U-M and Kenyan Partners Discuss Gaps in Global Aging Research, Data and Policy

In Kenya's national health surveys, children are broken out by single years of age. Women and men of working age get their own detailed brackets. Everyone older than 50 gets one box: 50-plus. That gap was one of several data problems on the table in June when researchers and government officials from Kenya joined University of Michigan faculty at the Institute for Social Research for a roundtable on global health and population aging in low- and middle-income countries this summer. The discussion, hosted by the Center for Global Health Equity and the Longitudinal Study of Health and Ageing in Kenya, known as LOSHAK, brought together scholars and collaborators from Aga Khan University in Nairobi, Kenya's Ministry of Health, and the Kenya National Bureau of Statistics. 

"If you look at the Kenyan health statistics, statistics for the elderly are conspicuously absent," said Elias Nyaga, a statistician and demographer with the Kenya National Bureau of Statistics who leads its health statistics section and serves as the bureau's point person for LOSHAK. Conditions that affect mainly older adults, including dementia, go largely uncounted as a result. What little prevalence data exists comes from small samples, with almost nothing on the biological and social exposures, lifetime poverty, hard labor, or infectious disease, that shape how people age. 
Filling that gap is not just a matter of running more surveys. It requires rethinking what the surveys measure in the first place. Anthony Ngugi, a co-principal investigator for LOSHAK and professor of epidemiology and population health at Aga Khan University, where he chairs the Department of Population Health and serves as associate dean for research in the medical college, described what happened when his team tried to adapt a standard cognitive test, the clock drawing task, which is widely used to assess executive function. "When it's an analog clock, which is used in the West, you find that most people will not be able to do that," Ngugi said. "And that is not because they are cognitively impaired, but because they do not have any exposure to analog clocks." The problem, he explained, runs deeper than clock face familiarity. In Swahili time, the day is counted from sunrise, so what a Western clock marks as 7 a.m. becomes saa moja — literally "hour one" — because the day begins at 6 a.m. A participant may understand time perfectly within their own cultural frame and still be marked down by a researcher relying on an unadapted Western clock-drawing task. 

Without adapting the test, researchers would not be measuring executive function at all. They would be measuring familiarity with a Western object. Emily Briceño, a clinical associate professor at the University of Michigan Medical School and co-investigator on the Harmonized Cognitive Assessment Protocol, said the same hidden assumptions show up across nearly every standard cognitive tool, including tests that rely on paper, pencils and the expectation that people are used to working quickly under pressure. 

That tension between adapting tools locally and keeping them comparable across countries came up again and again. Ngugi argued it should not be treated as a problem to solve once and move past. "I don't think it's a bug that we need to fix," he said. "I think it's one of the features of working in comparative aging science." How much localization a study needs, he added, depends on the question being asked. A study comparing how a memory test performs across rural and urban Kenya can lean more toward localization. A study comparing memory across Kenya, India and South Africa needs a shared foundation first — harmonization across contexts is what makes the results meaningful. 

The roundtable kept returning to a different kind of foundation: partnership. LOSHAK grew out of conversations between faculty from University of Michigan and Aga Khan University that began six years ago, even before any formal agreement existed. Ngugi said the project's structure reflects that origin. Kenyan and Michigan investigators co-lead every dimension of the work, scientific and administrative, and government partners from the Kenya National Bureau of Statistics and the Ministry of Health contribute to weekly management meetings rather than joining only once a study is designed. "They are not just partners who are helping us implement the study," Ngugi said. "They have been involved from the very beginning, conceptualization and design." Several Kenyan colleagues have also traveled to Michigan to build data science skills, a piece Ngugi called central to making the partnership reciprocal rather than one directional. 

Lindsay Kobayashi, associate director for research at the Center for Global Health Equity and co-director of the Harmonized Cognitive Assessment Protocol (HCAP) International Network, pointed to what that kind of long term, regionally embedded data can reveal. Through a study in South Africa called HAALSI, built on a long running health surveillance platform in Agincourt, her team linked aging outcomes to a randomized cash transfer program that had ended years before HAALSI began sampling. Households that received the transfer saw slower memory decline and lower dementia risk over seven years of follow up. 

But when her team looked more closely, they found something else: older adults in those same households had worse mental health right after the payments stopped — a decline they believe was driven by the sudden loss of income, not by the cash transfer itself. "We saw really meaningful effects," Kobayashi said, "but it confounded us, because it went counter to what we think improved income should do for health." 

Technology drew similar caution. Akbar Waljee, director of the Center for Global Health Equity, said he sees artificial intelligence as one more tool layered onto an already overwhelming volume of administrative and health data, not a solution in itself. He pointed to Kenya's mobile money system, M-Pesa, as an example of how the country has leapfrogged infrastructure gaps before, and suggested similar leaps may be possible with health data. But he was quick to add a caveat: "We can't forget the human behind the advancement of technology, because that in itself can drive inequities." 

The discussion closed on a harder reality. Kenneth Langa, co-principal investigator of the Health and Retirement Study and the Harmonized Cognitive Assessment Protocol International Network, noted that shifting federal research priorities have made the funding climate for international aging research more uncertain than at any point in his career. Kobayashi described a new federal mechanism that requires a US based institution to lead international partnership grants, a structure that has created both new constraints and new room to design studies differently. Josh Ehrlich, the roundtable's moderator and a LOSHAK principal investigator who also serves as associate director of the Center for Global Health Equity, said conversations like this one, across institutions and across borders, are what keep him going. Closing the session, he said, "These kinds of conversations personally give me a lot of hope, and I am energized by our deep and genuine collaboration, as well as what it means for the future of health, data, and aging in Kenya and globally.”