Developing Measures of Pain Appraisal and Pain-Related Self-Efficacy for People Living with Chronic Pain [Internet]

Excerpt

Background: The Patient-Reported Outcomes Measurement Information System (PROMIS®) is a system of reliable and flexible measures of patient-reported health status that include several instruments for measuring different aspects of pain. However, 2 pain-related constructs are missing: pain catastrophizing (PC) and pain-related self-efficacy (PRSE). Pain catastrophizing is a form of “cognitive distortion” that can lead to negative affect and can amplify symptoms and distress. PRSE refers to a person's confidence in living well with chronic pain and successfully managing the impact of pain on their lives. Understanding and measuring these aspects of chronic pain is important for designing treatments that can improve quality of life.

Objectives: The purpose of this study was to develop brief, flexible, psychometrically sound, patient-centered, and clinically meaningful measures of PC and PRSE.

Methods: Development of both item banks followed the PROMIS methodology. Patient advisors worked with investigators on every aspect of the study. A panel of pain researchers defined the PC construct. Focus groups and cognitive interviews with people living with chronic pain reviewed the definitions and provided feedback on meaningfulness and clarity of all items. Individuals with chronic pain recruited from various sources (online recruitment companies, previous studies, research registry, pain clinic) completed the candidate items via online or paper surveys. Pain Catastrophizing Scale and Pain Self-Efficacy Questionnaire were used as legacy measures (ie, existing gold standard self-report measures of PC and PRSE). Test–retest stability data were collected from a subset of respondents 40 to 80 hours after initial administration of the candidate items. Reliability was evaluated using intraclass correlation (ICC), and items were examined for unidimensionality, local dependence (redundancy), and differential item functioning (DIF). Item response theory (IRT) was used to calibrate the items.

Results: Final banks included 24 PC and 29 PRSE items calibrated on a demographically (eg, gender, age, race) and clinically (ie, various chronic pain conditions) heterogeneous sample of people with chronic pain (n = 795). The PC instrument was renamed “Pain Appraisal Scale” (PAS) after feedback from patient groups indicated “catastrophizing” was perceived as stigmatizing. Items for short forms were selected by considering item parameters and content coverage. Six items that captured the trait continuum were selected for each short form. Correlations between the 6-item short forms and the full item bank scores (PAS r = 0.99; PRSE r = 0.85) and test–retest reliability (PAS ICC = 0.93; PRSE ICC = 0.90) were excellent. Correlations of short form scores with legacy measures of related constructs were high (>0.8).

Conclusions: Six-item short forms measure well across the PC and PRSE continuums and the scores are highly correlated with the full item bank scores. The results support the validity and reliability of the PAS and PRSE short forms. Short form scores are on the same metric and directly comparable to full item bank scores, are brief, and are well suited for research and clinical practice. The short form and the full item bank will be available publicly and free of charge at uwcorr.washington.edu.

Limitations and Subpopulation Considerations: If people with additional chronic pain conditions were included, they may have provided different perspectives on PRSE and PC. Analyses suggest scores are primarily driven by the level of PRSE and PC, rather than demographic characteristics (eg, age, gender).