8
Key Takeaways
In the final session of the workshop, Rebeca Wong (University of Texas Medical Branch at Galveston; workshop planning committee chair) invited each of the moderators to give a brief summary of key points from their sessions and from the workshop in general. Following these remarks, representatives of the Pan American Health Organization (PAHO) and National Institute on Aging (NIA) shared their perspectives on the workshop and on next steps. She then opened the floor to all workshop participants to share their thoughts and key takeaways.
SESSION KEY POINTS
The Role of Inequality
Socioeconomic positions, family arrangements, and environmental factors are all key determinants of health and well-being, said Yaohui Zhao (Peking University; workshop planning committee member). These are also the areas where public policies can make an impact; these actionable areas were the key focus of the workshop. Research in low- and middle-income countries (LMICs) is valuable in many ways; in particular, there are heterogeneous exposures that are useful for learning more about cause and effect. There is a need to pay more attention to external validity and transportability of findings, but these challenges can be met by collecting the right data and using the right method. There are several barriers to accessing high-quality data; these include lack of data infrastructure, lack of research capacity, changes in data sharing policies, and lack of funding for research.
Zhao identified several areas where there are opportunities for near-term action to improve research in LMICs. First, international collaboration is critical, both for sharing data and for working together on environmental issues that affect everyone. Second, there are new sources of data and new ways to leverage existing data. However, she noted that it can be difficult for users to assess the quality of data; she suggested that a role for NIA would be to assess sources of data and make the assessment publicly available. Third, there is technology for data sharing. Creating “data enclaves” in a central location, as well as in individual countries, can facilitate data sharing even with new regulations. Zhao emphasized the importance of collaborating in order to avoid duplication of efforts to create and improve data sharing technologies.
Conceptual and Methodological Barriers
One key issue, said Ayaga Bawah (University of Ghana; workshop planning committee member), is a need to collect new data, particularly longitudinal data, and combine these new data with existing data. Sources of existing data include surveys that are conducted in many of the LMICs, as well as other routine datasets. Combining data sources can allow for triangulation to get a better understanding of the issues that are being investigated. All three presenters in the session emphasized the need for training LMIC researchers, particularly in the areas of survey design, data management, data analysis, and longitudinal data analysis. Speakers also emphasized the need to build partnerships and work collaboratively in ways that include LMIC scientists in an inclusive manner. Finally, said Bawah, they all emphasized the need to recognize there is a lot of diversity within and among LMICs, and it is important to recognize the differences and to account for these in research efforts.
Research and Policy Interventions
Since the rate of population aging in LMICs has accelerated in recent years, Wong said, speakers emphasized the importance of studying future older adults (i.e., younger people) in order to detect patterns that will affect them as older adults. For example, high rates of obesity and diabetes may have serious consequences as these populations age. There is a need for longitudinal studies that collect data both before and after policies are implemented in order to measure the impact of policies. Speakers mentioned that there are policy databases in higher-income countries and that similar efforts would be very valuable for LMICs. Harmonization of data was called out as being important, but speakers also noted that it is essential to pay attention to context-specific variables and to collect data that are
important, even if they cannot be harmonized with other data. Studies need to be designed in order to capture variability in exposure to the intervention (e.g., a new policy), and to enable cross-national comparisons when appropriate.
The Role of Family
In looking at LMICs, Emily Agree (Johns Hopkins University; workshop planning committee member) first stressed the importance of understanding the centrality and reliance on family caregiving for older adults with chronic illness or disability, as well as home-based care work. Within this area, she said, the speakers gave a lot of attention to understanding family caregivers and how caregiving affects their health, labor force participation, and other outcomes. It is important to look not just at individual caregivers, but also to understand care networks and how both family and nonfamily members work together to take care of older family members. In addition, family resources are an important factor when considering how unmet needs may lead to further disability and poor health.
Second, said Agree, living arrangements and social isolation are significant factors in understanding the well-being of older people in LMICs. However, there is a need for more investigation because a lot of research is imbued with assumptions about what specific living arrangements mean and who benefits from them. There is a need to dive deeper in order to understand the meaning and the value of multigenerational households, of living alone, and other living arrangements, and to identify the nuances that really reflect families’ experiences.
Use of Existing Data
Wong said that the speakers emphasized conducting more research on life-course trajectories because there are limited data in this area. There was also a major emphasis on leveraging existing studies by adding aging content. Engaging with existing studies—whether longitudinal or not—presents a lower-cost opportunity to efficiently increase research on aging in LMICs.
