Avoiding the False Negative Results in LC Method Robustness Testing by Modifications of the Algorithm of Dong and Dummy Factor Effects Approach

Abstract

The algorithm of Dong and error estimation based on a priori declared negligible effects (dummy factor effects approach) were evaluated for testing robustness, in case they provide contradictory results. Robustness of LC method for separation of ropinirole and its impurity was determined. Plackett-Burman design was applied to unravel the significant/influential factors. Four selected responses (retention factor of ropinirole and impurity, selectivity and resolution) were analyzed by the standard algorithm of Dong and dummy factor effects approach. The interpretation of the results for retention factor of ropinirole and resolution was inconclusive, since the two applied approaches identified different significant factors. The potential false negative or false positive results were avoided reanalyzing the results by the modified approaches. Modification of the dummy factor effects approach included omitting the large dummy effects values, and modification of the algorithm of Dong included calculating the margin of error values for all different number of factors included, and then selecting the lowest one to be the critical effect. The applied reanalysis allowed final important factor identification and non-significant intervals for significant effects calculation.