Steve van de Weijer, Catrien Bijleveld, Doreen Huschek

• Background: Previous research has shown that offenders are at increased risk to die prematurely, but the etiology of this association is still unknown. Moreover, most previous studies use relatively short follow-up periods and do not take into account variation within the offender population with respect to frequency, timing and types of offenses. Method: Using conviction data for a number of families at high-risk of offending born on average in 1932, we study mortality in both offenders and non-offenders, from a similar socio-economic background, until 2007. We condition on life expectancy of the parents, age, gender, year of birth and marital status. We investigate associations between mortality and offending for different types of offenses: violent offenses, property offenses, weapons offenses, drugs offenses and driving under influence. Results: In general, offending sample members were not significantly more likely to have died than non offending sample members. Compared to the general population, however, both the offending andBackground: Previous research has shown that offenders are at increased risk to die prematurely, but the etiology of this association is still unknown. Moreover, most previous studies use relatively short follow-up periods and do not take into account variation within the offender population with respect to frequency, timing and types of offenses. Method: Using conviction data for a number of families at high-risk of offending born on average in 1932, we study mortality in both offenders and non-offenders, from a similar socio-economic background, until 2007. We condition on life expectancy of the parents, age, gender, year of birth and marital status. We investigate associations between mortality and offending for different types of offenses: violent offenses, property offenses, weapons offenses, drugs offenses and driving under influence. Results: In general, offending sample members were not significantly more likely to have died than non offending sample members. Compared to the general population, however, both the offending and non offending sample members were at increased risk to die. Sample members who were convicted for driving under the influence of alcohol or weapons offenses were at increased risk to die prematurely compared to non-offending sample members. Conclusions: The relationship between offending in general and mortality is largely spurious. Limitations: The use of official conviction data might have influenced the results. (C) 2015 Elsevier Ltd. All rights reserved.…