Određivanje doze leka koja izaziva odgovor po principu 'sve ili ništa' kod 50% ispitanika (ED50) uz pomoć probita

Abstract

There is a number of drugs which produce responses that are clinically visible according to the principle 'all-or-nothing', i.e. the effect on certain dose either happens, or not. In order to calculate the dose of such drug that would cause response in 50% of exposed patients (effective dose 50 or abbreviated ED50), probit analysis is necessary. The essence of this application of probit analysis is formation of linear regression between logarithm of the drug doses and probits, i.e. numbers of standard deviations that responses to corresponding doses deviate from the most frequent response (assuming normal distribution of responses to certain doses of the drug). ED50 is further calculated from equation of this linear regression as the dose corresponding to probit value five.