Scientific evidence against a nurse accused of deliberately harming patients

These cases are characterized by an absence of witnesses, an absence of material evidence, an absence of plausible motive and no general agreement that a crime has even occurred. The prosecution is based necessarily on scientific evidence. This evidence has two components: the first deals with speculation on how the supposed crimes may have taken place, examples such as insulin poisoning, overfeeding and air embolism being the most common. The second is based on statistics. The statistics, motivated by an unanticipated cluster of adverse events, are believed to show that a crime must have taken place. Furthermore, the statistical correlation between the presence of the accused and the timing of the supposed crimes is such that the most likely culprit is the accused. In this presentation we focus exclusively on the statistical evidence since the other scientific evidence, however interpreted or understood, does not point a finger at anyone in particular. Most of this presentation can be followed by a general audience. Some technical issues are described but without going into deep mathematical detail.