Incidence of frailty: a systematic review of scientific literature from a public health perspective.

INTRODUCTION: Because of the dynamic nature of frailty, prospective epidemiological data are essential to calibrate an adequate public health response. METHODS: A systematic review of literature on frailty incidence was conducted within the European Joint Action ADVANTAGE. RESULTS: Of the 6 studies included, only 3 were specifically aimed at estimating frailty incidence, and only 2 provided disaggregated results by at least gender. The mean follow-up length (1-22.2 years; median 5.1), sample size (74-6306 individuals), and age of participants (≥ 30-65) varied greatly across studies. The adoption of incidence proportions rather than rates further limited comparability of results. After removing one outlier, incidence ranged from 5% (follow-up 22.2 years; age ≥ 30) to 13% (follow-up 1 year, age ≥ 55). CONCLUSIONS: Well-designed prospective studies of frailty are necessary. To facilitate comparison across studies and over time, incidence should be estimated in person-time rate. Analyses of factors associated with the development of frailty are needed to identify high-risk groups.

Prevalence of frailty at population level in European ADVANTAGE Joint Action Member States: a systematic review and meta-analysis.

INTRODUCTION: Although frailty is common among community-dwelling older adults, its prevalence in Europe and how this varies between countries is unclear. METHODS: A systematic review and meta-analysis of literature on frailty prevalence in 22 European countries involved in the Joint Action ADVANTAGE was conducted. RESULTS: Sixty-two papers, representing 68 unique datasets were included. Meta-analysis showed an overall estimated frailty prevalence of 18% (95% confidence interval, CI, 15-21%). The prevalence in community (n = 53) vs non-community based studies (n = 15) was 12% (95% CI 10-15%) and 45% (95% CI 27-63%), respectively. Pooled prevalence in community studies adopting a physical phenotype was 12% (95% CI 10-14%, n = 45) vs 16% (95% CI 7-29%, n = 8) for all other definitions. Sub-analysis of a subgroup of studies assessed as high-quality (n = 47) gave a pooled estimate of 17% (95% CI 13-21%). CONCLUSIONS: The considerable and significant heterogeneity found warrants the development of common methodological approaches to provide accurate and comparable frailty prevalence estimates at population-level.