Mode
Text Size
Log in / Sign up

Machine learning predicts recreational therapy engagement in veteransVeterans at Risk of Isolation Can Be Found Early

AI-generated summary of the cited source, checked by automated accuracy review. How we work

Key Takeaway
Consider using machine learning to identify veterans at risk of low recreational therapy engagement, but validate findings in larger samples.

This cross-sectional observational study included 57 veterans aged 18 years and above from a single New York State Veterans Home. Researchers developed machine learning models (Random Forest, Decision Tree, Gradient Boosting, Logistic Regression) to predict recreational therapy engagement, defined as high participation or any participation. For high participation, Random Forest achieved F1-scores of 0.860 ± 0.347; for any participation, balanced accuracy was 0.619 ± 0.081. Feature importance analysis revealed activity preference diversity (Gini importance: 0.293) and total preferences (0.254) as primary predictors for high participation, while facility tenure (0.268) was strongest for any participation. Veterans with preference diversity >4.5 activities and satisfaction scores >3.84 had a 100% observed probability of sustained high participation (n=5; 95% CI: 47.8% to 100%). Safety and tolerability were not reported. A key limitation is the small subgroup size for the 100% probability finding, warranting cautious interpretation. Clinically, these predictive models could help identify veterans at risk of low participation early in residency, enabling tailored engagement strategies.

Veterans at Risk of Isolation Can Be Found Early

New tools help staff spot veterans who might stop joining activities before it happens.

Imagine walking into a new home for the first time. You have to learn the rules and find your way around. It feels overwhelming. Now imagine you are a veteran who has served your country. You arrive at a long-term care facility. You expect to find friends and hobbies. Instead, you might feel alone.

This feeling of isolation is common. Many veterans struggle to find their place in a new environment. They might miss group events or stop going to activities they once loved. Staff members want to help. They want everyone to feel connected. But they often do not know who is at risk until it is too late.

But here is the twist. Researchers have found a way to see these risks coming. They used special computer programs to look at past data. These programs can spot patterns that humans might miss. The goal was simple. Find the veterans who need help before they stop participating.

How does this work. Think of a factory assembly line. Each worker checks a specific part. The computer acts like a smart inspector. It looks at many different pieces of information at once. It checks how many hobbies a person likes. It checks how long they have lived at the facility. It also looks at what kind of activities they prefer.

The study looked at fifty-seven veterans. They were all at one specific home in New York. The team asked them to fill out a survey. The survey asked about their age and their interests. It also asked if they liked group games or quiet spiritual events. The researchers built two types of prediction models. One predicted high participation. The other predicted any level of participation.

They tested five different computer methods. Random Forest was the best performer. It correctly identified most veterans who would join activities. It also found those who would stay away. The computer learned that variety matters. Veterans who liked many different types of activities were more likely to show up.

This doesn't mean this treatment is available yet.

The findings show that time matters too. New residents who had been there less than one and a half years were at higher risk. They had not yet adapted to the new routine. Their list of preferred activities was also smaller. This combination made them less likely to join in.

Group activities were another big factor. Veterans who enjoyed doing things with others were more engaged. Spiritual activities also played a role. These findings give staff a clear map. They can see which factors drive engagement. They can focus their energy where it counts most.

What does this mean for you. If you are a veteran or a caregiver, this is good news. It means staff can be proactive. They can reach out to new arrivals quickly. They can offer a wider range of hobbies. They can encourage group events to build community. This approach helps prevent loneliness before it starts.

Of course, there are limits to what we know. This study only looked at one facility. The group was small. The computer model worked well for this specific place. But we need to test it elsewhere. We need to see if it works in different states and different homes.

The next steps are clear. Researchers will try to use these models in more places. They will see if the predictions hold up. If they do, staff can use these tools during admission. They can create personalized plans for every new resident. This ensures no one falls through the cracks.

The path forward is hopeful. Technology can help us care for veterans better. It gives us the power to act early. We can build a community where everyone feels welcome. We can make sure every veteran finds their place.

Study Details

Study typeCohort
EvidenceLevel 3
PublishedApr 2026
View Original Abstract ↓
There is limited understanding regarding the factors that predict recreational therapy engagement among veterans in long-term care facilities. We aimed to develop and validate machine learning models to predict recreational therapy participation and identify key factors influencing engagement patterns among veterans in long-term care facilities. In this cross-sectional observational study, we used data from 57 veterans aged 18 years and above at the New York State Veterans Home at Oxford. Data were collected through a comprehensive self-administered survey capturing demographic characteristics, participation patterns, and activity preferences. Two binary outcome variables were constructed: high participation and any participation. Five machine learning algorithms (Random Forest, Decision Tree, Gradient Boosting, and Logistic Regression with L1 and L2 regularization) were systematically evaluated using Leave-One-Out Cross-Validation for high participation and 5-fold Stratified Cross-Validation for any participation. Feature selection was implemented using SelectKBest with f_classif scoring, and class imbalance was addressed through balanced weighting techniques. Random Forest emerged as the optimal algorithm for both prediction tasks, achieving F1-scores of 0.860 ± 0.347 for high participation prediction and balanced accuracy of 0.619 ± 0.081 for any participation prediction. Feature importance analysis revealed activity preference diversity (Gini importance: 0.293) and total preferences (0.254) as the primary predictors of high participation, while facility tenure (0.268) was the strongest predictor of any participation. Veterans with preference diversity >4.5 activities combined with satisfaction scores >3.84 achieved 100% observed probability of sustained high participation [n = 5; 95% exact binomial CI: (47.8%, 100%)], though this estimate should be interpreted cautiously given the small subgroup size. New residents (≤1.5 years) with limited preferences demonstrated the highest risk for non-participation. Group activities (Gini importance: 0.143) and spiritual activities (Gini importance: 0.100–0.101) emerged as significant predictors across both models. This research provides the first proof-of-concept demonstration of a machine learning approach for predicting recreational therapy engagement among veterans in long-term care facilities, establishing methodological feasibility and generating testable hypotheses for prospective multi-site validation. Activity preference diversity and facility tenure serve as primary determinants of participation, with a critical 1.5-year adaptation period identified for intervention targeting. These predictive models can be applied during admission or early in residency to identify veterans at risk of low participation, enabling recreational therapy staff to implement tailored, proactive engagement strategies before disengagement occurs.
Free Newsletter

Clinical research that matters. Delivered to your inbox.

Join thousands of clinicians and researchers. No spam, unsubscribe anytime.