Introduction

Cannabis use during pregnancy is on the rise within the USA. Recent legislative changes related to cannabis regulation and legalisation have made cannabis available for purchase through legal means within several states in the USA shifting social norms around use.1 Correspondingly, a recent study found that pregnant women perceive cannabis use to be low-risk.2 While exact prevalence rates of cannabis use can differ depending on factors like study population, criteria for use and data-collection methods,3 4 evidence consistently indicates an overall rise in cannabis use during pregnancy in the USA.5 6 For instance, data based on self-reported use from a representative survey in the USA showed past month cannabis use rates for pregnant people doubled between 2002–2003 and 2016–2017 when adjusted for age, race/ethnicity and income, going from 3.4% adjusted prevalence to 7.0%.7 This increase could lead to a set of unanticipated consequences in health outcomes for parents and children, which are only minimally understood. For example, cannabis use during pregnancy has been associated with an increased risk of adverse parent–child health outcomes. Specifically, cannabis use in pregnancy has been linked to preterm birth, low birth weight and being small for gestational age8 and cannabis exposure during pregnancy has been associated with affective symptoms and attention deficit hyperactivity disorder in children.9 Additionally, tetrahydrocannabinol, the main ingredient in cannabis that produces psychoactive effects, can be passed to infants through breast milk.10 It is important to note that there are limitations to the current data on cannabis use during pregnancy, which limit understanding. Specifically, some research has been conducted in animals that can be poor surrogates, and studies in humans are confounded by use of multiple substances and other sociodemographic factors which can also lead to adverse outcomes, making it difficult to isolate the contribution of cannabis specifically.11 While outside of the scope of the current review, the field would thus benefit from more research on the impact of perinatal cannabis use to ascertain a fuller understanding of the risks. However, overall trends in current data are worrisome. Thus, the American College of Obstetricians and Gynaecologists recommends against the use of cannabis during pregnancy and breastfeeding (‘perinatal cannabis use’).11 However, there are few updated guidelines on best practices for intervention in the context of increasing normalisation of use among pregnant people.

Supporting patients in changing their perinatal cannabis use behaviours requires a nuanced approach that goes beyond education about health risks. This is not to understate the importance of provider-delivered education about cannabis use. This education is an important intervention and is itself an area rife for future research as current provider-delivered education about perinatal cannabis is limited by a lack of provider training.12–14 However, education alone is likely insufficient to support behaviour change around perinatal cannabis use. Some patients, even after being educated about the risks, continue to use cannabis due to its perceived benefits, such as increasing relaxation, improving sleep and reducing anxiety.15 16 In some cases, pregnant people see these perceived benefits as outweighing the health risks,17 18 making it difficult to discourage use. Addiction research has also shown that abstinence-only approaches have limited effectiveness for such patients, as trying to ‘right’ an individual’s behaviours or restrict their choices can induce reactivity that leads them to be more firmly set in their behaviours.19–23 Therefore, for patients who are unwilling or unable to cease use, alternative methods to abstinence-only approaches are needed.

One evidence-based approach for promoting health when complete substance use cessation is not possible is harm reduction. Unlike abstinence-only approaches, in cases where abstinence is not feasible, harm reduction views any improvement or reduction in harm as a clinical success.24 Harm reduction has been supported by the Substance Abuse and Mental Health Services Administration,25 and the clinical guidelines for the perinatal use of other substances, like nicotine and opioids, have incorporated harm-reduction approaches.26 27 Institutions including the White House and the National Institute on Drug Abuse have also promoted harm reduction for substance use,28–30 and there is evidence that perinatal patients are already employing harm-reduction strategies for cannabis use.18 Additionally, perinatal patients have expressed interest in harm-reduction education,31 and training literature created in Canada32 33 indicates medical providers too want guidance on the application of harm reduction to perinatal cannabis use. As such, harm reduction may not only be the best evidence-based treatment available for individuals who are unable or unwilling to completely discontinue use but is also increasingly asked for and endorsed across the patient, provider and institutional levels.

