You Might Also Like This: Programmers Seek to Streamline Person-Centered Care Questionnaire

With the goal of offering better person-centered care, nursing homes use the Preferences for Everyday Living Inventory (PELI-NH), a long questionnaire to assess resident preferences. This can be time-consuming for staff and potentially burdensome to residents. In a study that earned a 2019 silver Mather LifeWays Innovative Research on Aging Award, researchers aimed to develop a system that would reduce the number of questions asked of residents, while still collecting all relevant information. This system would use machine learning similar to that used by Netflix or Pandora to recommend content users might enjoy, based on their preferences for other movies or music.

The investigators recruited 255 residents from 28 nursing homes in the eastern United States to participate in interviews about their social, activity, and care preferences. The interviews consisted of 72 questions from a questionnaire commonly used to assess resident preferences.

The combined responses from all residents were then used to develop the recommender system, based on resident responses to a set of 16 core questions. The way this would work is to ask all residents the same 16 core questions, then the recommender system would suggest about 10 other questions to ask residents, based on their responses to the core questions. The remaining 46 questions would be less relevant to the resident and would not be asked.

In testing the effectiveness of the system, the researchers found that recommendations of additional questions matched actual resident preferences four out of five times. If nursing home staff were to use a system like this, they would only need to spend time assessing residents’ preferences for activities that are actually relevant. This type of system is still in its early stages and the researchers pointed out that other factors, such as regional or community characteristics, may influence preferences as well and would need to be incorporated into the system.


Want to keep up with recent research that’s relevant to aging services? Use the form below to subscribe to our monthly InvestigAge email.



Gannod GC, Abbott KM, Van Haitsma K, Martindale N, & Heppner A. A machine learning recommender system to tailor preference assessments to enhance person-centered care among nursing home residents. The Gerontologist (2019); 59(1): 167-176.


Mather Institute provides at-a-glance summaries of award-winning research from the past year, offering fresh, evidence-based ideas for senior living and aging services industries.

Download FREE Copy

    Add insight to your inbox

    Join our email list to receive information about the latest research from Mather Institute. Just complete the form below to subscribe.

    Thank you!

    You are now subscribed to the email list.
    A confirmation has been sent to the email you provided.

    Continue to Website Share with a Friend