Skip to main content
. 2026 Jul 10;14:1880890. doi: 10.3389/fpubh.2026.1880890

Table 4.

The relationship between PSQI-based sleep quality and fatigue among men and women aged 60 + years in this study (N = 5,484).

Characteristics PSQI-based sleep qualityb OR (95%CI) for experiencing fatiguea
% (n/N) of physical fatiguec Model 1† Model 2‡ % (n/N) of mental fatiguec Model 1† Model 2‡
Overall Poor 70.7 (863/1221) 1 1 62.6 (764/1221) 1 1
Good 55.6 (2,370/4263) 0.52 (0.45, 0.60) 0.61 (0.53, 0.70) 47.8 (2039/4263) 0.55 (0.48, 0.63) 0.65 (0.56, 0.74)
Age (years) 60–69 Poor 66.9 (421/629) 1 1 58.2 (366/629) 1 1
Good 54.0 (1,364/2528) 0.58 (0.48, 0.70) 0.63 (0.52, 0.77) 46.8 (1,183/2528) 0.63 (0.53, 0.75) 0.72 (0.59, 0.86)
70–79 Poor 71.0 (279/393) 1 1 64.6 (254/393) 1 1
Good 56.9 (750/1318) 0.54 (0.42, 0.69) 0.61 (0.47, 0.79) 48.1 (634/1318) 0.51 (0.40, 0.64) 0.57 (0.45, 0.73)
80+ Poor 81.9 (163/199) 1 1 72.4 (144/199) 1 1
Good 61.4 (256/417) 0.35 (0.23, 0.53) 0.46 (0.29, 0.72) 53.2 (222/417) 0.44 (0.30, 0.63) 0.58 (0.39, 0.86)
Gender Men Poor 70.4 (376/534) 1 1 62.9 (336/534) 1 1
Good 54.8 (1,180/2153) 0.51 (0.42, 0.63) 0.61 (0.50, 0.76) 47.4 (1,020/2153) 0.53 (0.44, 0.65) 0.66 (0.53, 0.81)
Women Poor 70.9 (487/687) 1 1 62.3 (428/687) 1 1
Good 56.4 (1,190/2110) 0.53 (0.44, 0.64) 0.60 (0.49, 0.73) 48.3 (1,019/2110) 0.57 (0.47, 0.67) 0.64 (0.53, 0.78)
Area Urban Poor 68.9 (348/505) 1 1 62.6 (316/505) 1 1
Good 55.6 (912/1639) 0.57 (0.46, 0.70) 0.66 (0.52, 0.82) 49.4 (809/1639) 0.58 (0.48, 0.72) 0.70 (0.56, 0.87)
Sub-urban Poor 71.9 (515/716) 1 1 62.6 (448/716) 1 1
Good 55.6 (1,458/2624) 0.49 (0.41, 0.58) 0.57 (0.47, 0.69) 46.9 (1,230/2624) 0.53 (0.45, 0.63) 0.62 (0.52, 0.74)

a: OR = odds ratio, CI = confidence interval. b: PSQI-based sleep quality was categorized as “good (PSQI ≤7)” or “poor (PSQI >7).” c: Fatigue was assessed using Pittsburgh Fatigability Scale (PFS), with PFS≥15 and 13 indicating high physical and mental fatigability, respectively. † Model 1 was an unadjusted mixed-effect logistic regression model with PSQI sleep quality as the single predictor and adjustment for neighborhood-level clustering effects. ‡ Model 2 was a multivariate mixed-effect logistics regression model with adjustment for age (where applicable), gender (where applicable), educational attainment, marital status, occupation, body weight status, smoking, drinking, physical activity, meat consumption, vegetable consumption, self-reported histories of main chronic diseases (diabetes, hypertension, COPD, stroke, coronary heart disease, kidney disease, cancer, abnormal lipid profile), frailty, depression, and neighborhood-level clustering effects in addition to those included in Model 1.