia risk. Not being married or having a job and being depressed came into the statistical models to predict dementia risk in the recent (but not 2006) cohorts.

These risk factors are all similar to the factors associated with dementia risk reported in Europe and higher income countries. So radically different known risk factors probably did not contribute to this at least doubled reported risk of dementia in Indonesia.

Kangkung and tofu

It may be that unmeasured factors in our and the more recent studies contributed to the increased risk of dementia – such as increased pollution and the possibly highly prevalent, but under-investigated, role of present infectious disease.

For instance, we previously found that consumption of green vegetables was associated with an increased dementia risk. This is in contrast to what was found in Western cohorts (Hogervorst, 2020), as green leafy vegetables contain folate, which should be a protective vitamin helping against dementia. However, in Indonesia, much of the green vegetable consumption on Java consists of the popular water spinach, Kangkung. Stir fried with garlic onion and soy or oyster sauce, it's very tasty and widely consumed. Kangkung grows abundantly in muddy soil on lands irrigated by rivers.

The rivers on Java are notoriously polluted, many being black and full of trash before they reach the ocean. President Jokowi recently pledged quicker persecution of polluting offenders. In addition, local advocates are trying to stop waste dumping in the rivers. But could the long-term industrial pollution of lands by river water have made the foods dependent on the that water (e.g. rice, Kangkung) dangerous to consumption?

Kangkung was investigated for its excellent ability to absorb pollutants including heavy metals. Heavy metals have also been found to accumulate in brains of people with dementia, and would be neurotoxic. This possible risk factor of eating polluted foods needs to be further investigated in Indonesia, as blanket advice on promoting green vegetable consumption – like we do in the UK – may not be warranted.

Another food which is usually thought of as very healthy is tofu. I had actually come to Indonesia to investigate benefits of tofu, as this soybean curd food is highly oestrogenic. Earlier research suggested that oestrogens can be good for brain health, but my work at Oxford and Cambridge suggested that high oestrogen levels and oestrogen treatment for older people did not benefit brain function. But while high tofu consumption also increased risk of dementia in Indonesian older people (older than 68 years of age, when eaten a couple of times a week), fermented whole soybean food (tempe) and fruit consumption reduced risk in dementia.

Local governments were quick to respond to these findings. Primary care centres (Puskesmas) advised older people who came to visit to eat more fruit and tempe, and provided exercise and other psychosocial programmes, which were also found to be protective against dementia. In one region we advised older people via the Puskesmas to eat more tempe, the traditional fermented whole soybean food, and to substitute this for tofu. When we re-assessed people after two years, those who had increased their tempe intake had reduced dementia risk (Hogervorst et al., 2011). In animal models we confirmed that tempe could protect the brain against dementia markers (Kridawati et al., 2016). These studies were funded by the British Council/Newton fund.

While some people in the media suggested we were sponsored by the meat industry, we were not the first to report that high tofu consumption could increase dementia risk. We replicated the findings in China and in Japanese older people in Hawi'I (Soni et al., 2016). Tofu can also increase risk for thyroid disorders, which were endemic in one area in particular: Borobudur, close to Yogyakarta. This poor and deprived rural area had the highest number of people with dementia, but also low iodine in the soil, according to local public heath researchers such as Professor Toni Sadjimim. This can lead to thyroid disease.

When we analysed whether thyroid function explained the detrimental association of tofu on brain function in Japanese older men in Hawi'i, this was not found to be the case there. We still need to test this in Borobudur. In Cambridge, only half of those older people whose thyroid levels we tested were aware that they had thyroid disease, but that did affect their mental function – including their rate of memory decline and dementia risk (Hogervorst et al., 2008). Worsening poor thyroid function in this area could explain how tofu could further increase dementia risk.

Multi-disciplinary medical examinations are thus needed, to also rule out other potentially treatable or contributing conditions to memory issues seen in possible reversable dementias. The reversable factors can include poor nutrition, as recent Indonesian cohorts reported that being underweight was a risk factor for dementia in Indonesia. A lack of B12 vitamin (mainly found in meats) and folate (found in green leafy vegetables, citrus fruit and liver) is associated with dementia. In addition, underlying infections (of the urinary tract, or bronchia) can reduce cognitive function, but these often reverse to normal after these infections are treated with broad-spectrum antibiotics. Covid-19 showed us that potentially infectious factors can increase risk for longer term cognitive impairment.

In the villages

Our recent ESRC-funded anthropological Indonesian investigators lived in the villages and followed older people and their families over time to investigate the nature of care systems of dependent older people. Cognitive testing was quickly abolished by our staff. Their feedback suggested that people with little education, not having been exposed to many 'school-like' testing situations, felt intimidated, humiliated or irritated by this formal questioning approach. According to our staff, the older people did not feel free to just stop testing, or be motivated to try and answer the questions and engage with the tests.

If cognitive tests are to be used, continued training of testers is very important, to ensure a standard protocol which reduces inter-tester variability. People need sufficient time to get to the answer. The attitude of the tester should be respectful and relaxed, making sure that performance is optimised, but without priming. Lastly, the environment should be quiet and private, without distraction and interference from helpful carers – not always easy to guarantee in the villages. In our previous studies our testers were trained and regularly supervised. Variation in data caused by inconsistent testing is notorious in surveys. In our multi-site US collaborations we could not analyse cognitive data from our computerised tests because even these had not been used correctly.