Although existing reviews have examined patient and provider perspectives as well as provider approaches to perinatal cannabis use, major gaps remain in the field’s understanding of harm reduction for perinatal cannabis use. One prior review has focused on counselling practices by clinicians and providers regarding perinatal cannabis use.34 Another has reviewed the literature on cannabis with respect to pregnant women’s experiences and thoughts regarding use.17 A third has focused on clinical approaches to reduce cannabis use.35 A fourth examined educational resources for medical providers about cannabis use and reproduction.36 However, to our knowledge, no prior review has culled concrete clinical practices that comprise harm reduction for perinatal cannabis use. A likely contributor to this lack of understanding is the fact that the concept of harm reduction is not consistently defined, and strategies which comprise harm reduction may not always be labelled as such. This makes it challenging to comprehensively collect articles using search terms meant to pull for harm reduction specifically.

This scoping review will address this gap by looking at the scope and coverage of articles describing harm reduction clinical practices in the context of perinatal cannabis use. We will do so by collecting all references discussing perinatal cannabis use that have been published since the discovery of the endocannabinoid system, and then screening them for references describing concrete clinical practices that comprise harm reduction for this population. As perinatal cannabis use is an area of research with emerging evidence, the methodology of a scoping review is ideal to get an overview of the current state of research, summarise existing findings, and help identify the particular areas where current research is lacking.37 38 Thus, by systematically mapping the state of the literature, this scoping review will advance our understanding of the existing evidence on harm reduction for perinatal cannabis use and provide avenues for future clinical research. It is important to note that, when operationalising harm reduction in this study, we are focused on approaches that can be used when a patient is unwilling or unable to completely abstain from cannabis use. Thus, while abstinence can be part of a harm reduction approach, we are focused on identifying practices to reduce harm when cessation from cannabis is not possible. As mentioned, knowledge of such practices is particularly needed as not all patients are able to completely abstain, and at present the field lacks an understanding of what approaches can be applied to promote health in these cases.

Methods and analysisOverview

This review is guided by the JBI Manual for Evidence Synthesis39 and the Arksey and O’Malley methodology for scoping reviews,38 as updated by Levac and colleagues. In line with this guidance, an iterative and reflexive approach will be taken, particularly during the screening and data-extraction phases as these may become more refined as we engage with reference materials. The protocol follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses for Scoping Reviews (PRISMA-ScR).40 A systematic search of the literature was initially conducted to identify English language articles authored between January 1990 and 2023 present in these databases as of 22 September 2023: PubMed (National Library of Medicine), Embase (Elsevier), Web of Science Core Collection (Clarivate), APA PsycINFO (EBSCO), CINAHL (EBSCO) and the Cochrane Central Register of Controlled Trials (CENTRAL) (Wiley). Preceding submission of the results for publication this search will be run again to ensure capturing later publications relevant for the review. Google search results will also be hand-searched for patient-facing materials. Additional grey literature sources include clinical trials, preprints and conference proceedings that were not excluded from the database search results. Literature will be included if it describes clinical approaches involving harm reduction for perinatal cannabis use. Two reviewers will independently conduct title and abstract screening, followed by a full-text screening of references meeting the title/abstract criteria. Data extraction will focus on clinical practice descriptions, data-collection dates, publication details, legal status of cannabis in the data-collection location and reported outcomes associated with harm-reduction practices. The quality of the eligible studies will be assessed using the MMAT.41 42

The protocol, and any potential revisions, has been registered and can be accessed through the Open Science Framework (OSF Registries) (https://osf.io/wb3jc). Covidence will be the software used for screening and tracking decisions.43

Stage 1: identification of the research question

The purpose of the review is to understand the state of the field regarding clinical practices comprising harm reduction for perinatal cannabis use. As mentioned, the concept of harm reduction is not consistently defined, and strategies which comprise harm reduction may not always be labelled as such, which makes it challenging to comprehensively collect articles using search terms meant to pull for harm reduction specifically. In addition to describing the state of the field, this review will allow for the identification of key clinical research priorities to further development of harm reduction approaches. Accordingly, our research question is: What is the scope and coverage of existing harm reduction approaches for perinatal cannabis use? It is important to note that, as mentioned above, we are specifically focused on concrete clinical strategies that can be applied to reduce harm when complete cannabis cessation is not possible, as this is an area with a particular lack of understanding at present.

