Intergenerational continuity of child physical abuse: how good is the evidence?

Background There is widespread belief that individuals who were physically abused during childhood are more likely to abuse their own children than th ose who were not abused, but the empirical studies examining this belief ha ve not been systematically reviewed. The aim of this study was to evaluate systematically, based on eight methodological standards derived from a hypo thetical randomised controlled trial, the design of studies investigating t he intergenerational transmission of child physical abuse. Methods We reviewed studies published between 1965 and 2000 in English that provided information about physical maltreatment in two generations and in cluded a comparison group. Two investigators independently assessed whether each study met the methodological standards. Findings In the ten studies identified (four cohort, one cross-sectional, a nd five case-control), the relative risks of maltreatment in the children o f parents who were abused during childhood were significantly increased in four studies (relative risks 4.75-37.8), but in three other studies the rel ative risks were less than 2. Most study reports provided a clear descripti on of abuse of parents during childhood and abuse of their children. Five s tudies failed to avoid recall and detection bias; five did not ensure that controls were not themselves maltreated; eight did not provide adequate fol low-up; and in six the report did not state whether the enrolled parent was responsible for the maltreatment. Most studies did not control for interve ning factors, such as sociodemographic characteristics during the time of a buse of the parent generation and at the time their children were abused. O nly one study met all eight criteria (relative risk of abuse transmission 1 2.6 [95% CI 1.82-87.2]) and one met six (1.05 [0.53-2.06]). Interpretation The one study that met ail eight methodological standards pr ovided evidence for the intergenerational continuity of child physical abus e, but that which met six standards did not support the hypothesis. Use of our model and methodological standards should improve the scientific qualit y of studies examining the effects of risk factors for adverse outcomes tha t may continue across generations.