2. Current challenges in modeling the social determinants of
health and aging call for comparative approaches
Socially and economically disadvantaged persons experience a higher
accumulation of risk factors (low income: Bor et al., 2017; Hirai et
al., 2012; migrants: Riosmena et al., 2014; racial minorities: Shah et
al., 2020; Wallace, 2015; Williams and Chiquita, 1995) and increased
mortality (Chetty et al., 2016; National Center for Health Statistics,
2016). Yet, common approaches to modeling health trajectories (e.g.,
growth curve models, latent class models) often report similar health
profiles between socially disadvantaged and advantaged groups (Brown et
al., 2012; Gueorguieva et al., 2009; Markides and Coreil, 1986) or
favorable health trajectories in traditionally marginalized groups
(i.e., Migrant Effect and the Hispanic Paradox; Franzini et al., 2001;
Markides and Rote, 2019; Quiñones et al., 2011). This divergence between
evidence of cumulative risk and increased mortality on the one hand, and
methodological approaches that model average health trajectories of
disadvantaged groups on the other, suggests that current forecasting
methods for predicting the progression of individual health may fail to
capture critical aspects of human sociality and thus not accurately
capture human health trajectories (Engelman and Jackson, 2019). Another
issue when modeling the social determinants of human health and aging
concerns limitations in handling missing data when this missing data is
not random. As individuals age and their health deteriorates,
longitudinal studies suffer non-random reductions in the number of
participants due to mortality and other sources of attrition (Vaupel,
2010). Thus, many studies of human health and aging may have biases in
that robust individuals will remain in the study into very old age,
whereas those who die earlier or have unknown fates will not be included
(Jackson et al., 2019). Finally, survey responses – a common method in
human studies – are influenced by many other factors such as the
personal perceptions of respondent and interviewers (Courtenay, 2000;
Davis et al., 2010; Dowd and Zajacova, 2010; Gunasekara et al., 2012;
Salazar, 1990; Sorlie et al., 1992; Williams and Chiquita, 1995) that
can limit our understanding of the causal mechanisms of individual
health and consequent aging. We argue that animal models for the social
dimensions of health and aging can therefore provide a new and valuable
opportunity to test novel biodemographic perspectives on analyzing
individual health and improve methods to forecast health over the life
course that can be applied to humans.