Original Article Criterion validity of the Pittsburgh Sleep Quality Index and Epworth Sleepiness Scale for the diagnosis of sleep disorders Takeshi Nishiyama a,⇑ , Tomoki Mizuno b , Masayo Kojima c , Sadao Suzuki c , Tsuyoshi Kitajima d , Kayoko Bhardwaj Ando e,f , Shinichi Kuriyama e,f , Meiho Nakayama e,f a Department of Public Health, Aichi Medical University, Nagakute, Japan b Nagoya City University Medical School, Nagoya, Japan c Department of Public Health, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan d Department of Psychiatry, Fujita Health University, Toyoake, Japan e Department of Otolaryngology, Nagoya City University Graduate School of Medical Sciences, Nagoya, Japan f Good Sleep Center, Nagoya City University Hospital, Nagoya, Japan article info Article history: Received 29 September 2013 Received in revised form 4 December 2013 Accepted 9 December 2013 Available online xxxx Keywords: Pittsburgh Sleep Quality Index (PSQI) Epworth Sleepiness Scale (ESS) Obstructive sleep apnea (OSA) Periodic limb movement disorder (PLMD) Rapid eye movement sleep behavior disorder (RBD) Narcolepsy abstract Objectives: (1) To examine criterion validity of the Pittsburgh Sleep Quality Index (PSQI) and Epworth Sleepiness Scale (ESS) using obstructive sleep apnea (OSA), periodic limb movement disorder (PLMD), rapid eye movement sleep behavior disorder (RBD), and narcolepsy as criterion standard. (2) To summa- rize the evidence for criterion validity of the ESS for the diagnosis of OSA by a meta-analysis that com- bines the current and previous studies. (3) To investigate the determinants of the PSQI and ESS scores. Methods: The PSQI and ESS as well as the Hospital Anxiety and Depression Scale (HADS), which measures anxiety and depression levels, were administered to 367 patients consecutively referred to a sleep clinic. They underwent overnight polysomnography (PSG) and the multiple sleep latency test if narcolepsy was suspected. Results: The area under the receiver operating characteristic curves for the ESS and PSQI (and its sub- scale) were <0.9, meaning that these questionnaires were not highly accurate for predicting the four sleep disorders. The meta-analysis found that the ESS had no value in identifying OSA. The variable that most strongly influenced PSQI or ESS scores was the HADS score. Conclusion: The PSQI and ESS should no longer be used as a screening or diagnostic instrument for the four PSG-defined sleep disorders, especially in a low-risk population. Ó 2014 Elsevier B.V. All rights reserved. 1. Introduction Many self-report questionnaires have been developed for mea- suring various aspects of sleep disturbance [1]. Two of the most widely used are the Pittsburgh Sleep Quality Index (PSQI) [2] and the Epworth Sleepiness Scale (ESS) [3]. The PSQI and the ESS were originally designed to measure sleep quality and subjective day- time sleepiness, respectively, but they were not designed to screen for a specific sleep disorder [2,3]. Nevertheless, they have been widely used in clinical settings, with the expectation that the PSQI can identify persons at high risk for insomnia or that the ESS can identify persons at high risk for OSA or narcolepsy by identifying persons with excessive daytime sleepiness [4,5]. Although a previous study examined the criterion validity of the PSQI using actigraphy, sleep diary, and a questionnaire of depres- sion as criterion standard [6], there have been no studies using polysomnographic sleep abnormalities as criterion standard. On the other hand, several studies have been conducted to examine the criterion validity of the ESS for the diagnoses of OSA and nar- colepsy [7–12]. For example, the original developer of the ESS used the case-control method in which a group of typically diagnosable patients with narcolepsy is compared with a group of unquestion- ably healthy subjects [5]. In this situation, the ability of the ESS to discriminate between the two groups may be overestimated. The failure to include an appropriately broad spectrum of diseased and non-diseased subjects in the study population may give falsely high sensitivity and specificity and falsely elevate the area under the receiver operating curve (AUC). This effect is known as spec- trum bias [13]. http://dx.doi.org/10.1016/j.sleep.2013.12.015 1389-9457/Ó 2014 Elsevier B.V. All rights reserved. ⇑ Corresponding author. Address: Department of Public Health, Aichi Medical University, 1-1 Yazako, Nagakute, Aichi 480-1195, Japan. Tel.: +81 561 62 3311; fax: +81 561 62 5270. E-mail address: nishiyama@minos.ocn.ne.jp (T. Nishiyama). Sleep Medicine xxx (2014) xxx–xxx Contents lists available at ScienceDirect Sleep Medicine journal homepage: www.elsevier.com/locate/sleep Please cite this article in press as: Nishiyama T et al. Criterion validity of the Pittsburgh Sleep Quality Index and Epworth Sleepiness Scale for the diagnosis of sleep disorders. Sleep Med (2014), http://dx.doi.org/10.1016/j.sleep.2013.12.015