Review Reports - The Potential for Fetal Alcohol Spectrum Disorder Prevention of a Harmonized Approach to Data Collection about Alcohol Use in Pregnancy Cohort Studies

Round 1

Reviewer 1 Report

Thank you for the opportunity to review this study. Asking women about how much alcohol they consume in pregnancy is important for the mother and child. Whilst there is some interesting information in the paper it requires more focus and the specific aims more clarity. If the aim is to identify the most useful way of measuring and comparing ways of asking about alcohol use in pregnancy, there should be some method of comparison between questions/ methods. As it stands the paper is more an audit of methods used in some cohort studies and screening tools than a research question.

I have reviewed the article and outlined my suggestions below.

Introduction

The authors examined several cohort studies to determine the range of questions asked about alcohol use in pregnancy. They note that use of that the confidential research context (that is use of previously collected cohort studies) will be helpful, perhaps because of the anonymity (although this is not specifically stated), in that women are more likely to disclose their drinking patterns during pregnancy. Please justify.

The authors also note that several research cohorts can be bought together to increase sample size so there is adequate representation in regional areas. However, how well these studies can achieve this is not discussed further. That is, the underlying rationale for using these methods is not addressed by analysis in the paper in the paper.

I am unclear as to the purpose of the information about screening. The authors note (page 2 line 71) that this will define how questions about how alcohol use in pregnancy should be asked. In the methods section, however, there is an overview of some of the screening tools that have been used, but no discussion about what might be considered ‘gold standard’.

Materials and methods

First para. What information is taken from the screening tools? How is best practice identified?

What information is presented about provincial perinatal surveillance efforts? The stated aim “We use these findings to make recommendations on guidelines for data collection about women’s alcohol use by future pregnancy cohorts” is unclear. What cohorts are the authors referring to? How will the recommendations be made? Will the validity and reliability of measures be compared?

Page 3 Line 113 This para is unclear. Identification of current practices would require an audit of particular time/place.

Results

The section on current measures of alcohol use in pregnancy in the cohort studies assessed is a description of the wording used in the questions and whether some measures could be combined to increase sample size on particular items. As noted a more critical analysis of this information is needed. Comments on the section on screening and surveillance are similar, each section should be a critical analysis of current practice and specific recommendations.

Author Response

Introduction	The authors examined several cohort studies to determine the range of questions asked about alcohol use in pregnancy. They note that use of that the confidential research context (that is use of previously collected cohort studies) will be helpful, perhaps because of the anonymity (although this is not specifically stated), in that women are more likely to disclose their drinking patterns during pregnancy. Please justify.	Added a sentence to describing confidentiality of the research context and its impact on safety to report.
	The authors also note that several research cohorts can be bought together to increase sample size so there is adequate representation in regional areas. However, how well these studies can achieve this is not discussed further. That is, the underlying rationale for using these methods is not addressed by analysis in the paper in the paper.	A section on the underlying rationale for using these methods has been added to the introduction.
	I am unclear as to the purpose of the information about screening. The authors note (page 2 line 71) that this will define how questions about how alcohol use in pregnancy should be asked. In the methods section, however, there is an overview of some of the screening tools that have been used, but no discussion about what might be considered ‘gold standard’.	There is no gold standard for screening for basic alcohol use in pregnancy. A sentence is added to this effect Note the validated tools for assessing problematic use in pregnancy were validated in 1994, and the Provincial and Territorial (P/T) surveillance experts note that they not well received by women and health care practitioners. Since we are not focussing on problematic alcohol consumption, we made brief reference to this in the paper.
Materials and methods	First para. What information is taken from the screening tools? How is best practice identified?	We have described the screening tools and questions only in this paper – an analysis of the data elicited will be described in a later paper. We make recommendations based on triangulating current practice and standardized screeners for the general population.
	What information is presented about provincial perinatal surveillance efforts? The stated aim “We use these findings to make recommendations on guidelines for data collection about women’s alcohol use by future pregnancy cohorts” is unclear. What cohorts are the authors referring to? How will the recommendations be made? Will the validity and reliability of measures be compared?	The information about provincial perinatal surveillance efforts is described in section 3.1.2. We are referring to the recommendations made in the discussion section, which are indented for both the further questioning of women in the longitudinal studies now in place, as well as any future cohorts funded. Specific recommendations have also been added to the results section.
	Page 3 Line 113 This para is unclear. Identification of current practices would require an audit of particular time/place.	Wording clarified to say we ‘present’ these two types of data collection and discuss them.
Results	The section on current measures of alcohol use in pregnancy in the cohort studies assessed is a description of the wording used in the questions and whether some measures could be combined to increase sample size on particular items. As noted a more critical analysis of this information is needed. Comments on the section on screening and surveillance are similar, each section should be a critical analysis of current practice and specific recommendations.	We reorganized the sections 3.1 and 3.2 to present the information on the standardized screeners and the current perinatal surveillance efforts first in an effort to more effectively make our argument of how these might inform the data collection in the research context. Specific recommendations related to the research context have now been added to each subsection of 3.2 as well as being summarized in the discussion section

Reviewer 2 Report

This is a very well written paper, with a valid rationale and clearly presented methods and results (if somewhat atypical as they are not the standard results of statistical analyses, but more a review of data available and their format). The authors have extensively reviewed how prospective Canadian cohorts collected data on prenatal alcohol exposure, alongside what standardised tools are available and also how Canadian surveillance systems collect this information.

