Elsevier

Women's Health Issues

Volume 19, Issue 1, January–February 2009, Pages 45-51
Women's Health Issues

Article
Assessing Sleep During Pregnancy: A Study Across Two Time Points Examining the Pittsburgh Sleep Quality Index and Associations with Depressive Symptoms

https://doi.org/10.1016/j.whi.2008.10.004Get rights and content

Background and Purpose

Sleep quality seems to be an antecedent to depressive symptoms during pregnancy. We sought to 1) examine the psychometrics of the Pittsburgh Sleep Quality Index (PSQI) in pregnancy; 2) examine whether sleep quality predicted increases in depressive symptoms; and 3) compare PSQI scores across 3 or 2 levels of depressive symptoms.

Methods

Each of the 252 participants completed the Beck Depression Inventory (short form) and a sleep quality measure at mid and late pregnancy.

Results

PSQI total scores showed good internal consistency and construct validity. An improved model of the internal structure of the PSQI in pregnancy was found with 1 factor labeled Sleep Efficiency, a second labeled Night and Daytime Disturbances, and an Overall Sleep Quality component associated with, but separate from, both of these 2 factors. Although PSQI scores showed moderate stability over time, sleep disturbance scores increased in late pregnancy. Importantly, PSQI prospectively predicted increases in depressive symptoms.

Conclusions

Findings suggest that the PSQI is useful in pregnancy research. Findings also support the idea that sleep problems are prospective risk factors for increases in depressive symptoms during pregnancy. Practitioners are advised to screen for sleep quality during early pregnancy.

Introduction

Sleep problems are frequently reported by pregnant women (Dzaja et al.,, Lee and Gay, 2004, Schweiger, 1972), possibly because of the occurrence of pregnancy-related physical symptoms or discomforts (nausea, back pain, increased urinary frequency), hormonal changes, enlargement of the fetus, and/or shortness of breath (see Lee, 1998). There is now evidence that sleep quality earlier in pregnancy may contribute to the development of higher levels of depressive symptoms later in pregnancy (Skouteris, Germano, Wertheim, Paxton, & Milgrom, 2008). Similarly, sleep patterns in late pregnancy are associated with elevated symptoms of depression in the first few weeks post birth (Wilkie and Shapiro, 1992, Wolfson et al., 2003). If indeed sleep quality is an antecedent to depressive symptoms, as these studies suggest, it may be important to screen for and address sleep difficulties during pregnancy to prevent increases in depressive symptomatology at a life stage when women's well being is particularly important (Skouteris et al., 2008).

One measure of sleep quality that has established test–retest reliability and validity in nonpregnant samples is the Pittsburgh Sleep Quality Index (PSQI; Backhaus et al., 2002, Buysse et al., 1989, Carpenter and Andrykowski, 1998). The PSQI measures quality and patterns of sleep, including difficulties related to subjective overall sleep quality, latency, duration, and disturbance; habitual sleep efficiency; use of sleep medication; and daytime sleep dysfunction over the past month. To our knowledge, only 2 recent studies have used the PSQI to assess subjective sleep quality in pregnant women. With a small sample of pregnant women (n = 19), Okun, Hall, and Cussons-Read (2007) administered the PSQI at approximately 12, 24, and 36 weeks gestation and found that PSQI scores did not differ significantly across pregnancy. Whereas the psychometric properties of the PSQI were not evaluated in that research, they were in a study by Jomeen and Martin (2007) who administered the PSQI to a group of 148 women at a mean gestation of 14 weeks. Women classified as having minor/major depressive symptoms at this early pregnancy time point, based on concurrent Edinburgh Postnatal Depression Scale scores, had significantly poorer sleep quality, than women with no depressive symptoms in all PSQI scales with the exception of sleep disturbances (sleep medication data were not analyzed owing to zero means). Moreover, as expected, the PSQI subscales and global sleep quality component scores were associated with higher scores on the Edinburgh Postnatal Depression Scale but were not correlated with age (except for a positive correlation with the sleep duration subscale), revealing convergent and divergent validity, respectively. Internal consistency was acceptable using the 7 component subscales scores, but improved slightly from 0.73 to 0.76 after excluding the sleep medication subscale which few women endorsed.

