N. Sirven's research while affiliated with Paris Descartes, CPSC and other places

Population aging increases the need to better understand the determinants of hospitalization in order to improve the quality of services. However, usual health measures often prove to be poorly predictive of hospitalization. The aim of this research is to determine whether physiological health measures designed for older adults, namely the Short Physical Performance Battery (SPPB) (Guralnik et al., 1994), frailty phenotypes (Fried et al., 2001), and profiles of functional decline (Lunney et al., 2003), are better predictors of hospital use. We use longitudinal data from the International Mobility in Aging Study (IMIAS), carried out between 2012 and 2014 in four countries (Canada, Brazil, Colombia, Albania) among individuals aged 65–74 at baseline (n=1724). Differences between health systems provide additional insights into the determinants of hospitalization. Health profiles from 2012, besides other confounders, are used to explain hospitalization in 2014. SPPB scores are computed using objective measures of gait speed, chair-stands, and balance. Frailty phenotype variables are created according to Fried’s classification into robust, pre-frail, and frail. Four profiles of functional decline are created: (1) terminal illness – cancer and at least one ADL disability; (2) organ failure – heart or lung disease and at least 2 ADL disabilities, (3) low reserve – stroke or more than 2 ADL disabilities, and (4) other respondents. Sample attrition is corrected by means of a Heckman selection Probit model. Our results indicate that a score below 8 on SPPB, the pre-frailty phenotype, and a profile of organ failure are significantly associated with hospitalization. Cross-country differences are discussed.

Although frailty is a well-known concept in the health science, its operational definition is still debated. From a diversity of models, two emerged in the early 2000 from epidemiological studies conducted in large population-based aging cohorts: the frailty phenotype (Fried 2001), and the frailty index (Rockwood 1994) measuring the accumulation of deficits and death. These tools have been applied as screening instruments in clinical settings to guide individual decision-making and orient treatments. New interrogations are raised by the use of instruments developed to screen frailty in epidemiological research for assessing individual situations. Open questions remain on the feasibility of frailty screening, the properties of screening tools, the relevance of an integration of socioeconomic dimensions in screening tools, and the effectiveness of interventions targeting frailty. This article presents an overview of current perspectives and issues around frailty screening in populations and in individuals.