Teams from Amsterdam UMC, the University of Amsterdam and the Netherlands Forensic Institute (NFI) have developed a technique that improves the accuracy of time-of-death estimates at crime scenes to within less than an hour of the subject’s death.

The new method measures potassium levels in the subject’s eye, analyzes protein in the subject’s tissue and measures the body temperature of the subject in up to four different locations on the body via thermal camera or using sensors adhered to the body. That data is then fed into a model along with other information about the crime scene and the victim (the type of clothing or environment, for example).

Researchers tested this new model on bodies where the actual time of death was known to the researchers. The model managed to perform better than current time-of-death estimations, placing the time of death on average within 45 minutes of the actual time of death.

Current methods for estimating time of death are less reliable according to researchers, relying on cooling of the body following death as determined via rectal temperature. This, combined with the subject’s body weight and ambient conditions point to a location on a table used by forensic investigators that suggests how many hours — typically a window of several hours — have passed since the time of death. However, such tables tend to be inaccurate, failing to take into consideration factors such as body structure and shape despite similar weights, for instance, which might impact outcomes. Additionally, obtaining a rectal temperature is an invasive process that could potentially destroy evidence.

Accurate time of death estimates are critical, particularly in homicide cases, revealing details to help investigators solve a crime. As such, the research teams intend to improve upon the accurate time of death estimates with the addition of motion photogrammetry where the body is photographed from all different angles to create a 3D model.

The research appears in Science Advances.