Reply to Comments on Child Abuse Misdiagnosis
Author Information
Author(s): van Gemert Martin J. C., Vlaming Marianne, Nikkels Peter G. J., van Koppen Peter J.
Primary Institution: Amsterdam University Medical Centers
Conclusion
The use of Bayes' Theorem in child abuse assessments may lead to incorrect conclusions about physical abuse probabilities.
Supporting Evidence
- The likelihood ratio for child abuse was estimated to be 10-100, but the actual probability of abuse was very low.
- The authors argue that Bayesian statistics are misused in child abuse assessments.
- The case involved a girl with bruises and rib fractures, which were later attributed to her mother's medical condition.
Takeaway
This study discusses how a girl was wrongly suspected of child abuse due to misinterpretation of medical data, highlighting the need for better methods in such assessments.
Potential Biases
The authors suggest that the LECK's anonymous methodology may lead to uncritical acceptance of their conclusions.
Limitations
The study critiques the methodology of the LECK child abuse program without providing new empirical data.
Participant Demographics
A 2-month-old girl of Dutch-Armenian descent.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website