Associations of Anxiety, Physical Activity, Parity, and Sleeping Position with Sleep Quality among Third-Trimester Pregnant Women in Palangka Raya, Indonesia

Poor sleep in late gestation has been associated with adverse maternal and neonatal outcomes. Psychological, behavioral, obstetric, and sleep-practice factors have usually been examined separately, leaving their relative contributions within a single population unclear. This study aimed to examine the associations of anxiety, physical activity, parity, and sleeping position with sleep quality among third-trimester pregnant women in Palangka Raya. A cross-sectional study was performed on a sample of 66 third-trimester pregnant women visiting a primary healthcare department in Palangka Raya, Central Kalimantan, Indonesia, who were recruited over a period of four weeks. The quality of sleep was evaluated by the Pittsburgh Sleep Quality Index (PSQI) that generated scores in the form of continuous global score, where higher scores suggest lower sleep quality. The level of anxiety was assessed via Hamilton Anxiety Rating Scale, while physical activity was calculated through Pregnancy Physical Activity Questionnaire as metabolic equivalent-hour per week. Simultaneously, parity and habitual position of sleep were assessed as dichotomous; the key analysis was completed with multiple linear regression model with all four factors put simultaneously with validity of coefficient estimates being examined by means of 5,000 bootstraping and BCa 95% confidence intervals, as well as through two variations as sensitivity analyses. On average, the PSQI score was 7.23±2.26, while 80.3% of the participants showed the score higher than 5. The four-factor model was significant, F (4,61) = 36.496, p<0.001, R²=0.705. Supine sleeping was independently associated with a 4.249-point higher PSQI score than side sleeping (p<0.001; BCa 95% CI 2.929 to 5.846) and remained so when parity was excluded (B=3.324; BCa 95% CI 2.617 to 4.084). Parity reached conventional significance (B=1.140, p=0.040). However, it was not robust to resampling (BCa 95% CI: −0.106 to 2.608), and its coefficient reversed sign when sleeping position was removed, indicating a suppression artifact due to the near-complete overlap between the two variables (Cramer's V=0.817). Anxiety and physical activity were associated with PSQI score bivariately but not after adjustment; excluding sleeping position rendered the anxiety coefficient significant and bootstrap-robust (B=0.086, p=0.009), indicating an association inseparable from sleeping position rather than absent. In this sample, habitual sleeping position was the only correlate robustly associated with sleep quality, supporting attention to sleeping practices in antenatal assessment; the magnitude was unusually large for observational data and requires replication.