May 2, 2014

survey of adverse childhood experiences and health-harming behaviours

A paper published in BMC Medicine suggested a link between adverse experiences and events in childhood, and health harming behaviours as an adult. This analysis was accompanied by a roundup.

Title, Date of Publication & Journal

National household survey of adverse childhood experiences and their relationship with resilience to health-harming behaviors in England.

2 May 2014, BMC Medicine

Claim supported by evidence?

The paper does not prove the claim that adverse childhood experiences lead to later problems in life. The paper suggests relationships, but the p-values are unreliable.

Summary

• The analysis takes no account of the sampling – 10 regions, 200 geographical areas. The statistical model must take the regions and geographical area into account. Unfortunately this means that the statistical analysis is flawed and the p-values could be highly misleading.
• It is unclear how the modelling calculates the proportion nationally of the events due to adverse childhood experiences. These figures should not be relied on without further details.

Study Conclusions

The paper observes strong relationships. However, the p-values are unreliable.

Strengths/Limitations

Strengths

• They take into account major confounding variables
• Large scale study
• Detailed breakdown of figures.
• Observed relationships in the expected direction.

Limitations

• Observational – we cannot determine cause and effect. It is not possible to say that altering adverse childhood experiences would alter the outcomes. There are liable to be (many) other factors playing a part.
• Self-reporting – there could be biases and distortions in self-reporting:
• Some individuals may be more likely to report slaps as abuse. If this happens for both childhood experiences and outcome variables, this could distort the observed relationships.
• Did the authors check for the “wonky vote”? – Some individuals (“for a laugh”) may report as having everything. Even a small proportion could distort the results.
• The questionnaire is about an individual – it is not clear how (an) individual was selected within a household.
• 16,000 households initially selected – 4,010 completed questionnaire. So approximately 25% of initially selected. Table 3 shows that distribution across deprivation

Before The Headlines is a service provided to the SMC by volunteer statisticians: members of the Royal Statistical Society (RSS), Statisticians in the Pharmaceutical Industry(PSI) and experienced statisticians in academia and research.  A list of contributors, including affiliations, is available here.