Original Articles
Psychometric evaluation of the pittsburgh sleep quality index

https://doi.org/10.1016/S0022-3999(97)00298-5Get rights and content

Abstract

The Pittsburgh Sleep Quality Index (PSQI) measures sleep quality and disturbance retrospectively over a 1-month period using self-reports. Although the PSQI has been used in a variety of populations, published psychometric data are limited. The goal of this study was to examine psychometric properties of the PSQI among four populations: bone marrow transplant patients (n=155); renal transplant patients (n=56); women with breast cancer (n=102); and women with benign breast problems (n=159). Results supported PSQI internal consistency reliability and construct validity. Cronbach’s alphas were 0.80 across groups and correlations between global and component scores were moderate to high. PSQI scores were moderately to highly correlated with measures of sleep quality and sleep problems, and poorly correlated with unrelated constructs. Individuals with sleep problems, poor sleep quality, and sleep restlessness had significantly higher PSQI scores in comparison to individuals without such problems.

Introduction

Subjective assessment of sleep quality and disturbance is important to a variety of researchers and clinicians because diminished sleep quality and the presence of sleep disturbance can profoundly impact quality of life and may be associated with physical and/or emotional illness 1, 2. Sleep disturbance and poor sleep quality may result from physical discomfort, side effects of medications, and other aspects of physical illness, and can be related to psychiatric disorders such as depression, anxiety, and schizophrenia.

Although various methods are available to study subjective sleep quality, many do not provide a comprehensive assessment of sleep quality and sleep disturbance. Methods such as single item scales [3], simple visual analog scales 4, 5, temporal sleep logs for recording time of sleep onset and awakening [6], or sleep diaries 7, 8 tend to assess only one or two components of sleep quality. Although standardized questionnaires provide a more comprehensive assessment of sleep quality, relatively few such questionnaires exist. Of three standardized measures of sleep quality found in the literature, the Karolinska Sleep Diary [9], the Verran and Snyder-Halpern Sleep Scale [10], and the Pittsburgh Sleep Quality Index (PSQI) [1], the PSQI is the most widely used 2, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20.

The PSQI was originally designed for use in clinical populations as a simple and valid assessment of both sleep quality and disturbance that might affect sleep quality [1]. According to the scale’s authors, advantages of the PSQI include abilities to: (a) determine patterns of sleep dysfunction over a 1-month period through assessment of both qualitative and quantitative data; and (b) calculate a simple, global score that conveys both the number and severity of sleep problems [1]. The PSQI consists of 19 items that produce a global sleep quality score and the following 7 component scores: sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbance, use of sleeping medications, and daytime dysfunction. Items and component scores were designed to represent standard areas that are assessed by clinicians when individuals report sleep complaints. PSQI items use varying response categories that include recording usual bed time, usual wake time, number of actual hours slept, and number of minutes to fall asleep, as well as forced-choice Likert-type responses. Psychometric properties of the PSQI were supported using data collected from 52 healthy subjects and 96 individuals with sleep problems [1].

Published psychometric information for the PSQI has been limited despite widespread use of the scale in a variety of populations. The PSQI has been used among healthy individuals of various ages [2], persons with Parkinson’s disease [14], trauma survivors [13], bereaved spouses [17], and patients with affective or panic disorder, depression, or social phobias 15, 16, 17, 18, 19. Apart from the original article cited earlier [1], only three additional studies have included information on reliability and validity of the PSQI and these studies are limited in three ways: (a) by the amount of psychometric information reported; (b) by small sample sizes (n’s<50); and (c) by a focus on healthy or elderly individuals, rather than more diverse clinical populations 11, 12, 20. In one study of 18 cognitively intact, elderly, nursing home residents, 19-day test–retest reliability (range 3–44 days) was high for global scores (r=0.82) and moderate to high for component scores (r’s=0.45–0.84) [12]. In another study, correlations between the PSQI and a measure of vigor from the Circadian Type Questionnaire provided estimates of validity [11]. Vigor was moderately correlated (r=−0.43) with PSQI global scores (worse sleep quality associated with less vigor) among 35 healthy individuals aged 20–30 years. In a third study, PSQI scores were used to validate 20 individuals’ self-ratings as either poor or normal sleepers [20].

Given the limitations of the available psychometric data and the fact that the PSQI is appropriate for use in a variety of clinical populations, additional psychometric data from larger and more diverse populations would be useful. The goal of this study was to examine psychometric properties of the PSQI within four clinical populations: bone marrow transplant (BMT) patients; renal transplant patients; women with breast cancer (BC); and women with benign breast problems (BBP). Data from these populations were available to us because we had used the PSQI in previous quality-of-life research. The purpose of this study was to examine internal consistency reliability and construct validity (convergent, discriminant, and known groups validity) of the PSQI within each of the aforementioned populations. Comparisons between groups were not made due to wide variability in demographic, disease, and treatment variables across groups.

Section snippets

Method

The first of four samples included in the psychometric analyses consisted of 158 adult BMT survivors (59 women and 99 men). Survivors had received BMT as treatment for cancer at one of five BMT centers and were in disease remission at time of interview. The second sample included 57 adult renal transplant survivors (31 women and 26 men). The third sample included 107 women diagnosed with nonmetastatic breast cancer recruited from a comprehensive breast care clinic during routine follow-up

Completeness of data

Because participants represented populations in which fatigue and other physical problems could potentially interfere with completion of questionnaires, frequency distributions for individual items were examined to assess percentages of missing data in each group. In all groups, less than 4% of data on individual items were missing. With the exception of the renal transplant group, the most commonly missed items pertained to reasons for having trouble sleeping on the sleep disturbance component.

Discussion

Results suggest the PSQI can be successfully self-administered to clinical populations. With the exception of the BC group, less than 4% of data on individual items were missing. In all groups except the renal transplant group, missing data were most prevalent among the eight-item sleep disturbance component. This component consists of reasons for having trouble sleeping, such as cannot get to sleep within 30 minutes, feel too hot, and have pain. Response categories for these items included:

References (29)

  • C. Trenkwalder et al.

    L-dopa therapy of uremic and idiopathic restless legs syndromea double-blind, crossover trial

    Sleep

    (1995)
  • T. Akerstedt et al.

    The subjective meaning of a good sleepan intra-individual approach using the Karolinska Sleep Diary

    Percept Motor Skills

    (1994)
  • R. Snyder-Halpern et al.

    Instrumentation to describe subjective sleep characteristics in healthy subjects

    Res Nurs Health

    (1987)
  • D.J. Buysse et al.

    Quantification of subjective sleep quality in healthy elderly men and women using the Pittsburgh Sleep Quality Index (PSQI)

    Sleep

    (1991)
  • Cited by (1189)

    View all citing articles on Scopus
    View full text