Abstract
Background/Objectives: Depressive symptoms, malnutrition and falls are common geriatric syndromes with shared risk factors, yet they are usually studied separately. We aimed to identify latent subgroups defined by their co-occurrence among older adults in Northeast Thailand and the factors associated with membership. Methods: In this cross-sectional survey (October–December 2025), 12,115 community-dwelling adults aged ≥60 years in 39 subdistricts of Maha Sarakham Province were interviewed. Latent class indicators were the Thai 9-item depression scale, Mini Nutritional Assessment Short-Form, calf circumference, recent falls and gait impairment. Associated factors and concurrent correlates were estimated with bias-adjusted three-step models and cluster-robust standard errors. Results: Depressive symptoms affected 8.4%, malnutrition risk 24.5% and fall risk 17.0%; all three co-occurred in 2.2%, 6.2 times more often than expected under independence, mainly reflecting pairwise associations. Three classes emerged: robust (77.3%), nutritional risk (16.2%) and complex multidomain (6.5%). The complex multidomain class featured gait impairment (78.2%) and malnutrition risk (70.2%), with elevated depressive symptoms (40.9%) and recent falls (26.8%). Membership was associated with age ≥80 years (adjusted odds ratio 4.39, 95% CI 2.68–7.18), not working (4.23, 2.59–6.90), high household debt (2.45, 1.29–4.65), ≥2 chronic conditions (3.54, 2.23–5.64), polypharmacy (2.45, 1.73–3.47) and per additional home hazard (1.44, 1.29–1.61). Members more often rated their health as worse than peers (38.6% vs. 1.8% in the robust class). Conclusions: An estimated 6.5% of older adults (5.0–8.5% across closely agreeing specifications) formed a subgroup characterised by mobility impairment, malnutrition risk and socioeconomic disadvantage. Interpreting existing screening results jointly may help target integrated primary care, social welfare and home-modification support; prospective evaluation is needed.

This publication has 32 references indexed in Scilit: