Abstract
The experimental design extracts valuable information about the main effects and interactions from the least number of experiments. The current work constructs a solid-state sensor for selective assay of Ondansetron (OND) in pharmaceutical dosage form and plasma samples. During optimization, the Design Expert® statistical package constructed a custom design of 15 sensors with different recipes. We fed the software with the experimentally observed performance parameters for each sensor (slope, LOQ, correlation coefficient, and selectivity coefficient for sodium ions). The computer software analyzed the results to construct a prediction model for each response. The desirability function was adjusted to optimize the Nernstian slope, minimize the LOQ and selectivity coefficients, and maximize the correlation coefficient (r). The practical responses of the optimized sensor were close to those predicted by the model (slope = 60.23 mV/decade slope, LOQ = 9.09 × 10-6 M, r = 0.999, sodium selectivity coefficient = 1.09 × 10-3). The sensor successfully recovered OND spiked to tablets and human plasma samples with mean percentage recoveries of 100.01 ± 1.082 and 98.26 ± 2.227, respectively. Results were statistically comparable to those obtained by the reference chromatographic method. The validated potentiometric method can be used for fast and direct therapeutic drug monitoring of OND co-administered with chemotherapeutic drugs in plasma samples.