The innovative Health Assessment Tool (HAT) optimally predicts mortality and hospitalizations in older adults

Development of the HAT in the older population
We live in an era where the world’s population ages unrepentantly. This demographic shift presents a significant challenge for healthcare systems worldwide. There is a growing need for a comprehensive health assessment of older adults to ensure they receive appropriate care and support to live their lives agreeably. To address this challenge, researchers from the Aging Research Center (ARC) at Karolinska Institutet (KI) have developed the HAT, which integrates five key health indicators – walking speed, cognition, chronic diseases, and basic and instrumental activities of daily living. Information from these indicators is summarized into a health status score ranging from 0 to 10, with higher scores indicating better health. Since HAT was developed in an urban, wealthy area of Stockholm, this study aimed to validate it in three other Swedish cohorts of older adults aged 60+ years living in suburban, rural, and urban areas across the country.

Image by freepik

The HAT optimally predicts mortality and hospitalizations
The HAT was tested in four aging cohorts from the Swedish National study on Aging and Care – Kungsholmen (SNAC-K), Skåne (SNAC-S), Blekinge (SNAC-B), and Nordanstig (SNAC-N). The results showed that the HAT has high predictive accuracy in all geographical cohorts for critical health outcomes such as mortality and hospital admissions. More specifically, it was highly predictive of mortality over 16 years, as well as for 1-year, 3-year, and 5-year mortality. Moreover, its predictive capacity was high for 1-year and 3-year unplanned hospitalizations.

As a clinically intuitive and externally valid tool, the HAT holds promise for better healthcare planning and personalized decision-making for older patients. By comprehensively and meaningfully assessing older adults’ health, the HAT can guide healthcare professionals in making informed decisions about the care and support each patient needs.

Ahmad Abbadi, the first author, and Amaia Calderon Larrañaga, the last author of the study. Photographers: Gunilla Sonnebring and Stefan Zimmerman.

Congratulations Ahmad on your results! What was it like working with NEAR data?
I got to work with data from four different SNAC sites, which was an excellent opportunity to learn about different aging cohorts. Thanks to the extensive support from NEAR data managers, I could focus mainly on data analysis, even if I was also actively involved in the data harmonization process. It was a great experience.

Most unexpected research finding:
That would be two findings. The first is how well the HAT performed across all sites, and how predictably it performed based on cohort size (better predictive capacity with increased sample size). The second is how well the harmonized dataset worked. In the early stages, we had doubts about whether it would outperform the predictability of the HAT across individual sites because of the heterogeneity between cohort sites. However, this was the case for all outcomes and stratified analyses. Moreover, although the area under the curve (AUC) in the harmonized dataset was slightly lower than in the development cohort (i.e., SNAC-K), it surpassed the predictive capacity of all comparable published tools for mortality and hospitalizations.

Best tips for working with NEAR data:
Taking the time to be familiar with the datasets and reading about the methods of data collection (of the different sites) can help you have a realistic expectation of which variables are available. For example, we did not have gait speed (walking speed) at baseline for SNAC-B. We had to do statistical imputations to overcome that. However, knowing that from the start helped us have a clear strategy for working with those data. Also, the NEAR team provided great support; whenever additional data was needed, the team stepped in to diligently facilitate this. Therefore, I would add having good communication with the NEAR team as an additional consideration.

Best tips for improving overall health:
Understanding that health in old age is complex and multifaceted requires interventions and tools that capture this complexity. As a clinician doing research, I cannot take my clinical lenses off. I use them frequently when thinking about the research I am doing and how it can help patients and people in the community. Generally, I think research needs to bridge clinical gaps and have clear clinical applicability. I think that the HAT is an excellent example of this philosophy; it is clinically intuitive, easy to use, can capture the complex and diverse health spectrum of older adults, and can predict varied outcomes that aid in clinical decision-making. 

Best tips to regain focus:
Having a clear goal in mind and remembering that our research is meant to benefit people. It was also helpful to understand how the research I was doing fitted into the larger scheme of work being done at ARC, and more specifically in the team of Amaia Calderón-Larrañaga (my supervisor). I would also add that having a supportive supervisor (who mentors you) helps keep you motivated, and I was blessed to work with Amaia on this project because she is the perfect embodiment of a supervisor who is likewise a mentor and an enabler.


Abbadi A, Kokoroskos E, Stamets M. et al. Validation of the Health Assessment Tool (HAT) based on four aging cohorts from the Swedish National study on Aging and CareBMC Med. 2024; 22 (236)