Bayesian Longitudinal Plateau Model of Adult Grip Strength with Multiple Imputation for Missing Response and Covariates

• Activity Number: 440
• Type: Contributed
• Date/Time: Wednesday, August 5, 2009 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Ramzi W. Nahhas*+ and Stefan A. Czerwinski
• Companies: Lifespan Health Research Center and Lifespan Health Research Center
• Keywords: longitudinal data analysis ; Bayesian ; multiple imputation ; growth curve ; grip strength ; change-point

Abstract

Low muscle strength is related to frailty in later life. Higher peak strength and a slower rate of decline with aging can delay the onset of sarcopenia. The purposes of this study were to investigate the longitudinal pattern of change in grip strength during adulthood controlling for sex, body size, and secular trend, and to assess associations between the pattern of change and various risk factors. Serial data from the Fels Longitudinal Study were analyzed using a Bayesian mixed model. Multiple imputation was used for missing values of the response and covariates with Bayesian posterior inferences combined using multiple imputation rules. A cohort adjusted plateau model was fit estimating the rate of increase, plateau timing, peak strength, and rate of decline. For males, birth weight was significantly related to the adult trajectory. Supported by NIH grants HD012252, AR052147.