Identifying symptom-specific intervention targets for multidimensional frailty in community-dwelling older adults with chronic diseases: A network-based simulation approach.
TL;DR
Balance difficulty and memory problems may represent candidate symptom-prioritized targets for frailty management among older adults with chronic diseases and require validation in contemporary longitudinal cohorts and intervention studies.
Abstract
Background
Evidence remains limited regarding the symptom network structure of multidimensional frailty and simulation-based intervention targets among community-dwelling older adults with chronic diseases.
Methods
This cross-sectional study included 1478 community-dwelling adults aged 60 years and older with chronic diseases who were recruited from 22 communities in Jinan, China, between August 2015 and December 2016. Multidimensional frailty symptoms were measured using the 15-item Tilburg Frailty Indicator. The Ising model was used to estimate the frailty symptom network, and the expected influence was calculated to identify central symptoms. The NodeIdentifyR algorithm was used to simulate alleviating and aggravating interventions and estimate changes in network sum scores. Network accuracy and stability were assessed, and exploratory subgroup analyses were conducted by gender and age.
Results
This study included 1478 participants (67.39% women, mean age = 69.41 years); 59.81% were classified as frail. Difficulty in maintaining balance, difficulty in walking, and feeling down had the strongest expected influence. The expected influence index showed good stability, with a stability coefficient of 0.75. In the simulation analyses, alleviating problems with memory produced the largest projected reduction in the network sum score, from 5.74 to 4.88. Aggravating difficulty in maintaining balance produced the largest projected increase, from 5.76 to 7.29.
Conclusions
Balance difficulty and memory problems may represent candidate symptom-prioritized targets for frailty management among older adults with chronic diseases. Given the cross-sectional, simulation-based design and the earlier data collection period, these findings should be interpreted as hypothesis-generating and require validation in contemporary longitudinal cohorts and intervention studies.