Lectures and seminars Sfo Epi Seminar in Biostatistics: Dr Nadine Binder

Zoom Link: https://ki-se.zoom.us/j/67418842539

*Webinar format: unless you are on the panel you will be muted and with no video. Participants can raise their hands during the Q&A to be recognized to ask questions.

Abstract

Some observational studies find that dementia incidence in Western nations is declining. A prominent recent example is Satizabal et al. (2016), who analysed Framingham Heart Study (FHS) data over four non-overlapping 5-year epochs ranging from 1977 (beginning of Epoch 1) to 2008 (end of Epoch 4) and seemingly found a linear decline over time. Recently, we challenged this finding on the basis that bias may have resulted from the failure to adequately account for potential disease onset in the period between last observation and death, a bias which we refer to as missing disease information due to death (MDID) bias. Re-analysing the FHS data using spline-based analytic methods permitting adjustment for the probability of disease onset following last observation, we did not find convincing evidence for a decline in dementia incidence over the four epochs (Binder et al 2019). 

Yet, our replication study, having otherwise closely adhered to the data structure of the original publication, still leaves room for improvement. First, we note that the classification of calendar time into 5-year epochs is both unnecessary and arbitrary; comparison of the survival curves for each epoch, as provided in Binder et al (2019), clearly shows that the conclusion of a linear decline in dementia incidence would not have held had e.g. a 4-year follow-up period been used. Second, two separate cohorts (the ‘original’ FHS cohort and a cohort of their offspring) were combined for analysis, which may be inappropriate if they differ markedly. 

Thus, a more suitable approach for analysing the question of how dementia incidence has evolved over time in the FHS may be to consider the two cohorts separately, and to dispense with the epoch structure by analysing age as the base time. This results in the analysis of two cohorts which are both ageing over time and being subject to death in greater numbers over time without replenishment from younger participants. If MDID bias is occurring, the effect of this would therefore be to underestimate the incidence of dementia cases to a greater extent over time. This problem requires the use of statistical methods based on the illness-death multi-state model, such as spline-based penalized likelihood as employed in our earlier study. 

In this talk, we will present so far unpublished findings of the new analysis strategy aiming to reevaluate dementia incidence over time in the Framingham Heart Study cohort. 

References 

Incidence of Dementia over Three Decades in the Framingham Heart Study.
Satizabal C, Beiser AS, Seshadri S
N. Engl. J. Med. 2016 07;375(1):93-4

A multi-state model based reanalysis of the Framingham Heart Study: Is dementia incidence really declining?
Binder N, Balmford J, Schumacher M
Eur. J. Epidemiol. 2019 Nov;34(11):1075-1083

If you have any questions, please contact Erin Gabriel at  Erin.Gabriel@ki.se