PSYCHOMETRIC EVALUATION OF FLOOD DISASTER MANAGEMENT QUESTIONNAIRE: CONFIRMATORY FACTOR ANALYSIS AND ITEM RESPONSE THEORY ANALYSIS

Abstract

Flood disaster is the commonest natural disaster with huge impact on healthcare services in Malaysia. The FloodDMQ-BM© questionnaire was developed as a tool to assess the knowledge, attitude, and practice of healthcare providers regarding patient management during a flood disaster. We aim to further validate the FloodDMQ-BM© questionnaire by using Confirmatory Factor Analysis (CFA) and Item Response Theory (IRT).This cross-sectional study involved doctors, nurses and paramedics working in the Emergency Department of Hospital Universiti Sains Malaysia, Hospital Raja Perempuan Zainab II and Hospital Kuala Krai. Respondents were required to complete the FloodDMQ-BM© questionnaire. The responses were analysed by using CFA and IRT to establish its validity and reliability. A total of 215 respondents participated in this study. CFA analysis with Maximum Likelihood Robust as the estimation method, on the attitude and practice components resulted in good factor loadings (>0.5) in nearly all items and excellent model fit indices values (CFI = 0.96-0.98, TLI = 0.95-0.96, SRMR = 0.04-0.05, RMSEA = 0.07). Meanwhile, IRT analysis on the knowledge section showed a good two-way marginal fit based on S-X2, and a good model fit with RMSEA of 0.08. Based on the 2PL model by using the IRT assessment of the knowledge section, one item in the knowledge section (K3) was removed (chi-squared residual >4) resulting in improved model fit. The included items had well-standardized loadings (>0.3) and marginal reliability of 0. 651.Our results confirmed that the FloodDMQ-BM© questionnaire displayed valid and reliable psychometric properties.