Unraveling disease patterns in older adults starting long-term care in Japan and their future health outcomes

As the population continues to age, improving the quality and efficiency of health care services for older adults has become increasingly important. However, the older adult population is diverse, often suffering from multiple diseases and various combinations of diseases.

Developing appropriate intervention methods for this diverse demographic is challenging. Using unsupervised machine learning techniques, researchers classified individuals aged 65 years or older who had started using long-term care into clinical types based on 22 diseases. The work is published in the journal Scientific Reports.

The individuals were newly enrolled in long-term care insurance services in Tsukuba City in Ibaraki Prefecture and Sammu City in Chiba Prefecture. The researchers examined the association between the classification (called the “clinical subtype”) and prognosis by clinical subtype.

Six clinical subtypes were identified in Tsukuba City:

1. Musculoskeletal and sensory diseases
2. Cardiac diseases
3. Neurological diseases
4. Respiratory diseases and cancers
5. Insulin-dependent diabetes
6. Others.

The same classification was reproduced after analyzing the data from Sammu City.

In terms of prognosis, cardiac disease, respiratory disease/cancer, and insulin-dependent diabetes were found to incur a higher mortality risk than musculoskeletal and sensory diseases. Furthermore, cardiac disease, respiratory disease/cancer, and others were associated with a worsening care-need level.

The findings of this study are relevant not only to individuals requiring care but also to their families and staff(s) involved in their care. The research and development of optimal interventions for each clinical subtype identified in this study could also influence health care policy.