Jomeen and Martin (2007) further examined the internal structure of the PSQI. Confirmatory factor analyses of several competing models suggested the best fitting model comprised 2 correlated factors. Factor 1 included overall sleep quality, sleep latency, sleep duration, and sleep efficiency subscales; Factor 2 included sleep disturbance and daytime dysfunction subscales, which reflected having problems sleeping for reasons such as feeling hot or cold, snoring, coughing, pain, night-time or early morning waking, and having difficulty staying awake and functioning properly during the day.

Changes over time in PSQI scores were not assessed by Jomeen and Martin (2007). Given prior reports that sleep becomes more disturbed in later pregnancy (Dzaja et al.,, Pien and Schwab, 2004), one would expect the PSQI to reflect this poorer sleep quality in late pregnancy. Although Okun et al. (2007) found that PSQI scores did not differ significantly across time points during pregnancy, the mean PSQI total score was noticeably higher in the third trimester compared with the first and second. Hence, it is possible that their small sample size (n = 19) precluded meaningful differences reaching significance statistically.

The aim of the current study was to extend on Jomeen and Martin's (2007) study with a larger sample size examining 2 time points during pregnancy (mean of 18 and 33 weeks gestation) instead of just one; to examine Jomeen and Martin's 2-factor model at 2 time points; to assess construct and predictive validity of the PSQI; and to examine the 15-week stability in PSQI scores. In relation to internal structure of the PSQI, we examined internal consistency and factor structure at 2 antenatal time points to confirm the component structure of the PSQI across this life stage. Construct validity was expected to be supported by moderate correlations between PSQI and depression scores. Validity of PSQI scores was also assessed by examining whether scores increased over the course of the pregnancy, as would be expected across extended time points (for a review of the literature see Dzaja et al., 2005; Pien & Schwab, 2004). Data collection at 2 time points also allowed stability of sleep problems to be examined over time, which has implications for test–retest reliability of PSQI scores. Given laboratory (Lee, 1998, Lee et al., 2000) and questionnaire (Field et al., 2007) findings of moderate stability in sleep profiles for women across pregnancy, we predicted significant moderate, but not high, correlations over an average 15-week period between the second and third trimesters. At the first time point, PSQI scores were also compared across 3 degrees of depressive symptoms (none, mild, and moderate/severe) instead of only 2 (none versus mild to severe as in Jomeen and Martin's study), enabling comparison of sleep difficulties across more varied degrees of severity in depressive symptomatology.

We also assessed whether sleep quality predicted increases in levels of depressive symptoms over and above baseline levels of depressive symptoms. There were 2 purposes of examining prediction of changes in depression scores, the first being to examine the predictive validity of PSQI scores in pregnancy, as noted. The second purpose was to confirm data suggesting that poor sleep quality is a risk factor for increases in depressive symptoms during pregnancy (Wilkie and Shapiro, 1992, Wolfson et al., 2003), which would have clinical implications. If the PSQI effectively predicts worsening of depressive symptoms, it could form part of a screening package for predicting antenatal depressive syndromes.

Section snippets

Participants

The sample consisted of 252 pregnant women ranging in age from 18 to 42 years with a mean age of 31.67 (standard deviation [SD], 4.55) years. The majority of women (83.7%) reported that their pregnancy was planned and 47.2% were primiparous. Most women were born in Australia (81.3%) and the majority (73.8%) had a tertiary degree; 40.1% reported an annual family income of >AUD$95,000 (approximately USD$85,500) and 16.7% reported earning a family income of <AUD$45,000 (approximately USD$38,100).

Internal consistency

Cronbach's alpha for the PSQI (items were the 7 component subscale scores) was .70 and .76 at T1 and T2, respectively, exceeding Kline's (2000) criterion, and sufficient for a 7-item scale. Removing the sleeping medications component (only 7 women at T1 and 14 at T2 had scores > 0) improved Cronbach's alpha to .72 and .78 for T1 and T2, respectively (item total rs from .30 to .72).

