Psychometric Classical Theory Test and Item Response Theory Validation of Patients’ Knowledge, Attitudes and Practices of Uncomplicated Malaria Instrument

Abstract

Objective: To develop and validate Patients’ Knowledge, Attitudes and Practices Instrument for Uncomplicated Malaria (PKAPIUM) through Classical Test Theory (CTT) complemented by Items Response Theory (IRT). Methods: A draft 31-items’ scale was developed using relevant variables from literature and initially screened by six experts before it was used to collect data from 300 patients receiving treatment for uncomplicated malaria in Primary Health Care (PHC) facilities in Plateau state, Nigeria. An orchestrated classical and modern psychometric approach including CTT and IRT was then used to validate the draft instrument through IBM® Statistical Package for Social Sciences (SPSS®) version 23 and Analysis of Moment Structures (AMOS™) software version 22 and Bond and Fox software®, respectively. Results: The 31-items’ draft scale showed good Item’s Content Validity Index (I-CVI) (> 0.8) with good Universal Agreement (UA) level of Scale Content Validity Index (S-CVI/UA) (0.9 – 1) and average CVI (S-CVI/Ave) (0.98 – 1). The CTT and Rasch analyses resulted in retention of twenty one items distributed under Knowledge, Attitude and Practice (KAP) constructs, with Average Variance Extracted (AVE), square root AVE, chi-square, Standardized Root Mean square Residual (SRMR), Root Mean Square Error Approximation (RMSEA), items’ infit Mean Square (MNSQ), Infit Standardized Z-scores (infit Zstds), Point-Measure Correlation Coefficients (PTMEA Corr), Cronbach’s alpha, items’ and person’s reliability indices within accepted limits. Conclusion: The new scale was considered valid and reliable for assessing patients’ knowledge, attitudes and practices on uncomplicated malaria. Key words: Cronbach’s alpha, Factor analysis, Infit and outfit indices, Persons and items reliability, Point-measure correlation coefficients, Standardized Zscore