Simulating the multicausality of Alzheimer's disease with system dynamics

In Alzheimer's disease (AD), cognitive decline is driven by various interlinking causal factors. Systems thinking could help elucidate this multicausality and identify opportune intervention targets.

We developed a system dynamics model (SDM) of sporadic AD with 33 factors and 148 causal links calibrated with empirical data from two studies. We tested the SDM's validity by ranking intervention outcomes on 15 modifiable risk factors to two sets of 44 and 9 validation statements based on meta-analyses of observational data and randomized controlled trials, respectively.

The SDM answered 77% and 78% of the validation statements correctly. Sleep quality and depressive symptoms yielded the largest effects on cognitive decline with which they were connected through strong reinforcing feedback loops, including via phosphorylated tau burden.

SDMs can be constructed and validated to simulate interventions and gain insight into the relative contribution of mechanistic pathways.