Searching for a serial killer on a hospital ward.

Area of Science:

• Forensic Science
• Medical Statistics
• Nursing

Background:

• Exposing serial killers in healthcare settings presents significant challenges, as highlighted by high-profile cases.
• Previous reliance on statistical evidence in legal proceedings has faced scrutiny due to inadequate data analysis.
• Investigating suspicious deaths requires objective and robust methodologies to ensure accurate identification of perpetrators.

Purpose of the Study:

• To evaluate the applicability of statistical methods, adapted from diagnostic test evaluation, in identifying a nurse potentially involved in suspicious patient deaths.
• To analyze the correlation between a convicted nurse's work shifts and unexplained patient fatalities in a geriatric unit.

Main Methods:

• Applied statistical metrics (sensitivity, specificity, likelihood ratios) to analyze 29 deaths occurring during a 7-month period in a geriatric ward.
• Compared the work shifts of a convicted nurse, particularly night shifts, with the timing of patient deaths.
• Examined death certificates for discrepancies and conducted medico-legal autopsies on exhumed patients where necessary.

Main Results:

• A significant proportion of deaths (55%) occurred during night shifts, with the nurse in question working a disproportionately high number of these shifts.
• The nurse's shifts coincided with 38% of the deaths, placing her within the top 10% of staff shifts associated with fatalities.
• Suspicious, unexplained blood glucose fluctuations or hypoglycemia-mimicking comas were noted in 55% of deaths occurring during or immediately after her shifts, with one case supported by autopsy.

Conclusions:

• Statistical methods, similar to those used for diagnostic accuracy, can effectively target investigations into potential serial killings within hospital environments.
• While these statistical approaches can identify individuals requiring further scrutiny, they cannot definitively determine guilt.
• Such statistical analyses are valuable tools but must be complemented by thorough case-by-case examinations of circumstances and causes of death, especially in high-mortality settings.