Importance of modelling decisions on estimating trajectories of depressive symptoms and co-morbid conditions in older adults: Longitudinal studies from ten European countries

Abstract

Background International comparisons of trajectories of depressive symptoms in older adults are scarce and longitudinal associations with co-morbid conditions not fully understood. Objective To compare trajectories of depressive symptoms from participants living in 10 European Countries and identify ages at which the associations of co-morbid conditions with these trajectories become more relevant. Methods Latent growth curve models were fitted to depressive symptoms scores from participants of the Survey of Health and Retirement in Europe (SHARE) initiative (combined n = 21,253) and co-morbid conditions modelled as time varying covariates. To identify the ages at which the association between co-morbid conditions and depressive symptoms was significant the Johnson-Neyman (JN) technique was used. Results The shape of depressive symptoms trajectories varied between countries, and was highly dependent on modelling decisions. The association between the average number of co-morbidities reported over time and depressive symptoms was consistent and positive across countries and ages. Conclusion International differences in ageing-related trajectories of depressive symptoms emerged. The longitudinal association of co-morbid conditions with trajectories of depressive symptoms was found, but the results overall suggest that modelling decisions could greatly influence the outcomes, and should thus be interpreted with caution.

The effect of multimorbidity on changes in health-related quality of life following hip and knee arthroplasty

“…Fourteen out of eighteen studies carried out in a community setting found that multimorbidity was positively related to depressive symptoms over time. For example, the largest study (Marroig et al, 2019) The results showed a significant positive association between multimorbidity across time and the average number of depressive symptoms at ages 60, 65 and 70 in all countries studied. One study from the US (Sutin et al, 2013) with a follow-up duration of 30 years investigated the trajectory of depressive symptoms and associated factors that may contribute to increases in depressive symptoms across adulthood.…”

“…, 2007Fauth et al, 2012;Stommel et al, 2004;Turuba et al, 2019;Wu et al, 2012) focused on older adults aged 65 and above and the rest on individuals aged 50 and above. The percentage of females ranged from 45.6% to 72%, although the proportion was unclear in four studies (Chou, 2008;Marroig et al, 2019;Roberts et al, 1997; (Table 4.1).…”

“…In the three studies, multimorbidity was measured with the Comorbidity Index (CMI) (Curyto et al, 1999), and the Charlson Comorbidity Index (CCI) (Eliyan et al, 2020;Sutin et al, 2013), respectively, which account for both the number of conditions and severity of the conditions. Depressive symptoms were measured by self-reported symptoms using the Centre for Epidemiologic Studies Depression (CES-D) scale in fifteen studies; the NSHAP Depressive Symptoms Measure (NDSM) in one study (Eliyan et al, 2020), the Geriatric Depression Scale in two studies (Feng et al, 2013;Oslin et al, 2002); the EURO-D symptom scale in one study (Marroig et al, 2019); and the 12 items adapted from Primary Care Evaluation of Mental Disorders (PRIME-MD) mood disorders section in one study (Roberts et al, 1997) (Table 4.1).…”

Coordinated data analysis: Knowledge accumulation in lifespan developmental psychology.

“…These issues are not always insurmountable, and indeed can be an asset when handled appropriately, but do require mindful decision making and transparency. It is also important to keep in mind that, with any analysis, including coordinated analysis, differences in modeling decisions will have potential impacts on the output obtained and the conclusions drawn (Marroig et al, 2019).…”