Death Scene
© Lbusca / Getty Images
Estimating someone's time of death can be a complicated and vague affair, with important ramifications - especially if it's suspicious.

And let's face it, as wryly noted on Coroner Talk, perpetrators are highly unlikely to check the time they committed a homicide and wouldn't let on even if they did.

Scientists say they have now devised a simplified, versatile numerical model to predict body cooling and thus the time elapsed since death (the post-mortem interval), reporting their findings in the journal Science Advances.

Putting their model to the test in deceased people, they projected post-mortem intervals with an average 38-minute error, a considerable improvement on the three to seven-hour discrepancy produced by the current gold standard method.

This is a big deal, as determining someone's time of death "remains one of the most important challenges in forensic medicine to date," write Leah Wilk, from the University of Amsterdam, Netherlands, and co-authors.

Current methods involve invasive sampling of body tissue or fluids followed by extensive lab tests or examining pathophysiological indicators at the crime scene - of which changes in body temperature are most used.

Time of Death
© MC Aalders, G Edelman, L Wilk, Amsterdam UMC, location AMC, Amsterdam, The Netherlands.
The gold standard model relies on rectal temperature, which assumes different human bodies follow the same cooling curve and doesn't allow for varying real-life conditions.

To overcome these limitations, the new model integrates a vast range of variables derived from the body, environment, clothes coverage, contact surface, submersion in water and environmental changes before and after the body's discovery.

It also enables body temperature to be computed externally, limiting risk of contaminating bodily tissues.

To validate the model, they recorded changes in skin temperature of the stomach, chest, forehead and thighs of four human bodies over post-mortem intervals ranging from five to 50 hours and compared these with their predicted body cooling curve.

More than 80% of predictions deviated from the actual time of death by no more than an hour.

The findings have broad applicability, as the deceased people ranged from 60 to 94 years of age and had vastly different sizes and body composition.

The authors say this makes their model a vastly improved, forensically relevant method for estimating when someone took their last breath.

This could make all the difference in criminal investigations.

Consider, for instance, that a suspect claims to have been home with their partner until 2pm before heading to work then returning at 8pm to find their partner dead on the floor.

If the medical examiner can only estimate the time of death as between 11am and 4pm, there is some room for doubt. If, however, it is determined that death occurred between 1pm and 2pm, then there really is a suspect.