Stage 2: identification of relevant studies

Using a search strategy advised by a health sciences librarian (third author), we will locate all articles associated with cannabis and the perinatal period (pregnancy and lactation) and screen that corpus for the types of approaches that comprise harm reduction. We will ‘bookend’ our search from 1990 to the present, as the 1990s saw the discovery of the endocannabinoid system,44 and the first passing of legalised medical cannabis in the USA.45 The sociocultural and legal landscape regarding cannabis has been rapidly changing since this time, with increasing numbers of states legalising cannabis and decreasing public perception of harm.46 47 Our initial search was 1992 to 2023, to identify English language articles authored between January 1990 and 2023 and present in these databases as of 22 September 2023; however, preceding submission of the results for publication, this search will be run again to ensure capturing later publications relevant for this review.

As we are seeking a comprehensive understanding of existing research and clinical guidelines related to harm reduction for perinatal cannabis use, empirical (original research, reviews) and grey literature documents (conference proceedings, conference papers, preprints and patient-facing materials) will be included in the search. We have searched the following key databases for medicine and social science for empirical peer-reviewed materials: PubMed, Embase (Elsevier), Web of Science Core Collection (Clarivate), APA PsycINFO (EBSCO), CINAHL (EBSCO) and Cochrane Central Register of Controlled Trials (CENTRAL) (Wiley). In line with guidance from the Cochrane Handbook for Systematic Reviews of Interventions,48 both Medline and Embase will be included and searched to find clinical studies, with PubMed used to find additional articles not indexed in Medline. The other databases will be searched to include relevant subject-specific information, with APA PsycINFO selected for its behavioural and substance use content, CINAHL for its extensive collection of clinical journals covering allied health fields and Web of Science Core Collection for its broad geographic scope. Query strings for each database can be found in online supplemental file 1. Of note, the strategy pairs controlled vocabulary of subject headings (Medical Subject Headings (MeSH), Emtree and APA Thesaurus of Psychological Index Terms, etc) available for each database with relevant keywords appearing in title, abstract and author keywords. For grey literature, Cochrane CENTRAL includes clinical trials, and Google and Google Scholar will be hand-searched for patient-facing materials and preprints (OSF Pre-Prints, MedRxiv and BioRxiv). In addition to articles, preprints and conference proceedings were included among results of the empirical database searches.

As a search validation procedure, a ‘test set’ of nine articles was selected, which the study team knew the search strategy would have to be sensitive enough to identify. After running the search, we checked that these articles were included in the results. As mentioned, to ensure we have not missed useful findings, prior to publication of the results of the search we will run our search terms again.

Stage 3: screening and study selection

A three-step review process will be undertaken to select references. In step 1, RIS files will be exported from each database into EndNote for data management and de-duplication.49 References will then be uploaded to Covidence for screening.43 Duplicates potentially missed by EndNote will be identified and removed after upload to Covidence. References will then be prescreened by a trained research assistant (second author) with training and supervision (by first author). Prescreening will exclude basic science work, such as biochemistry of cannabis and work performed in animal analogues, so the remaining references only include behavioural and clinical studies conducted in humans. Any additional duplicates found will also be manually removed.

Step 2 entails titles and abstract screening, which will be conducted by two reviewers (first and second author) using Covidence. A title/abstract is eligible for full-text screening if it:

Discusses the perinatal population (inclusive of pregnancy and breastfeeding/lactation)

Discusses cannabis use (inclusive of polysubstance use so long as cannabis is among the substances mentioned)

Seems likely to have description of examples or recommendations of harm reduction treatment approaches. These are operationalised as involving one or both of: (a) concrete actions that an individual can take that reduce health harm associated with use (modifiable risk factors; a la the lower risk cannabis use guidelines50) and (b) principles guiding healthcare provider delivery of an intervention (a la harm reduction principles for healthcare settings51). Articles describing clinical approaches that are abstinence-only will be excluded. Articles including only description of correlates of perinatal cannabis use or outcomes associated with perinatal cannabis use will also be excluded.