Another important factor relates to the feasibility of using Western tests for cognitive assessment in rural areas, where these are not culturally transferable, both in content as in form. For instance, some of these dementia assessment tests ask for identification and naming of objects, which are perhaps less commonly seen in Indonesian rural deprived areas (tutu, tassels, callipers, etc). Many of the tests used, such as the Mini-Mental State Examination, are heavily affected by education, with those who have obtained less education needing lower adapted cut-offs for dementia screening.

The adapted Hopkins Verbal Learning Test we used against a gold standard dementia assessment tool independent of education associated well with the dementia diagnoses, unlike the MMSE. The Newcastle-developed instrument for community-based Identification and Intervention for Dementia in Elderly Africans (IDEA) in Tanzania and other such instruments, may also offer better objective measurements, with better acceptability and user-friendliness.

From testing to care

Dementia diagnostics isn't just about objective testing, though. According to the diagnostic criteria, carers need to report on how much the cognitive impairments affect daily life and social interactions. If it does not do this, technically people do not have dementia, but instead would be diagnosed with minor neurocognitive disorder or mild cognitive impairment, a precursor stage of dementia. Our carer reports from the questionnaires showed that only half of people who scored below cognitive test-cut-off scores were thought by their carers to have issues with cognition and that this made them dependent. Combining data from the tests and the reports from carers of dependency led to an overall estimate of 6 per cent in the over 60s who had dementia. When visiting the rural areas, this was echoed by the village elders. Using a mapping exercise with the village elders and systematically asking about dementia symptoms in villagers suggested a similar prevalence of dementia as that seen in other European and other Western cohorts. 

Importantly, cognitive symptoms were not considered as important issues for the villagers. They were thought to be a normal sign of ageing. But beyond these cheerful gatherings discussing care with village elders, it remains to be seen whether actual care for people with dementia comes without issues. Recent generations are perhaps less patient, with many young people moving away, and older women in particular are left behind by themselves. These often-widowed women, with low education and living in poverty, are at highest risk for frailty and dementia. With increasing migration of the younger population to urban areas, how rural isolation will affect care for older people is a potential problem for the future. We can hopefully mitigate this by ensuring that young girls receive good education, and are able to take care of themselves financially over their lifetimes.

Photo: "Working with my then Indonesian PhD Dr Atik Kridawati who now heads the University of Respati: very proud of her work!"

Eef Hogervorst is a Professor of Psychology and Director for Dementia research at Loughborough University. Eef also holds visiting Chairs in Indonesia where she has conducted research since 2005 funded by the British Council/Newton Trust, ESRC and Alzheimer's Research UK. [email protected]

Key sources

Hogervorst, E., Temple, S., & O'Donnell, E. (2023). Sex Differences in Dementia. Current Topics in Behavioral Neurosciences. Springer, Berlin, Heidelberg.
Hogervorst, E. et al. (2021). Dementia and Dependency vs. Proxy Indicators of the Active Ageing Index in Indonesia. Intern J Environ Res and Public Health, 18(16), 8235.
Hogervorst, E. et al. (2011a). Validation of Two Short Dementia Screening Tests in Indonesia. In Jacobsen, S.R. (ed) Vascular Dementia: Risk Factors, Diagnosis and Treatment. NY: Nova Science Publishers, Inc.
Hogervorst, E., Sabarinah, B. et al .(2011b). Optimal Ageing, Dementia, Gender and Socioeconomic Status. In Ageing, Gender, Health and Productivity, p.60–75. UI Press, Jakarta.
Hogervorst, E., Mursjid, F., Priandini, D. et al. (2011). Borobudur revisited. Brain Research, epub Oct 28 2010, Mar 16, 1379, 206-12.
Hogervorst, E., Huppert, F., Matthews, F. & Brayne, C. (2008). Thyroid function and cognitive decline in the MRC Cognitive Function & Ageing Study. PNEC, 33(7), 1013-22.
Hogervorst, E. (2020). Vegetable, fruit and low to moderate alcohol intake is associated with better cognition in middle-aged and older Hispanics/Latinos. J Nutrition, 150(6), 1352–1353.
Kridawati, A., Hardinsyah, H., Sulaeman, A., Rahardjo, T.B.W. & Hogervorst, E. (2020). Tempe, Tofu, and Amyloid-β 1–40 Serum Levels in Ovariectomized Rats. JAD, 1-6.
Kridawati, A., Sulaeman, A., Damanik, R. et al. (2016). Tempe Reversed Effects of Ovariectomy on Brain Function in Rats: Effects of Age and Type of Soy Product. J steroid biochem mol biology, 160, 37-42.
Soni, M., White, L.R., Kridawati, A., Bandelow, S. & Hogervorst, E. (2016). Phytoestrogen consumption and risk for cognitive decline and dementia. J steroid biochem mol biology, 160, 67-77.
Subramaniam, M., Chong, S.A., Vaingankar, J.A., et al. (2015). Prevalence of Dementia in People Aged 60 Years and Above: Results from the WiSE Study.  Journal of Alzheimer's Disease, 45(4), 1127–1138.