Role of Environmental Exposures
LMICs have the greatest vulnerability to the effects of climate change, climate shocks, and environmental exposures, and older adults are among the most vulnerable within LMICs, said Mary Ganguli (University of Pittsburgh; workshop planning committee member). There are multiple levels and mechanisms through which climate factors affect health, and how factors related to aging are key to the increased vulnerability of older adults.
There are major risks from household air pollution from traditional indoor cooking and heating methods, and how it contributes to poor health in LMICs, in addition to the effects of climate change and outdoor pollution. Speakers also discussed the synergistic relationship between environmental hazards, exposure, and vulnerabilities and described opportunities and challenges for adaptation and mitigation.
PAN AMERICAN HEALTH ORGANIZATION PERSPECTIVE
Patricia Morsch (Pan American Health Organization) first offered a brief overview of the work of PAHO. PAHO works as two organizations, she said, serving as both the health organization of the inter-American system and as the regional office for the World Health Organization (WHO) in the Americas. As part of this work, PAHO provides technical coordination for the countries in the Americas on public health, working directly with ministries of health.
Morsch explained that she works in the Healthy Aging Program, the main focus of which is implementing the United Nations Decade of Healthy Aging. There are four main areas of action, she said, and these areas are very closely related to the topics discussed at the workshop:
- protecting older adults’ rights and combatting all forms of ageism;
- promoting age-friendly environments;
- creating or providing integrated person-centered care; and
- long-term care.
The Decade of Healthy Aging plan acknowledges that in order to work in these areas of action, there is a need for “enablers.” These enablers include increasing data information and innovation, building capacity, and increasing leadership and collaboration. All of these important ideas were discussed at the workshop, said Morsch. PAHO collaborates with many partner organizations in order to use evidence-based information to support countries in guidance, policies, and practices.
NATIONAL INSTITUTE ON AGING PERSPECTIVE
Minki Chatterji (NIA) said that the workshop was a “colossal [...] task,” but it was well carried out. She said that the presentations and discussion will be enormously helpful to NIA as it builds a portfolio over the next decade to address aging.
Chatterji identified key takeaways from the workshop, divided into five categories: inequality, family, environment, data needs, and training. There is a lot of inequality not just across but also within countries, and social
gradients in health are “not going to go away.” What is needed, she said, is to look for policies that can mitigate the unequal negative effects on health. The research agenda at NIA needs to include research on how different policy environments can affect inequality. In terms of family structure, living arrangements, and aging, said Chatterji, it seems that a mixed-methods approach might be necessary for learning more about how family and aging interact. It is important to determine what sort of data are needed and to employ both quantitative and qualitative methods in order to capture these data in a comprehensive, nuanced way.
One particular area that seems ripe for further research is the grandparent–grandchild relationship. In the aging world, people sometimes “fall into this trap” of thinking of older adults as people who need to be taken care of, but we also need to think about older adults as people with purpose who are capable of taking care of others.
For environmental research, there is a tension between collecting the data necessary to intervene appropriately (e.g., collecting data on specific disease outbreaks, specific pollution issues) while protecting people’s privacy and complying with data regulations. Chatterji said this is an issue that she is “concerned about” and that needs to be addressed.
In terms of data needs, there are 3 main approaches that could be taken: (1) start Harmonized Cognitive Assessment Protocol (HCAP) studies in new countries, (2) initiate studies in locations where data are lacking, and (3) leverage cohorts that were not originally designed to study aging. Each has its benefits and drawbacks, and all should be considered as ways to build the evidence base on aging in LMICs. Finally, she said, training was mentioned multiple times at the workshop. It is clear that there is a need for building capacity in the field and working within LMICs to train and support researchers. Chatterji closed by asking workshop participants to give feedback on what she had identified as key takeaways and to add their thoughts and perspectives on the research agenda going forward.
GENERAL DISCUSSION
Measuring Gradients
As Chatterji mentioned, social gradients in health are important to measure and understand, said Jacquiline Avila (University of Massachusetts Boston). She emphasized the importance of looking at how patterns change in gradients with each new wave of research, as this will provide even more information on the impact of interventions on inequalities. Nikki Sudharsanan (Technical University of Munich) agreed and said that it is important to measure both the absolute level of an outcome (e.g., mortality) and to look at the inequalities between groups and the gradient.