Some minor comments follow.

- The authors push for using cohort data collected in a research context to inform prevention strategies for FASD, citing stigma and other barriers to collecting accurate self-reported data in clinical or surveillance contexts. What evidence do they have that the (potentially heavily selected) population samples participating in cohort studies will provide more valid estimates of prenatal alcohol consumption, and furthermore that these will also be generalisable? Have any efforts been made to triangulate or link the information from cohorts with the information from routine surveillance systems? What about the inclusion or exclusion of certain marginalised groups in the population and in particular ethnic minorities including First Nations? It is also not clear how data from a population based cohort study could directly help diagnose children with FASD or support them (as claimed at the end of the first paragraph of the Discussion) - often there are ethical issues around identifying or diagnosing participants in a cohort study, and what about the informed consent?

- The authors should include mention of the biometric perfomance of different questionnaires and standardised tools for assessing prenatal alcohol exposure, for a complete picture on the 'strengths and limitations' of each tool as claimed in the second paragraph of the Discussion - what is the gold standard? What is known about how each questionnaire/tool fares compared to the gold standard? What is the reliability of each questionnaire/tool? This is an important aspect to consider when deciding which tool to adopt going forward, but also in choosing how to harmonize current data from cohort studies where standard validated tools like the AUDIT-C aren't available.

- Since cohort studies use prenatal alcohol exposure data to provide estimates of risk for various offspring outcomes (not just to report on the prevalence of drinking behaviour), the authors should mention how some tools/questionnaires could bias these estimates, based on their known limitations (eg a tool what is known to lead to systematic underreporting/overreporting of alcohol use could bias estimates in this or that direction etc....).

- A more extensive discussion of the issues of cultural sensitivity, diversity and inclusion in relation to prenatal alcohol exposure questions should be included.

- The authors don't comment on collecting information about partner's drinking behaviour before and during pregnancy. This is another important and related set of variables, and could per se influence offspring risk, but at present few datasets include these data (to this effect, see https://www.ncbi.nlm.nih.gov/pubmed/30055422).

Author Response

Reviewer 2:

	- The authors push for using cohort data collected in a research context to inform prevention strategies for FASD, citing stigma and other barriers to collecting accurate self-reported data in clinical or surveillance contexts. What evidence do they have that the (potentially heavily selected) population samples participating in cohort studies will provide more valid estimates of prenatal alcohol consumption, and furthermore that these will also be generalisable? Have any efforts been made to triangulate or link the information from cohorts with the information from routine surveillance systems? What about the inclusion or exclusion of certain marginalised groups in the population and in particular ethnic minorities including First Nations? It is also not clear how data from a population-based cohort study could directly help diagnose children with FASD or support them (as claimed at the end of the first paragraph of the Discussion) - often there are ethical issues around identifying or diagnosing participants in a cohort study, and what about the informed consent?	While the data will not likely provide a more valid estimate of prenatal alcohol consumption, they have the advantage to provide the opportunity to collect longitudinal information on alcohol consumption: before, during and after pregnancy. While the information collected will not be fully generalizable, cohorts are more useful to investigate associations. This has been addressed in the introduction. To our knowledge no efforts have been made to triangulate the data from clinical surveillance and research sources. Unfortunately, we do not have good data from either of these sources, and that is why we saw it a beneficial to prepare an article that offers recommendations for improvement in the research context. (We are also working with the P/T groups in a national committee) We are preparing a second article on the findings from the research cohorts, by demographic status and agree this is very important. We changed the sentence to show that we are not using this data in diagnostic efforts, but simply that knowing of maternal use is a criterion of FASD diagnosis and that ultimately knowing of levels, timing and frequency of alcohol consumption will be important in the long run.
	- The authors should include mention of the biometric performance of different questionnaires and standardised tools for assessing prenatal alcohol exposure, for a complete picture on the 'strengths and limitations' of each tool as claimed in the second paragraph of the Discussion - what is the gold standard? What is known about how each questionnaire/tool fares compared to the gold standard? What is the reliability of each questionnaire/tool? This is an important aspect to consider when deciding which tool to adopt going forward, but also in choosing how to harmonize current data from cohort studies where standard validated tools like the AUDIT-C aren't available.	We cannot provide the biometric performance, as the screeners such as the AUDIT C have not yet been validated with pregnant women, only the tools assessing problem use such as the TWEAK have. The TWEAK was validated in the 1990s when, for example, it was more acceptable to ask women if others have criticized them for their use, rather than asking them directly of their own assessment of how much their alcohol use had interfered in their lives. We see that pregnant women are able to report on their alcohol use with the same basic screening questions, as all women are, and that there are benefits to keeping the questions consistent, especially between the preconception and pregnancy periods. We are hoping that the AUDIT C, with efforts in Canada and Australia will become the gold standard.
	- Since cohort studies use prenatal alcohol exposure data to provide estimates of risk for various offspring outcomes (not just to report on the prevalence of drinking behaviour), the authors should mention how some tools/questionnaires could bias these estimates, based on their known limitations (eg a tool what is known to lead to systematic underreporting/overreporting of alcohol use could bias estimates in this or that direction etc....).	We can only advocate for best practice in asking about alcohol use in the research context, where over and under reporting is less likely – and hope that this will improve understanding of how/if alcohol use is associated with various outcomes The data from FASD diagnostic clinics which is now being consolidated in a national database, will hopefully serve to illuminate the linkage between levels and patterns of alcohol use and specific alcohol-related impairments.
	- A more extensive discussion of the issues of cultural sensitivity, diversity and inclusion in relation to prenatal alcohol exposure questions should be included.	We added a point in the discussion in this and agree it needs always to be emphasized
	- The authors don't comment on collecting information about partner's drinking behaviour before and during pregnancy. This is another important and related set of variables, and could per se influence offspring risk, but at present few datasets include these data (to this effect, see https://www.ncbi.nlm.nih.gov/pubmed/30055422).	We agree that is important information to collect, and we are currently preparing research summaries for the CanFASD Research Network on this topic. We felt it beyond the scope of this paper

