Network Analysis of Key Instrumental Activities of Daily Living and Cognitive Domains for Targeted Intervention in US Older Adults Without Dementia: Cross-Sectional Study

Abstract

Background: Cognitive impairment in older adults reduces independence and raises health care costs but can be mitigated through stimulating activities. Based on network theory, intricate relationships within and between clusters of instrumental activities of daily living (IADLs) and cognitive domains suggest the existence of central IADLs and cognitive domains, as well as bridge IADLs. Modifying these can significantly enhance daily living activities and cognitive functions holistically.

Objective: This study aims to identify central IADLs (key activities within the IADL network), central cognitive domains (key domains within the cognitive network), and bridge IADLs (linking IADL and cognitive networks). These insights will inform targeted interventions to effectively improve IADL and cognitive well-being in older adults.

Methods: A cross-sectional analysis of adults aged 65 years and older in the United States focused on 5 IADLs and 6 cognitive domains from the National Health and Aging Trends Study (NHATS). Network analysis identified central and bridge variables. Nonparametric and case-dropping bootstrap methods checked network stability. Network comparison tests assessed sex differences with Benjamini-Hochberg adjustments.

Results: Of the 2239 participants, 56.4% were female (n=976). We computed and tested 3 networks: IADL, cognition, and bridge-with correlation stability coefficients of 0.67, 0.75, and 0.44, respectively (all>0.25). Meal preparation was identified as the central IADL, with a centrality index of 3.87, which was significantly higher than that of other IADLs (all P<.05). Visual attention emerged as the central cognition domain, with a centrality index of 0.86, which was significantly higher than that of other cognition domains (all P<.05). Shopping was determined to be the bridge IADL, with a centrality index of 0.41, which was significantly higher than that of other IADLs (all P<.05). Notably, gender differences emerged in the IADL network, with stronger associations between laundry and meal preparation in females (1.69 vs males: 0.74; P=.001) and higher centrality in meal preparation among females (difference=1.99; P=.007).

Conclusions: While broad enhancements in all IADL and cognitive domains are beneficial, targeting meal preparation, visual attention, and shopping may leverage their within-network influence to yield a more pronounced improvement in holistic IADL, holistic cognition, and holistic cognition function through IADL interventions among older adults. Notably, meal preparation interventions may be less effective in males, requiring tailored approaches.

Keywords: network comparison; IADL; bridge variables; central variables; cognition function; cognitive domain; cognitive impairment; cognitive network; daily living activity; elder; elderly; holistic cognition; holistic cognition function; instrumental activities of daily living; intervention targets; network analysis; non-demented; older adults; stimulating activity.

Figure 1.. Network structure of instrumental activities of daily living (IADL) network, cognition network, and the bridge network.

Figure 2.. Expected influence centrality index and centrality bootstrapped difference tests for variables in the IADL network (Pane A and Panel B) and the cognition network (Panel C and Panel D). Gray boxes indicate variables that do not significantly differ from one-another. Black boxes represent variables that differ significantly from one another (α=.05). White boxes show the values of bridge expected influence. I1: difficulty in managing medication; I2: difficulty in managing laundry; I3: difficulty in managing shopping; I4: difficulty in managing meals; I5: difficulty in managing banking; C1: episodic memory; C2: executive function; C3: orientation; C4: psychomotor function; C5: visual attention; C6: working memory.

Figure 3.. Bridge expected influence centrality index (Panel A) and centrality bootstrapped difference tests (Panel B) for variables in the bridge network. Gray boxes indicate variables that do not significantly differ from one-another. Black boxes represent variables that differ significantly from one another (α=.05). White boxes show the values of bridge expected influence. I1: difficulty in managing medication; I2: difficulty in managing laundry; I3: difficulty in managing shopping; I4: difficulty in managing meals; I5: difficulty in managing banking; C1: episodic memory; C2: executive function; C3: orientation; C4: psychomotor function; C5: visual attention; C6: working memory.