Estimating child abuse numbers can overload social workers and put more pressure on families

Lauren Devine and Stephen Parker argue that estimated prevalence data of child abuse increases reporting, puts extra pressure on families but doesn't increase detection

Dr Lauren Devine & Mr Stephen Parker, Bristol Law School, University of the West of England, Bristol

The Children’s Commissioner’s report, Protecting Children from Harm: a critical assessment of child sexual abuse in the family network in England and Priorities for Action presents, as the authors stress, tentative and indicative data about the prevalence and incidence of child sexual abuse in the family network.

This follows the NSPCC’s 2011 report authored by Radford et al estimating the prevalence of child maltreatment. The Children’s Commissioner’s report uses Radford et al’s data although neither report can provide a confident estimate of prevalence.

Regardless, the reports influence public perception and policy, but are they helpful and how are the effects of policy change evaluated?

Undoubtedly public and professional perception of the prevalence of child abuse is important as it influences the culture in which professionals operate. Belief in a high level of undetected child abuse (sexual or otherwise), particularly in the family unit, creates a receptive culture for policies encouraging detection, reporting and intervention.

Recent policies are directed towards increased reporting, particularly via identifying ‘signs of abuse’, to be reported primarily by anyone who is under a section 11 Children Act 2004 duty.

At first glance this is a step forward in reducing child abuse. However, deeper examination raises questions: unintended consequences are revealed in the government data of reduced system efficiency, and numerous research studies indicate an increased risk of alienation of families feeling accused as opposed to supported, there is also a risk of social work overload and the fear of missing some of the estimated high number of undetected cases.

Policing parents

The report’s conclusions suggest there are a large number of undetected abusing families that need to be ‘identified’ and ‘prevented’. This conclusion legitimises further policy measures to police parents.

However, from the evidence of government annual statistics this is likely to result in a progressively more overloaded and less efficient system of referral and assessment if the negative impact of current policy is not taken into account. The hard data from government annual statistics is collated and reported in Rethinking Child Protection Strategy: Learning from Trends, which shows a counter-productive outcome from this prevailing policy drive.

There have been two key outcomes from recent policy, both noted in Rethinking Child Protection Strategy: Learning from Trends: Firstly there has been a progressive and statistically significant year-on-year drop in the proportion of child abuse (i.e. the number of substantiated cases as a ratio of referrals) that is addressed via social work involvement. The number of referrals has quadrupled since 1991/1992 but the number of referrals which amount to cases social workers consider to be abuse has remained relatively stable.

No increase in abuse detected

Secondly, the widely noted unreliability of identification of abuse from ‘signs’ has consistently been seen as a secondary question to the importance of referring on the basis of ‘concerns’. As a result, families are under more intense scrutiny now than ever before in England.

Despite the surveillance, prevalence estimates do not reflect a drop in abuse, and matching the prevalence estimates against the policy outcomes does not show any significant increase in the amount of abuse detected.

The implication of policy decisions reacting to prevalence estimates is fraught with complexity. In the Children’s Commissioner’s report the actual statistical analysis and its confidence limit is not included. The report explains that the analysis uses MSE (Multiple Systems Estimation).

This involves making assumptions about the probable level of overlap in recorded cases in different agencies (the police, social services and voluntary agencies) from a sample in time and geographic area. From this data a ‘dark number’ can be estimated. The detailed analysis used in the report is not published.

Therefore, not only does the level of undetected abuse remain a ‘dark number,’ the precise evaluation method used on the raw data is similarly ‘dark’.


Although this data collection and analysis is theoretically interesting, and attracts sensationalist reporting, it does little to reliably inform policy, or to increase public understanding of this complex area.

The media’s response to the 450,000 estimate of apparent undetected child sexual abuse in the family network was unhelpful. Not only was the estimate quoted as actual, rather than estimated prevalence data of child sexual abuse between close family members (not the less sensational ‘wider family network’), but it unsurprisingly resulted in one-sided calls for ‘urgent action’ to ‘identify and prevent’ sexual abuse.

The real power of prevalence estimates is not in finding out how much child abuse there is, but its use in persuading policy makers and the general public that policies and funds to detect, investigate and prevent child abuse are justified by increasingly intrusive means.

The question of balance within the system is not considered, nor is fairness towards social workers, who manage an increasing caseload in an uneasy culture where blame can easily be assigned.  Nor is the impact on families adequately considered if ‘parent’ becomes increasingly synonymous with ‘suspected abuser’.

Balanced policy must take account of positive and negative impacts when responding to these estimates. At present it is not clear that it does.


More from Community Care

Comments are closed.