Several adverse perinatal events were associated with an increased risk of violent and non-violent criminal convictions; low birth weight, smallness relative to gestational age and preterm birth with non-violent convictions were higher for men

Adverse perinatal events and offspring criminal convictions in men and women: A population-based study. Sofi Oskarsson et al. Journal of Criminal Justice, Volume 78, January–February 2022, 101879. https://doi.org/10.1016/j.jcrimjus.2021.101879

Highlights

• Adverse perinatal events were associated with an increased risk of violent and non-violent criminal offending

• Associations between some adverse perinatal events with criminal convictions were significantly higher for men than for women

•There was a dose-dependent association between adverse perinatal events with criminal convictions for both men and women

Abstract

Background: We examined associations of adverse perinatal events with offspring violent and non-violent criminal convictions in men and women.

Methods: All singleton births between 1973 and 1995 (n = 1,146,570 men, n = 1,085,217 women) were identified through Swedish population-based registers. Information about adverse perinatal events was retrieved from the Medical Birth Register. Outcomes were criminal convictions collected from the National Crime Register. We estimated absolute and relative risks of being convicted of criminal convictions using the Kaplan-Meier method and survival analyses for men and women separately. We also tested for differences in magnitudes of associations for men versus women.

Results: Several adverse perinatal events were associated with an increased risk of violent and non-violent criminal convictions in both men and women. Associations between low birth weight, smallness relative to gestational age and preterm birth with non-violent criminal convictions were statistically significantly higher for men than for women. There was a dose-dependent association between adverse perinatal events with violent and non-violent criminal convictions for both men and women, indicated by the strengthened magnitude of HR estimates with exposure to an increasing number of adverse perinatal events.

Conclusions: Adverse perinatal events are associated with violent and non-violent criminal convictions in men and women, with some differences in risk estimates between sexes. Findings are compatible with theoretical accounts implicating disruption of the neurodevelopment during the perinatal period.

4. Discussion

In this large-scale population-based study, we found that several adverse perinatal events were associated with an increased risk of violent and non-violent criminal convictions. These results are not only in line with findings from previous research (Liu et al., 2009), but also extend these. Specifically, we add to the existing literature by showing that the exposure to adverse perinatal events increases the risk of violent and non-violent criminal convictions, in both men and women. Further, we found evidence for a dose-dependent relationship between adverse perinatal events and criminal behavior, whereby the exposure to an increasing number of adverse perinatal events elevated the risk for violent and non-violent convictions in both men and women. Additionally, our findings suggest that there may be different adverse perinatal events that heightened the risk of violent and non-violent criminal convictions for men compared to women, even though they are few. For example, observed associations for low birth weight and preterm birth with non-violent criminal convictions were significantly higher for men than for women. These findings build on the existing knowledge regarding the potential importance of early risk factors for criminal offending and point to a dose-dependent relationship.

We also found that small head circumference at birth was associated not only with violent criminal convictions as reported in earlier studies (Ikäheimo et al., 2007) but also with non-violent criminal convictions. These findings were evident for both men and women. Small head circumference has in the previous literature been referred to as a minor physical anomaly (Denno, 1990). From a biopsychosocial criminological perspective, minor physical anomalies have been viewed as reflecting physical and neural maldevelopment of the fetus during pregnancy, potentially due to both genetic and environmental influences (Raine, 2002b; Raine, 2019). Minor physical anomalies have long been a known correlate of male aggression (Waldrop et al., 1978) and criminality (Raine, 2013), but the current results indicate that this association extends to criminal offending among women.

One previous study using a sample that overlapped partly with the present study's, reported a reduced risk of any criminal conviction among offspring who were born preterm (D'Onofrio et al., 2013). This association remained even among discordant siblings, pointing to an independent association between being born preterm and later criminal offending. Being born preterm would thus serve as a protective factor for later criminality. We found that in models adjusted for birth year, preterm birth increased the risk for violent and non-violent criminal convictions among men, but not women. However, in mutually adjusted models, preterm birth was associated with violent criminal convictions but not non-violent convictions, among men, but not women. The discrepancy in findings may be due to the fact that the present study examined men and women separately, whereas the previous study (D'Onofrio et al., 2013) used a total population sample with an adjustment for offspring sex. Another explanation may relate to the different outcomes studied: whereas the present study utilized violent and non-violent criminal convictions as separate outcomes and D'Onofrio et al. (2013) employed any criminal conviction as the outcome.

Taken together, these findings highlight the importance of examining men and women separately when studying associations between adverse perinatal events and criminal convictions, as well as studying violent and non-violent criminal convictions as separate outcomes. The male brain has been suggested to be more susceptible to influences early in life that can disrupt normative neurological development (Golding & Fitzgerald, 2019), and it is well-known that preterm birth is associated with behavioral and psychological problems in later life (Bhutta et al., 2002). Our findings suggest that the association between preterm birth and criminal offending should be further investigated to better understand whether this is a factor contributing to the higher overall rate of criminal convictions for men as compared to women.