Need for Multiple Sources and Methods
The modern causal inference toolkit sets a high bar for researchers, said Sudharsanan. In research on policies, it can be very difficult to determine causation, which presents an interesting challenge when seeking funding for projects. A quasi-experimental research study designed to prospectively evaluate the impact of a policy can fail for multiple reasons outside the researcher’s control, he said. The standard funding grant application to look at the effect of policies on health is inherently challenging because of the unpredictability of whether quasi-experiments would work.
Lindsay Kobayashi (University of Michigan) agreed and said that researchers need to think of quasi-experimental methods as one tool in a broad toolkit. Researchers need to consider creative methods and sources of data to evaluate policies, including observational data and linked data sources. Having a publicly available source of data on policies in order to enhance data linkages would be enormously helpful for informing robust, strong research questions and contextually grounding the work, she said.
Feinian Chen (Johns Hopkins University) agreed that mixed-methods research is necessary in this area, particularly in examining family relationships. Survey data can only “get you so far” when investigating quality of relationships and other within-family dynamics. Similarly, common tools such as time-use diaries can capture how people use their time, but they may not capture the entire experience of a caregiver’s emotional and cognitive labor.
Opportunities for Research
Morsch suggested two areas in which she sees opportunities for future research. First, PAHO and WHO use the term “healthy aging,” which is the ability to maintain the capacities and abilities to do what you value, she said. This definition incorporates both the environment of a person and their intrinsic capacities (e.g., cognition, locomotion). There is a group at WHO now working on measurements for healthy aging; the concepts are very broad and difficult to measure. More work is needed to determine how to best measure concepts such as “ability to participate,” “having purpose,” and “opportunities to grow.” Second, said Morsch, PAHO uses a life-course perspective when looking at inequalities. She explained that health is something built over time as a person builds on their opportunities and capacities. Inequalities early in life can have dramatic consequences for well-being later in life; for example, people who do not have the opportunity to get an education early in life are less likely to reach their full capacity for health. Looking at inequalities in aging through a life-course perspective is an area that is ripe for future research.
Importance of Nationally Representative Data
There is a tradeoff between using data that are nationally representative and data that are fit for the purpose you need, said Hans Peter Kohler (University of Pennsylvania). All things being equal, it would be great to always have nationally representative data, but these data are more expensive and burdensome to collect. In the LMIC context, said Kohler, there is a need to build on existing studies or cohorts that are useful for answering specific questions about aging, while also making systematic investments in nationally representative data.
Training
Chatterji asked for feedback on training researchers and what methods might work best to strengthen infrastructure in LMICs. Will Dow (University of California, Berkeley) responded that it is “super important” to consider how and where researchers will receive training. He told participants about a program at Berkeley that supports fellows from East Africa. The fellows come to Berkeley for a short time, followed by trainings in East Africa. There is a continued back-and-forth, he said, rather than a “one-and-done” approach. Supporting people over a longer time period is important, but it is unclear how this might fit into a typical National Institutes of Health funding mechanism. Agree added that one training model that can work well is to develop training around specific data. This is a way of bringing people together around research questions and creating community, which can be very beneficial for all involved. This could also be helpful for NIA, she said, in terms of locating training around data that NIA are supporting.
Strengthening Research Infrastructure
Several workshop participants offered ideas on how to strengthen research infrastructure:
- Form partnerships among studies or surveys (e.g., the Health and Retirement Survey International Family of Studies) in order to agree on common data elements and harmonization to facilitate cross-national comparative work or intracountry work on key questions. (Lis Nielsen, NIA)
- Revisit the idea of leveraging existing studies, even if it requires modifying or enhancing the datasets. (John Phillips, NIA)
- Find ways to share data while conforming to country- and region-specific data privacy laws. (Phillips)
- Link aging surveys to studies conducted earlier in life in order to capture information about early life (e.g., family history) to examine nuances and variation in effects. (Agree)
Ensure Sustainability of Data Collection
“Anybody can collect one wave” of data, said Wong, but the question is whether data collections can be sustained so that they are useful. When researchers start data collection in a new country or area, it is critical to think about how it will be continued. Sustainability is a major benefit of partnering with existing longitudinal surveys that are already launched, she said.
In adjourning the workshop, Wong thanked the organizers of the workshop, the moderators, speakers, and participants.
