Model for predicting which older people will need home care services

Disease and treatment 27. feb 2021 3 min Professor Rudi GJ Westendorp Written by Kristian Sjøgren

Researchers have developed a mathematical model for predicting which older people in Denmark will need municipal home care services. The General Data Protection Regulation (GDPR) hinders implementation.

Interested in Disease and treatment? We can keep you updated for free.

There was a time when the only criterion for being eligible for municipal home care services in Denmark was getting a birthday cake with 65 candles.

That was a while back, and today Denmark’s municipalities are only required to assess people’s need for home care services when they turn 70 years old if they live alone and when they turn 75 and 80 years regardless of their living situation. An altered demographic and fiscal situation determines these changes. However, this also means that many people younger than 80 years may need help but are not offered it.

Researchers from the University of Copenhagen have now developed a mathematical model that uses older people’s personal history to calculate who will probably need home care services and who will manage without them. Home care services can include nursing visits and help with shopping or practical tasks, such as cleaning and personal care, including help in taking medication.

The model has the potential to very precisely allocate municipal home care services to the people who need them most.

“Many people older than 65, 70 or even 80 years can easily manage on their own without municipal services in Denmark. But many people younger than 80 years have great difficulty in managing everyday life without help. The strange thing is that people are legally entitled to home care services if they need it, but we lack the tools to identify who these people are,” explains a researcher involved in developing the new model, Rudi Westendorp, Professor, Center for Healthy Aging, Department of Public Health, University of Copenhagen.

The research, with Sasmita Kusumastuti playing a key role, has been published in the Journal of the Royal Statistical Society: Series A (Statistics in Society).

Model assesses risk based on a person’s overall historical data

The starting-point for developing the new model was a desire to be able to identify the people with the greatest need for home care services.

The research was carried out in collaboration with the City of Copenhagen, which has many residents 65–80 years old but the City does not know who really needs a helping hand with their daily lives.

The researchers therefore developed a mathematical model that uses personal data to predict who is most likely to need home care services.

The model draws on data from several registries in Denmark, including data on age, hospital admissions and medical history, partner and family status, medication intake, length of education, job situation and much more.

The model assesses a person’s data history from the time they are born and into old age and then relatively accurately predicts their need for home care services.

“Using our model, we can reasonably accurately assess who will need home care services within the next year. We were surprised that the model is so accurate given how chaotic life can sometimes appear,” says Rudi Westendorp.

GDPR hinders implementation

So far so good.

But why is this model not already being implemented if it can help Denmark’s municipalities to identify older people who need to be contacted and may then need home care services?

The GDPR presents a challenge.

The model uses very sensitive personal data, and therefore the City of Copenhagen cannot use them arbitrarily and then contact people who have neither asked to have their data examined nor to be contacted.

“This would be a misuse of the public registries and the legal and ethical restrictions. So we have developed a tool that can benefit older people in Denmark and save municipal resources, but we cannot use it,” explains Rudi Westendorp.

People can grant access to personally sensitive data

However, this problem can be overcome, and the researchers are working on trying to find ways to use the model without compromising ethical rules or the data protection legislation.

One option is getting people to use the NemID common log-in system to consent to and authorize the municipality to apply the model to their personal data and to contact them if the model indicates that they are probably entitled to or will probably need home care services.

These data may not be used for other purposes.

“This procedure must be kept at arm’s length from the public sector, and individuals will have full control over their personal data. We can predict people’s needs if they choose to give permission for their data to be used. We have the algorithm and people can give us the data, and this is one way to implement the algorithm in the public sector,” says Rudi Westendorp.

Expanding the model nationwide

Rudi Westendorp is sure that the algorithm can be implemented very quickly if the final hurdles can be overcome.

The researchers want to improve the model, which is already good but can be better at making predictions.

Then the researchers want to implement the model in the City of Copenhagen, and if this proves successful, the model can be expanded nationwide.

“Ideally, we need to demonstrate that the model not only identifies the people who need home care services but also enables them to get these services. Society will benefit enormously from this,” concludes Rudi Westendorp.

Personalised need of care in an ageing society: the making of a prediction tool based on register data” has been published in Journal of the Royal Statistical Society: Series A (Statistics in Society). In 2018, the Novo Nordisk Foundation awarded a Challenge Programme grant to Rudi Westendorp for the project Harnessing the Power of Big Data to Address the Societal Challenge of Ageing.

The past century has witnessed a revolution. Less than a hundred years ago, the average Western life expectancy was 40; now it is 80. And there is no...

English
© All rights reserved, Sciencenews 2020