Perceived Nursing Interruptions During Medication Administration: Development and Validation of a Questionnaire Among Chinese Nurses.

BACKGROUND: Nursing interruptions during medication administration occur frequently, often leading to medication errors and workflow delays that threaten patient safety and reduce nurses' work efficiency. Current research has relied mainly on observation or interviews, which are labor-intensive, time-consuming, and prone to substantial variability. Practical, standardized instruments are needed to assess nursing interruptions during medication administration (NIMA). METHODS: This cross-sectional descriptive study developed and validated the questionnaire for measuring nursing interruptions during medication administration (Q-NIMA) through a multistep, theory-driven process based on the Systems Engineering Initiative for Patient Safety model. Items were generated from literature and expert input, refined by two Delphi rounds with 12 experts, and piloted with 30 nurses. Item analysis was conducted for comprehensive evaluation, and validity evidence was gathered based on test content, internal structure, and relations to other variables. The sample was split for exploratory factor analysis (EFA) using FACTOR software (based on polychoric correlations and ULS extraction) and confirmatory factor analysis (CFA) using AMOS. Reliability was assessed using internal consistency (Cronbach's α and ORION coefficients) and stability (test-retest reliability). RESULTS: We selected 478 nurses from a tertiary hospital in Hunan Province from July to August 2025. The final questionnaire comprised 25 items across five dimensions. Item analysis indicated that the questionnaire had strong discriminant validity. The content validity indexes were 0.944 at scale level and 0.833-1.000 at item level. EFA extracted five factors explaining 74.25% of the total variance. CFA demonstrated good model fit (X2/df = 2.455, CFI = 0.942, TLI = 0.933, RMSEA = 0.078). Criterion validity showed a significant positive correlation (ρ = 0.310, p < 0.001) between the Q-NIMA and the NASA-TLX. The scale demonstrated excellent reliability: the total Cronbach's α was 0.973, split-half reliability was 0.951, and test-retest reliability (ICC) was 0.992. The five dimensions were labeled: nurse professional competence, nurse psychophysical state, patient-related interruptions, workplace colleague factors, and work system factors. CONCLUSION: The Q-NIMA demonstrates robust psychometric properties, including high reliability, validity, and stability. Furthermore, it serves as a valuable tool to evaluate how nurses perceive nursing interruptions during medication administration. By aligning items with medication tasks and work-system elements, the Q-NIMA moves beyond event counting to support risk profiling and targeted improvement of medication administration.