Prospective GERiatric Observational (ProGERO) study: cohort design and preliminary results

The demographic changes in Brazil as a result of population aging is one of the fastest in the world. The far-reaching new challenges that come with a large older population are particularly disquieting in low- and middle-income countries (LMICs). Longitudinal studies must be completed in LMICs to investigate the social and biological determinants of aging and the consequences of such demographic changes in their context. Therefore, we designed the Prospective GERiatric Observational (ProGERO) study, a longitudinal study of outpatient older adults in São Paulo, Brazil, to collect data both on aging and chronic diseases, and investigate characteristics associated with adverse outcomes in this population. The ProGERO study takes place in a geriatric outpatient clinic in the largest academic medical center in Latin America. We performed baseline health examinations in 2017 and will complete subsequent in-person visits every 3 years when new participants will also be recruited. We will use periodic telephone interviews to collect information on the outcomes of interest between in-person visits. The baseline evaluation included data on demographics, medical history, physical examination, and comprehensive geriatric assessment (CGA; including multimorbidity, medications, social support, functional status, cognition, depressive symptoms, nutritional status, pain assessment, frailty, gait speed, handgrip strength, and chair-stands test). We used a previously validated CGA-based model to rank participants according to mortality risk (low, medium, high). Our selected outcomes were falls, disability, health services utilization (emergency room visits and hospital admissions), institutionalization, and death. We will follow participants for at least 10 years. We included 1336 participants with a mean age of 82 ± 8 years old. Overall, 70% were women, 31% were frail, and 43% had a Charlson comorbidity index score ≥ 3. According to our CGA-based model, the incidence of death in 1 year varied significantly across categories (low-risk = 0.6%; medium-risk = 7.4%; high-risk = 17.5%; P