How Can Neural Networks Improve Elderly Care?

The provision of elderly care is vexing many countries across the developed world as the transition of the baby boomer generation into retirement is stretching public finances in a multitude of ways.

New research from the University of Helsinki ponders how neural networks can help policy makers better optimize the provision of care to the elderly.

Traditionally, healthcare providers use risk-adjustment models that are in turn based upon data from previous years, to help them allocate funds in an effective and equitable way.  Such models are used in a wide range of countries, from Germany to the United States.  The researchers believe their neural network-based approach has the potential to significantly improve upon these models however.

The aim of the new approach is to take better account of the vagaries of healthcare, including the greater than average care required by elderly people.  Indeed, the early results from the model suggest that not only can it better allocate care to those who need it, but also save providers several million dollars.

Efficient care

The model was trained using patient data from the Register of Primary Health Care Visits of THL, which consists of all out-patient visit data for every Finnish citizen over 65 years of age.  The data, which was pseudonymized, enabled the team to train a deep learning model on such a rich dataset for the first time.

When the model was tested, not only did it produce strong results, but it also did so without needing a huge dataset in order to produce them.  Indeed, despite having around 1/10 of the data available to other models, it was able to produce superior results.

“Our goal is not to put the model developed in this research into practice as such but to integrate features of deep learning models to existing models, combining the best sides of both. In the future, the goal is to make use of these models to support decision-making and allocate funds in a more reasonable way,” the researchers say.

Machine learning models have been used before in healthcare in areas such as predicting demand of emergency departments, so it’s no surprise that AI can be equally useful in helping manage resources in elderly care too.  It’s an approach that the Finnish team can easily be scaled up to be deployed across a wide range of patient groups across the country.

Facebooktwitterredditpinterestlinkedinmail