Machine Learning Modeling of Low Quality Survey Responses to Predict Cognitive Impairment Among Older Adults
2024

Using Machine Learning to Predict Cognitive Impairment in Older Adults

Sample size: 12942 publication Evidence: moderate

Author Information

Author(s): Gao Hongxin, Schneider Stefan, Hernandez Raymond, Harris Jenny, Maupin Danny, Junghaenel Doerte U, Jin Haomiao

Primary Institution: University of Surrey, University of Southern California

Hypothesis

Can machine learning models predict cognitive impairment based on low-quality survey responses among older adults?

Conclusion

The study developed a machine learning model that can identify cognitive impairment in older adults using low-quality survey responses.

Supporting Evidence

  • The model achieved an AUC of 0.66 for identifying current cognitive impairment.
  • The model achieved an AUC of 0.70 for predicting dementia or mortality in the next 10 years.
  • The predictive performance was better in the 50-59 age group (AUC=0.72).
  • The predictive performance was also strong in the 60-69 age group (AUC=0.71).

Takeaway

Researchers created a tool that helps find older people who might have memory problems by looking at how they answer surveys, even if those surveys aren't about memory.

Methodology

The study used machine learning and psychometric methods to analyze survey responses from older adults.

Participant Demographics

Participants were aged 50 and above, with a focus on those aged 50-59 and 60-69.

Digital Object Identifier (DOI)

10.1093/geroni/igae098.1011

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication