Routine mortality monitoring was one of the recommendations in the Shipman Inquiry into Hyde, Greater Manchester GP Dr Harold Shipman, convicted in 2000 of killing 15 of his patients but believed to have killed over 200 in his 23-year career.
But applying three different types of statistical analysis to simulated data has shown that the system is unlikely to detect a mass murderer and more likely to cast suspicion on innocent GPs.
Routine data from NHS Scotland was used to gauge that on average, GPs tend to stay at a practice for three years, and therefore this has to be a reasonable time frame in which to detect a mass murderer.
They then stimulated a dataset of 405,000 patients in 75 practices and looked for suspiciously high annual rates of mortality - fixed at five to 10.
But even the most sensitive method - cumulative sum charts - only successfully detected more than half of suspicious cases if looking for 10 killings a year.
Moreover, it raised false alarms in more than 15 per cent of cases.
Comment below and tell us what you think