Limited expected value - an effective tool for analysing highly censored data?