5613177 AN INTERNATIONAL, LONGITUDINAL, OBSERVATIONAL STUDY TO EVALUATE THE PSYCHOMETRIC PROPERTIES OF A DAILY DIARY FOR ADOLESCENTS AND ADULTS WITH SICKLE CELL DISEASE

Abstract

Background: Vaso-occlusive crises (VOCs) are a serious complication of sickle cell disease (SCD). The Sickle Cell Pain Diary – Self Report (SCPD-S) was developed to capture the impact of VOCs, which occur sporadically and unpredictably. The daily diary evaluates the severity of SCD-related pain during and outside of a VOC and the impact of pain on patient-relevant concepts. Aims: A longitudinal observational study was conducted to evaluate the psychometric properties of the SCPD-S items that assess VOC impact. Methods: Participants were eligible for the study if they were ≥12 years old, had a verified diagnosis of SCD, experienced ≥1 VOC in the past 12 months, and lived in the United States, United Kingdom, or Kingdom of Saudi Arabia. Participants completed the SCPD-S daily for at least 3 months through a mobile application and at pre-specified intervals, completed additional surveys which served as criterion measures. Factor analyses were conducted to examine the conceptual framework of the SCPD-S VOC impact items. Psychometric properties of 3 different VOC impact scores were evaluated. Analyses included tests of internal consistency reliability, test-retest reliability, convergent/divergent validity, known-groups validity, and responsiveness. Results: The analytic sample included 299 participants (87% of enrolled participants) who had ≥1 VOC during the study period and missed ≤5 daily assessments in the month in which the VOC occurred. The median age was 30 (range: 12-67); participants were predominantly female (72%) and Black (83%). A unidimensional multi-level confirmatory factor analysis (CFA) yielded poor model fit, failing to support the hypothesized unidimensional construct of VOC impact (Table 1, Model 1). Multi-level exploratory factor analysis revealed a 2-factor model with sufficient model fit (Model 2), which was also supported by a CFA (Model 3). The 2 factors represent interference with general activities (Scale 1) and difficulty completing specific activities of daily living (Scale 2). Cronbach’s α was 0.87 for Scale 1 and 0.86 for Scale 2, indicating adequate internal consistency reliability. Tests of differential item function show that after controlling for VOC impact, participants’ responses to SCPD-S items do not depend on age or country of residence. Psychometric analyses were conducted on the Daily VOC Impact scores, which measure VOC impact over the past 24 hours. Test-retest reliability of Scale 1 Daily VOC Impact scores was satisfactory; results for Scale 2 were weaker (intraclass correlation coefficient=0.78 for Scale 1 and 0.66 for Scale 2). Scale 1 Daily VOC Impact scores were responsive to change; change scores were correlated with changes in symptom severity (ρ=0.41, P<0.001). Scale 2 was not responsive to change (ρ=0.27, P<0.001). Additional analyses provided support for scores that aggregate across days within a VOC, not for scores that aggregate across a fixed period.Conclusion: Findings support the use of the SCPD-S to characterize the daily impact of VOCs, confirming the need for a daily diary to capture variability in VOC impact across days. Compliance with the diary was high (87% completed ≥25 days in a month). While factor analyses supported a 2-factor model of VOC impact, Scale 1 performed consistently better than Scale 2 across analyses. As such, it is recommended that items comprising Scale 2 be dropped from the diary, leaving a single scale that is a reliable and valid measure of VOC impact. Reference 1. White et al, J Patient Rep Outcomes 2021; 5(1):1-12