The Frailty Concept

Abstract

An individual said to be frail if he or she is much more susceptible (exposed or infected) to adverse events than others. Clayton (1978) introduced the term frailty to indicate that different individuals are at risk even though on the surface they may appear to be quite similar with respect to measurable attributes such as age, gender, weight, etc. He used the term frailty to represent an unobservable random effect shared by subjects with similar (unmeasured) risks in the analysis of mortality rates. A random effect describes excess risk or frailty for distinct categories, such as individuals or families, over and above any measured covariates. Thus, random effect or frailty models have been introduced into the statistical literature in an attempt to account for the existence of unmeasured attributes such as genotype that do introduce heterogeneity into a study population. It is recognized that individuals in the same family are more similar than the individuals in different families because they share similar genes and similar environment. Thus, frailty or random effect models try to account for correlations within groups. Such groups arise naturally in those studies involving two or more failure times on the same individual subject. These failure times may be times to the recurrence of the same event or times to the occurrence of different types of events. Examples include the time sequences of asthmatic attacks, infection episodes, tumor diagnosis, tumor recurrences or bleeding incidents in individual subjects (Prentice et al. 1981).