Characterizing symptom clusters for older adults with urinary incontinence in home healthcare.

BACKGROUND

In home healthcare, limited understanding of symptom clusters among older adults with urinary incontinence may perpetuate disjointed assessments and negative outcomes.

OBJECTIVES

We aimed to identify symptom clusters and analyze differences by sociodemographic and clinical factors.

METHODS

We analyzed home healthcare episode data for adults ≥ 65 years old with urinary incontinence from the largest agency in the Northeastern United States. Symptoms were extracted from clinical notes using natural language processing. Symptom clusters were defined as distinct symptom experiences rather than requiring a specific number of symptoms. We employed hierarchical clustering, an unsupervised, data-driven method that iteratively identifies naturally occurring symptom subgroups without constraining the number of symptoms required. We used Chi-square tests to examine differences by sociodemographic and clinical factors, with the control group comprised of no symptoms.

RESULTS

Among 39,179 episodes (mean age 84, 74 % female, 61 % White), five clusters emerged: 1) anxiety (13.8 %), 2) all symptoms (5.8 %), 3) dizziness-anxiety (5.2 %), 4) constipation-anxiety-dizziness (3.7 %), and 5) no symptoms (71.5 %, control group). Compared to the no symptoms cluster, the anxiety cluster had more females and depression, all symptoms cluster had more cardiovascular disease, and dizziness-anxiety cluster had more fall history (all p < .001).

CONCLUSION

Comprehensively addressing symptom clusters is crucial for improving care. Future studies should examine symptom relationships and associations with health outcomes.