Abstracts

Confirmatory Factor Analysis of the Adult Epilepsy Self-Management Measurement Instrument: Scale Reduction

Abstract number : 2.372
Submission category : 17. Public Health
Year : 2021
Submission ID : 1826650
Source : www.aesnet.org
Presentation date : 12/5/2021 12:00:00 PM
Published date : Nov 22, 2021, 06:55 AM

Authors :
Cam Escoffery, PhD - Rollins School of Public Health; Yvan Bamps, PhD - Emory University; Robin McGee, PhD - Rollins School of Public Health; Archna Patel, MPH - Rollins School of Public Health; Nancy Thompson, PhD - Rollins School of Public Health; W Curt LaFrance, MD, MPH - Professor of Psychiatry and Neurology, Brown University; Regine Haardorfer, PhD - Rollins School of Public Health

Rationale: Epilepsy self-management activities have been reported to improve overall quality of life by decreasing seizure frequency, reducing health care costs, promoting seizure control, and increasing health and social outcomes. Few scales have been developed and tested to measure self-management behaviors for people with epilepsy (PWE). The Adult Epilepsy Self-Management Measurement Instrument (AESMMI) was created as a comprehensive scale covering relevant domains of self-management. The original scale had 113 items. This study aimed to develop an abbreviated version of the AESMMI using confirmatory factor analysis.

Methods: The data were from scale testing of a sample of adults with epilepsy who reported that a health provider diagnosed them with epilepsy or a seizure disorder from across the U.S. These adults completed a paper or online survey related to epilepsy history, self-management behaviors, and demographics. The SM items were grouped in 11 domains: 1) Healthcare communication; 2) Treatment management; 3) Coping; 4) Social support; 5) Seizure tracking; 6) Wellness; 7) Seizure response; 8) Safety; 9) Medication adherence; 10) Stress management; and 11) Proactivity. We calculated internal consistency (Cronbach's alpha) for the overall scale and the subscales. To reduce the scale, we removed items based low factor loadings and expert input. Confirmatory factor analysis was conducted hypothesizing the same factor structure as for the full scale in Mplus 8.4. All scale items were modeled as ordinal and thus the WLSMV estimator was used; all data were used in the analysis.

Results: The participants (n=422) were mostly female (73%), White (83%), and some had a college degree (42%). They had been diagnosed with epilepsy for 20.6 years (SD=15.0). The AESMMI was reduced to 37 items with 11 domains/factors. The overall Cronbach’s alpha for the reduced scale was very good at 0.915 and subscale alphas ranged from 0.536 to 0.915. Many subscales had adequate ( >0.7) to good ( >0.8) Cronbach’s alpha. However, four subscales with each three items had only (Seizure Response, Safety, Stress Management, & Proactivity) lower reliability. The model fit for the proposed reduced scale was good for all fit indices but Χ2 (p < .05): RMSEA=.041 (95%CI: [.037;.045]), CFI=.959, TLI=.953.
Public Health