SEM: Test of a correlated 2-factor model for PSQI

The best-fitting model from Jomeen and Martin's (2007) study (the 2-factor correlated model; Factor 1 included

Conclusion

The current study supported good psychometrics of the PSQI in a sample of women in middle and late pregnancy and provided a series of advances over the only prior study to evaluate the psychometric properties of the PSQI in pregnancy (Jomeen & Martin, 2007). Although the full PSQI showed good overall internal consistency, our study suggested that the PSQI is best viewed as comprising 3 main correlated factors. Our results produced an improved model of the PSQI internal structure compared with

Acknowledgments

The authors thank the women who participated in this research.

Helen Skouteris is a developmental psychologist who is currently working as a Research Fellow in the School of Psychology at Deakin University exploring factors that contribute to childhood obesity.

References (25)

  • M.S. Schweiger

    Sleep disturbance in pregnancy

    American Journal of Obstetrics and Gynecology

    (1972)
  • H. Skouteris et al.

    A prospective study of factors that lead to body dissatisfaction during pregnancy

    Body Image

    (2005)
  • Cited by (126)

    • Longitudinal and prospective assessment of prenatal maternal sleep quality and associations with newborn hippocampal and amygdala volume

      2022, Developmental Cognitive Neuroscience
      Citation Excerpt :

      This intriguing finding indicates that a plausible consequence of larger hippocampal volume may be heightened susceptibility to the postnatal environment. Second, sleep disruption is comorbid with maternal physical and mental health including depression (Skouteris et al., 2009), stress (Bublitz et al., 2018), diet (Van Lee et al., 2017), and physical activity (Tan et al., 2020) all of which influence the fetal brain (Dufford et al., 2021). Future work could consider the cumulative and synergistic influence of such processes on the developing brain.

    • Prenatal yoga and excessive gestational weight gain: A review of evidence and potential mechanisms

      2022, Complementary Therapies in Clinical Practice
      Citation Excerpt :

      Pregnant women often experience poor sleep quality in the first trimester and report even worse sleep in the third trimester [40,41]. Poor sleep quality has been associated with an increased risk of preterm birth [42], cesarean section and longer labor duration [43], and has been reported to be a risk factor for depression in both the prenatal and postpartum periods [44–47]. Several studies, in non-pregnant adult populations, have investigated the impact of sleep quality on weight outcomes with findings supporting a strong link between sleep restriction and weight gain or the development of obesity [48–51].

    • CBT-I in pregnancy

      2022, Adapting Cognitive Behavioral Therapy for Insomnia
    • Objective sleep efficiency but not subjective sleep quality is associated with longitudinal risk of depression in pregnant women: A prospective observational cohort study

      2021, International Journal of Nursing Studies
      Citation Excerpt :

      Although sleep disturbance and depression often coexist with sleep disturbance long been considered a part of the symptomatology of depression, emerging evidence has suggested a more influential, causal role of sleep disturbances in the development and maintenance of mood disorders (Baglioni et al., 2011). Few longitudinal studies have suggested that sleep disturbances precede depression in pregnant women (Skouteris et al., 2008, Skouteris et al., 2009, Yu et al., 2017). Skouteris and colleagues first reported results of their surveys over 200 Australia women during pregnancy starting from 15 to 23 weeks gestation.

    View all citing articles on Scopus

    Helen Skouteris is a developmental psychologist who is currently working as a Research Fellow in the School of Psychology at Deakin University exploring factors that contribute to childhood obesity.

    Eleanor Wertheim, Personal Chair at La Trobe University, conducts research into risk factors and prevention of body concerns and disordered eating, as well as conflict resolution processes.

    Carmela Germano has been a research assistant for the La Trobe University pregnancy and body image project and is completing a masters degree at La Trobe University.

    Susan Paxton, Professor at La Trobe University, conducts research in prevention, risk factors, intervention for body image and eating disorders and is past President of the Australian and New Zealand Academy for Eating Disorders.

    Jeannette Milgrom, Professor at University of Melbourne and Director of the Parent-Infant Research Institute, Austin Health, is an expert in postnatal depression and in the development of innovative specialized interventions for parent-infant mental health.

    Funded by an Australian Research Council Discovery Grant (DP0557181).

    View full text