One previous study has demonstrated associations between adverse perinatal events and later self-reported violent behavior in a dose-dependent manner, among both men and women (Murray et al., 2015). Our results are in line with these findings and suggest that when adverse perinatal events accumulate in the same individual (up to four for men and up to three for women in the present study), the risk for criminal convictions in the offspring increases for both men and women. The dose-dependent association between adverse perinatal events and offspring criminality is in line with biopsychosocial criminological theory, which suggest that adverse perinatal events contribute to a disruption of neuropsychological development, which in turn can heighten the risk for criminal behavior (Raine, 2002c). It is possible that the accumulation of adverse perinatal events in the same individual index disruption of fetal development during pregnancy more effectively than exposure to a single event. This may further be why we see relatively weak associations between individual adverse perinatal events with offspring criminal convictions, as well as low cumulative incidences, even though there are some indications that certain adverse perinatal events are more important than others (e.g., smallness for gestational age). Further research is needed to clarify the nature and etiological basis of the dose-dependent association between perinatal events and later criminal behavior in offspring, which in turn can inform clinical practice and possibly prevention efforts.

Interestingly, the adverse perinatal events that evidenced an association with criminal offending in the mutually adjusted models differed for men and women. For men, the summative index encompassed extremely low and low birthweight, small for gestational age, small head circumference, and post-term birth for non-violent criminal convictions, with the addition of preterm birth for violent criminal convictions. For women, the summative index encompassed of low birth weight, small head circumference, and post-term birth. The implication is that certain adverse perinatal events, when accounting for all others, are more important for men than for women and vice versa. We opted for an empirically driven approach that allowed for different adverse perinatal events to be included in the summative indices for men and women respectively. Our results highlight the need to differentiate between men and women in the study of adverse perinatal events in relation to criminal convictions, since men and women likely are at different risk of experiencing adverse perinatal events (Zeitlin et al., 2002), as well as engage in criminality. This approach was further supported by the HR estimates for the summative index in the total population (Table S3 in Supporting Information), which in general were not significant or in most cases exhibited overlapping CIs. The dose-dependent relationship between adverse perinatal events and offspring criminal behavior should be explored more extensively in future research to better understand the differences between men and women.

Some unexpected findings in our study should also be noted. Certain adverse perinatal events were associated with a decreased risk of violent and non-violent criminal convictions (e.g., breech presentation, assisted vaginal delivery). Breech presentation, as well as other adverse perinatal events, have previously been related to a lack of oxygen to the fetus (i.e., anoxia), which in turn has been described as a risk factor for criminal offending (Tibbetts, 2011). No study thus far has specifically examined breech presentation in relation to criminal offending, let alone different types of breech positions (e.g., frank breech, complete breech). Further research using other samples is needed to explicate the role of different breech presentations and other aspects of delivery on risk for later criminality.

In our sensitivity analyses, we stratified the full sample including men and women based on levels of SES. Previous work has reported associations between exposure to adverse perinatal events in combination with psychosocial adversities, such as maternal rejection (Raine et al., 1994; Raine et al., 1997) and a disadvantaged family environment (Piquero & Tibbetts, 1999) and later criminal behavior in the offspring. However, a few exceptions should be noted though (Murray et al., 2010; Murray et al., 2015). In the current study, HR estimates were largely unaffected by stratifying the sample into levels of SES, if anything they were somewhat attenuated for all groups.

Findings from the present study should be considered in the light of certain limitations. Some of our reported associations, particularly those for the female portion of the sample, need to be interpreted with caution because of the small number of individuals being exposed to particular adverse perinatal events, as well as the small portion of criminal offenders among females. Owing to these factors, HR estimates for females were less precise than those for males, with wider confidence intervals.

Another limitation is that, although the great majority of births in Sweden are recorded in the MBR, there are still 1–3% of all births during each of the past 20 years that are missing from the register (the National Board of Health and Welfare, 2021). While the MBR contains information of varying quality, the adverse perinatal events included in the present study have previously shown high validity (Källén & Källén, 2003). It is also important to acknowledge that our criminal convictions data relied on official records, which may not be representative of all men and women who have engaged in criminal activity. While registry data reduces the risk of misclassification in one way by limiting recall bias that is often associated with interview data, the results in the present study assume the same level of misclassification of criminal convictions for men and women. While more research is needed on this specific topic, especially in relation to registry data, there is some evidence for a more lenient treatment of female offenders as compared to male offenders (Doerner & Demuth, 2014). Lastly, we performed sensitivity analyses in the total sample of men and women, stratified on levels of SES. Ideally, this would have been done in men and women separately but was not possible due to statistical power restrictions.