Gerontologic Biostatistics: Time to Consider Best Practices?

• Activity Number: 95
• Type: Roundtables
• Date/Time: Monday, August 1, 2011 : 7:00 AM to 8:15 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Terrence E. Murphy*+
• Companies: Yale University School of Medicine
• Keywords: gerontologic biostatistics ; , non-informative censoring ; competing events ; joint models ; best practices

Abstract

The subdiscipline known as gerontologic biostatistics is centered around a systematic and comprehensive treatment of the special statistical challenges that arise when conducting research with older persons. These statistical challenges include the design and analysis of multicomponent interventions, the simultaneous modeling of multiple outcomes, and consideration for the timing and occurrence of the comorbidities that inevitably accompany the aging process of older persons. In this year's roundtable we pose the idea of whether it would be helpful to start categorizing different analytical approaches into a system of "best practices", a practice commonly used in industry. For purposes of discussion, we will examine the hierarchy of analytical preferences when performing longitudinal modeling of gerontological outcomes. These include simple survival modeling that assumes non-informative censoring, the use of competing risk models, and the application of joint modeling techniques to account for the predictable loss of followup due to death. Attendees will be asked to name an area of analysis within the subdiscipline where the establishment of "best practices" might be useful.