Analyzing evidence-based falls prevention data with significant missing information using variable selection after multiple imputation

Effectiveness of evidence-based fall prevention programs to reduce loneliness in the United States

Introduction

Falls are associated with activity limitations and injuries among older adults. An estimated 25% of older adults fall each year, and over 40% of older adults report they are lonely. Small group, evidence-based fall prevention programs are widely available in the United States and may be a strategy to improve social connectedness within our aging population. The purpose of this study was to identify the effectiveness of evidence-based fall prevention programs to reduce loneliness among older adults. Administration for Community Living (ACL) grantee data were collected in a national repository.

Methods

Data were analyzed from 12,944 participants across 12 fall prevention programs (e.g., A Matter of Balance, Stepping On, Tai Ji Quan, Otago Exercise Program, Bingocize) between January 2021 and July 2023. To assess loneliness, participants were asked, "how often do you feel lonely or isolated?" The response choices for this single 5-point item ranged from "never" to "always." A linear mixed-effects multivariable regression, with program type included as a random effect, was fitted to assess changes in loneliness before and after fall prevention workshops. The model controlled for program type and delivery site type as well as participants' age, sex, ethnicity, race, education, living alone, number of chronic conditions, number of falls in the three months preceding baseline, and workshop delivery site type and attendance.

Results

Significant reductions in loneliness scores were observed from baseline to post-workshop (p < 0.001), which were more pronounced among participants with more frequent baseline loneliness (p < 0.001). Participants who attended more workshop sessions reported reduced loneliness at post-workshop (p = 0.028). From baseline to post-workshop, loneliness increased among participants who lived alone (p < 0.001) and reported two or more falls in the three months preceding baseline (p =0.002). From baseline to post-workshop, compared to White participants, increased loneliness was observed among Black (p = 0.040), and Asian (p < 0.001) participants. Participants with more chronic conditions reported more loneliness from baseline to post-workshop (p = 0.004). Relative to participants who attended workshops at senior centers, increased loneliness was observed among participants who attended workshops at residential facilities (p = 0.034) and educational institutions (p = 0.035).

Discussion

Findings expand our understanding about the benefits of small-group fall prevention workshops to reduce loneliness among older participants. Results suggest that disease profiles, living alone, fall history, and workshop location (and attendee dynamic) may impede social connection among some participants. Beyond small group activities, purposive strategies should be embedded within fall prevention programs to foster meaningful interactions and a sense of belonging between participants. Other social connection programs, services, and resources may complement fall prevention programming to reduce loneliness.