The most frequent reason cited for withdrawal of an approved drug is toxicity, yet no simple solution exists to adequately predict such adverse effects. Compound prioritization and optimization during in vitro screening cascades need to be based on confidence, not only in efficacy and bioavailability, but also in safety. A wider number and diversity of potential molecular and cellular effects of compound interactions might affect safety than might affect efficacy or bioavailability. Accordingly, cytotoxicity assessment is less specific, more multiparametric, and extrapolatable with less certainty, unless there are specific safety signals indicated by the chemical structure or by precedents. Cytotoxicity assessments have been limited by their inability to measure multiple, mechanistic parameters that capture a wide spectrum of potential cytopathological changes. Assays with multiple parameters for key, multiple, and different features, such as in high content screening (HCS), are more predictive because they cover a wider spectrum of effects. Assays need to be applied to a large set of marketed drugs that produce toxicity by numerous and different mechanisms for assessment of correlation with human toxicity. This will enable determination of the concordance between in vitro and in vivo results. Multiparametric, live cell, prelethal cytotoxic HCS assays for assessing the potential of compounds for causing human toxicity address some of the limitations of traditional in vitro methods. Assays of this class were used to screen a library of drugs with varying degrees of toxicity and it was found that the sensitivity of the assays was 87%, whereas assay specificity was more than 90%, thereby minimizing false positives.