Psychometric testing and validation of the Chronic Pain Sleep Inventory

Patients with chronic pain often experience significant disruptions in sleep. A potential benefit of treatment aimed at pain relief is improved quality of sleep in patients with chronic pain.

The goal of this study was to evaluate the psychometric properties of the Chronic Pain Sleep Inventory (CPSI) and provide preliminary evidence of its construct validity in assessing sleep problems among patients with chronic pain.

Data came from four 12-week, multi-center, double-blind, randomized, placebo-controlled clinical trials of chronic low back pain and osteo-arthritis of the hip or knee. CPSI data were collected at baseline and 12 weeks. The 5 CPSI items measured trouble falling asleep (CPSI1), needing sleep medication (CPSI2), awakened by pain during the night (CPSI3) and in the morning (CPSI4), and overall quality of sleep (CPSI5) on a 100-mm visual analog scale. Exploratory and confirmatory factor analyses were conducted to evaluate the underlying dimensionality and structure of the CPSI scales. The known-groups method was used to assess validity by comparing CPSI item scores across groups of patients known to differ in the presence and severity of sleep problems as measured by the Physical Dependence Questionnaire.

A total of 2674 patients were included in the study (mean age, 57.6 years; 61.0% female; 80.2% white). The majority of the patients were treated for chronic pain related to osteoarthritis of the hip or knee (n=2294). Findings revealed a single sleep problems index can be scored from 3 of the 5 CPSI items (CPSI1, CPSI3, and CPSI4). All 3 items attributed sleep problems to pain, with high factor loadings (>0.80) and a high internal consistency (>0.90). Moderate to high correlations (r >or= 0.50) between CPSI items demonstrated convergent validity, and weak correlations (r<0.50) with other health-related scales (the Western Ontario and McMaster Universities Osteo-arthritis Index and the 36-item Short-Form Health Survey) demonstrated discriminant validity. All CPSI items showed greater ability to discriminate and respond to changes in the presence and severity of sleep problems than the other health-related scales.

CPSI items showed good construct validity, and the results support the scoring of a reliable single index from 3 of the 5 CPSI items that all attributed sleep problems to pain.