Alzheimer’s disease (AD) clinical trials typically employ an array of outcome measures in order to capture the high dimensionality of dementia. These outcomes are traditionally analysed separately, resulting in much useful clinical information being lost. For this reason, new analytical methods to recognize treatment responses, using techniques that better embrace the high dimensionality of the dementia syndrome, might prove useful.
This study aimed to ascertain the viability of detecting treatment signals using network analysis, a well-established set of techniques employed in a wide range of applications (e.g. ecology to stock markets) but to date little used in dementia clinical trials. Our specific objectives were to evaluate the feasibility and performance of analysing changes in the degree of network connectivity in recent clinical trials of Souvenaid®, a medical nutritional drink for the dietary management of mild AD.
Souvenir I and Souvenir II trials both showed significantly more connectivity in the treatment group compared to the placebo, with the difference being evident as early as 12 weeks. The increased connectivity in the Souvenir I and II studies suggest a more widespread treatment effect than when the measures are analysed individually. In this way, the network analysis approach allows overall treatment change to be demonstrated. By considering each treatment arm in a clinical trial as a network, the information from all outcome measures employed in the trial can be evaluated. Further work on translating this into clinically recognizable treatment effects is ongoing.