Background There is controversy over estimates of the civilian death toll associated with violent conflict in Iraq.

Objective To describe the epidemiology of media worker homicide in Iraq in order to describe its relationship with the homicide level for all civilians (and to identify risk factors).

Methods Data on the homicides occurring from 2003 to 2011 were systematically collated from five international databases and analysed.

Results In this time period, the annual number of homicides of media workers rose (peaking at n=47 in 2007) and then declined (n=9 in 2011). The peak years (2006–2007) for media worker homicides matched the peak years for estimated civilian fatalities. Out of a total of 194 homicides, 85% were Iraqi nationals and 61% worked for Iraqi media agencies (the latter increased over time relative to the first 5-year time period, p=0.02). Some were killed whilst reporting (36%), but most (64%), were targeted in other settings (eg, when commuting or at home). Common perpetrators were: political groups (51%), and coalition forces (9%), but this was often unknown (27%). None of the homicides during this period have been legally ‘solved’. For each targeted attack causing a media worker homicide, another two people were also killed on average (range 0–85).

Significance The relatively high quality of data on media worker homicides, suggests that routine surveillance of this sentinel population may provide an indicator of trends in societal violence. But in itself media worker homicide in Iraq is a major problem that needs urgent preventive measures.