The aims of this study are:

1. To synthesise and interpret existing prevalence studies, their methods and findings, summarising prevalence estimates in high-income countries like the UK.
2. To synthesise evidence on patterns of drinking and the ‘dose-response’ relationship between alcohol consumption during pregnancy and FASD.
3. To produce recommendations for future research on the prevalence of FASD.

To address the study aims, our approach comprises:

1. A scoping review of existing prevalence studies, describing and critically appraising the methods and synthesising estimates of prevalence to produce a plausible range of prevalence within which the English population is likely to lie.
2. A scoping review of studies of the association between volume of alcohol during pregnancy (‘dose-response’), stage of pregnancy (e.g. by trimester) and patterns of drinking during pregnancy (e.g. binge drinking) and the risks, levels and severity of FASD.
3. An online ‘roundtable’ event with multidisciplinary clinical and methodological experts to advise on possible study designs and methodology for a future prevalence study.
4. Reflecting on a range of options for future research, highlighting practical and ethical concerns and identifying strengths and weaknesses.

Key messages:

• Drinking during pregnancy can affect the developing foetus, risking Foetal Alcohol Spectrum Disorder (FASD). This can result in physical, mental, and behavioural problems for children, including learning difficulties.
• FASD is a preventable condition which can have a significant impact on life chances, and societal costs. • The UK has high rates of alcohol consumption, binge drinking and prenatal alcohol exposure.
• The scale of FASD and its impact on the population are believed to be under-recognised. There are currently no reliable estimates of the prevalence, incidence or cost impact of FASD in England.
• We reviewed published literature on reported FASD prevalence estimates and the relationship between quantity and patterns of alcohol drinking during pregnancy, and risks of FASD.
• Our review of international prevalence studies revealed a very wide range of estimates of national prevalence of FASD, notably affected by differences in study methodology.
  o Large sample studies based on passive surveillance or analysis of birth registries estimated prevalence of less than 1 per 1000 children.
  o Case ascertainment studies in school-aged children estimated much higher prevalence, although there was still a wide range (6 per 1000 to 80 per 1000).
• Given the very different methods used to measure exposure to alcohol, it was not appropriate to synthesise the prevalence estimates.
• Our review of dose-response and drinking patterns illustrates differing methods, drinking measures and FASD diagnosis thresholds, again making formal synthesis inappropriate.
• There is evidence that higher levels of drinking during pregnancy increases the risk of FASD, but the nature of the ‘dose-response’ relationship is unclear, and it is not possible to identify a threshold below which drinking can be viewed as ‘safe’. This supports the position of the UK Chief Medical Officers, who recommend that for women who are pregnant or planning a pregnancy, the safest approach is not to drink alcohol at all.
  o Even if it were possible to identify a clear relationship, sharing estimates of ‘dose response’ could feasibly weaken the clear, understandable current message that any drinking during pregnancy is potentially harmful.
• We consulted academic and clinical experts in FASD in a roundtable discussion about measuring prevalence of FASD.
  o The expert group believed that a national prevalence study is desirable to inform future policy and practice. 1
  o The group favoured an active school-based case ascertainment approach supplemented by statistical correction reflecting likely under-reporting, along with consent for future data linkage and follow-up.
  o This would cost around £1000 per child included, plus costs of training staff to screen children and provide advice and services for those newly diagnosed.
• Such a study would unavoidably have limitations. An existing pilot study was unable to recruit any children from specialist schools providing social, emotional and mental health support, and even in mainstream schools there was a lack of consent in high-risk groups such as children under local authority care.
• Questioning women about drinking during pregnancy requires a detailed and sensitive approach, reducing stigma and encouraging transparency.
• We question whether a more precise prevalence estimate would be an efficient way to influence policy and practice, compared with a modelled prevalence estimate.
  o Would it strengthen the prevention message – particularly given that guidance from the UK Chief Medical Officers clearly recommends total abstinence?
  o Should it be necessary for children to have a diagnosis of FASD to access services?
• We recommend that a first step should be to commission a modelling study, estimating the likely range of prevalence from existing information (including the prevalence estimates in our review).
• This should be supplemented with a value of information (VoI) analysis to assess the value of more research to increase the precision of prevalence estimates, in terms of the likely impact of resulting changes in policy and practice and usefully informing decisions on future research in this area.