Reviewer 3 Report

Well done to the authors. This is a well written and conceptualised study that has enormous potential to inform both policy and research. However there needs to be more said about this. For example, is standardisation possible in the clincal practice context? If so, how might this be addressed at a national level? What can be learned from efforts elewhere around the world e.g. the Australian context? Measuring risk in a standardised way can inform prevention efforts how? This should be addressed explicitly in the introduction and discussion sections.

Some further points for the authors to address:

Section 2.2 Please specify how both the structured screening tools and questions used in Canadian surveillance methods were identified. E.g. was a key word search conducted in the former and stakeholder consultations in the later?

The results section 3.1.1 and beyond (inclusive of Table 2) is not clear for the reader. In each section a number of cohorts are identified (e.g. 8 in 3.1.1, 10 in 3.1.2 and so on) however only 7 cohorts exist in Table 2. This needs clarity.

Lines 245 through 253: It is clear how the information is collected, but not how it is documented. Please address.

Line 305 – Once again, more needs to be said about how this might happen as per the initial comments.

I support accepting this paper for publication once the above comments are addressed.

Author Response

Reviewer 3:

Introduction	Is standardisation possible in the clinical practice context? If so, how might this be addressed at a national level? What can be learned from efforts elsewhere around the world e.g. the Australian context? Measuring risk in a standardised way can inform prevention efforts how? This should be addressed explicitly in the introduction and discussion sections.	Standardization in the clinical context is theoretically possible, but not realistic it requires all provinces and territories and other governments to agree to a common approach. We are working with the P/T surveillance experts on evidence informed, standardized questions related to cannabis use in pregnancy and hope there will be opportunity to discuss the importance of evidence informed, standardized approach to the alcohol questions as well. Of course, even if we have standardized questions there remain the challenges of getting them asked and answered in the clinical context, which to date has been an elusive goal. Ways in which the harmonization can improve the evidence to inform prevention efforts has been added to the introduction.
Materials and methods	Section 2.2 Please specify how both the structured screening tools and questions used in Canadian surveillance methods were identified. E.g. was a key word search conducted in the former and stakeholder consultations in the later?	Additional information about the methods used to identify screening tools and Canadian surveillance methods has been added to the Materials and Methods section.
Results	The section on current measures of alcohol use in pregnancy in the cohort studies assessed is a description of the wording used in the questions and whether some measures could be combined to increase sample size on particular items. As noted a more critical analysis of this information is needed. Comments on the section on screening and surveillance are similar, each section should be a critical analysis of current practice and specific recommendations.	We reorganized the sections 3.1 and 3.2 to present the information on the standardized screeners and the current perinatal surveillance efforts first in an effort to more effectively make our argument of how these might inform the data collection in the research context. Specific recommendations related to the research context have now been added to each subsection of 3.2 as well as being summarized in the discussion section
	The results section 3.1.1 and beyond (inclusive of Table 2) is not clear for the reader. In each section a number of cohorts are identified (e.g. 8 in 3.1.1, 10 in 3.1.2 and so on) however only 7 cohorts exist in Table 2. This needs clarity.	This is explained in section 2.2, and after the in-text introduction of the table in 3.2.1. Only seven cohorts included information about the actual question used to gather the data in the ReACH Catalogue, the remaining five only include data labels indicating the content of the variable, not the actual question. A note has been added to the table for clarification.
	Lines 245 through 253: It is clear how the information is collected, but not how it is documented. Please address.	We have noted this in the Methods section and in the Results section 3.1.2. The P/T get physicians to document information on the antenatal records, and the P/T representatives reported to the Centre of Excellence as to what info they gather and the challenges in getting this data.
	Line 305 – Once again, more needs to be said about how this might happen as per the initial comments.	Additional information on the potential benefits of data harmonization and co-analysis have been added to this section.