This article has a significant backstory. A highly influential study by Chadwick et al. in 1991 [1] asserted that reports of fatal short falls in children are

This article has a significant backstory. A highly influential study by Chadwick et al. in 1991 [1] asserted that reports of fatal short falls in children are invariably falsified and actually stem from violence. They reached this conclusion by noting a significantly higher (7 times more in their data) hospital patient fatality rate after short falls (< 1.20 m) compared to long falls (> 3 m). They interpreted this apparent mathematical paradox ("short falls are not 7x more severe that long falls!") as evidence that all reports of short falls were falsified, and that these deaths were violent murders. However, they overlooked many factors, including the fact that most nonfatal short falls do not result in hospital consultations; only the most severe ones do. Conversely, nearly all long falls (such as from several stories) are likely to prompt a hospital visit, even if the child appears uninjured. The hospital fatality rate is heavily influenced by this denominator, which does not accurately reflect the total number of short falls in a population—only the small subset that leads to hospital consultation. This article, along with others, has led many physicians worldwide to doubt reports of short falls whenever a child sustains severe or fatal intracranial injuries, often resulting in criminal prosecution and conviction. There have been numerous cases of miscarriages of justice [2], and the issue persists. In collaboration with mathematician Leila Schneps [3], our paper employs a particularly simple statistical model to quantify the impact of the hospital consultation rate confounding variable on the apparent hospital patient fatality rate. This is a small step towards helping physicians understand that reports of fatal short falls are not always falsified (though they may be, of course, but not always). There are many other scientific arguments, of course (case reports, including videotaped ones, biomechanical data, animal models, etc. [4]) This issue is similar to the famous case of Sally Clark, wrongly convicted because of statistical misunderstandings [5]. [1] https://journals.lww.com/jtrauma/abstract/1991/10000/Deaths_from_Falls_in_Children__How_Far_is_Fatal_.6.aspx https://journals.lww.com/jtrauma/abstract/1991/10000/Deaths_... [2] https://ncip.org/zavion-johnson/ https://ncip.org/zavion-johnson/ [3] https://www.sciencedirect.com/science/article/pii/S0096300324001504 https://www.sciencedirect.com/science/article/pii/S009630032... [4] https://cyrille.rossant.net/introduction-shaken-baby-syndrome-controversy/#sec-falls https://cyrille.rossant.net/introduction-shaken-baby-syndrom... [5] https://en.wikipedia.org/wiki/Sally_Clark https://en.wikipedia.org/wiki/Sally_Clark