Left-, Right-, and Interval-bivariate-censored Data: Evaluating Screening Mammography in the Prescence of Lead-time Bias, Length Bias, and Over-detection

• Activity Number: 302
• Type: Contributed
• Date/Time: Wednesday, August 11, 2004 : 8:30 AM to 10:20 AM
• Sponsor: WNAR
• Author(s): Jonathan D. Mahnken*+ and Wenyaw Chan and Daniel Freeman and Jean Freeman
• Companies: University of Kansas Medical Center and University of Texas and University of Texas Medical Branch and University of Texas Medical Branch
• Keywords: bivariate survival ; gamma frailty model ; censored data ; Monte Carlo integration ; screening ; mammography

Abstract

Of the large clinical trials evaluating screening mammography efficacy, none included women ages 75 and older. Recommendations on an upper age limit at which to discontinue screening are based on indirect evidence and are not consistent. Screening mammography was evaluated using observational data from the SEER-Medicare linked database. Measuring the benefit of screening mammography is difficult due to the impact of lead-time bias, length bias, and over-detection. A new method was developed to address this problem. The underlying conceptual model divided the disease into two stages; pre-clinical (T0) and symptomatic (T1) breast cancer. Treating time in these phases as a pair of dependent bivariate observations, (t0,t1), estimates were derived to describe the distribution of this random vector. Censored information was attained about (t0,t1) from each observation. After adding mild assumptions, the likelihood function was approximated using Monte Carlo integration and used to find MLEs and an estimate of the variance-covariance matrix. This method was also studied using simulations to ensure valid parameter estimates are attainable using this new approach.