After completion of title/abstract review of the empirical literature, the two reviewers will complete title/abstract review of the grey literature. For all references except those collected from hand-searching, title/abstract review will be completed using Covidence. 43 For references collected from hand-searching, title/abstract review will be tracked using a shared spreadsheet.

Step 3 will be a full-text review of any articles found to be relevant by both reviewers. Only those studies that describe specific clinical practices to reduce harm stemming from cannabis use in the population, without necessary discontinuation of use, will be kept. Of note, in the case of references discussing polysubstance use, references will only be considered eligible if they specifically discuss strategies to reduce harm stemming from cannabis use. We will not include articles where cannabis use is itself conceptualised as a harm-reduction strategy (eg, a pregnant person could decide to consume cannabis to stop consumption of another substance with more evidence of deleterious effects). During steps 2 and 3 any divergence in screener decisions will be resolved through discussion, and if disagreements persist, a third screener will be brought into the discussion.

Backward searching will be employed for any sources identified for full-text review, to identify sources cited in these texts and confirm their inclusion in the search results. Web of Science and PubMed will be used for forward searching to identify any relevant sources that have cited the text since its publication.

Stage 4: extraction of data from selected studies

Our data-extraction approach is influenced by a similar review on perinatal cannabis conducted by Panday et al.34 The information that we plan to extract covers basic information as well as the description of the harm reduction practice(s) and any reported outcomes associated with use of the harm reduction practices (see table 1).

Table 1

Data extraction for references

The extraction will be performed by a trained research assistant (second author). To ensure that extraction is performed correctly, the research assistant will first extract information on a subset of studies, and the study lead (first author) will review and provide feedback. This process will be repeated until information can be consistently extracted correctly, after which the research assistant will extract information from the remaining studies. If the article does not have enough information for extraction, we will reach out to the author for additional information and ask for permission to include this information in our extraction. If there is no response, we will attempt to contact them a second time a week later. All outreach efforts and results will be tracked and reported as appropriate.

Stage 5: collating, summarising, appraisal and reporting results

We will use the MMAT to assess the quality of references reporting on original research.41 42 This has been chosen as we expect a mixture of qualitative and quantitative data, and because it has been applied in similar work.34 For references not reporting on original research (such as guidelines and patient-facing materials) we will: (1) search the references for citations backing up their harm-reduction recommendations and (2) use a simplified evidence hierarchy grading system to rate the quality of the evidence backing up the recommendation.52 Evidence will be graded as follows: A (High Quality): Systematic reviews, meta-analyses, RCTs, B (Moderate Quality): Cohort studies or case-–control studies, C (Low Quality): Cross-sectional studies, case series, or case reports, D (Very Low Quality): Opinion or anecdotal evidence.

For synthesising findings, we will employ Sandelewki’s ‘qualitizing’ approach.53 This will result in a series of declarative ‘qualitized’ statements that synthesise the reference findings. This ‘qualitizing’ approach is based on methodology from systematic reviews and synthesis studies53 54 and is similar to the approach used previously in another review on perinatal cannabis.17 The statements should be understandable on their own and include both the findings and any relevant contextual information.17 As was done in similar previous work,17 this process will be carried out by two synthesists, and each statement will be written by one synthesist, and then verified by another synthesist. If there are any disagreements in the synthesis, a third study team member will be consulted.

Stage 6: consultation

After collating and summarising results, the authors will seek consultation from experts such as the Champlain Maternal Newborn Regional Program (located in Ontario) and Perinatal Services British Columbia, which developed harm-reduction practice resources for perinatal cannabis use in Canada.32 33 Additionally, one of our study team members has expertise in clinical interventions for harm reduction, substance use and health behaviour change, they (fourth author) will consult with other professionals in these fields for feedback on our preliminary results. These contacts can ensure that we are capturing the full scope of this topic and they can direct us towards other sources to include in our review. Data will be extracted from any additional sources and results re-collated and re-summarised to produce our final results.

Patient and public involvement

Patients/the public have not been directly involved in this project thus far. However, a member of our study team (fourth author) is a knowledge user of this scoping review, who currently provides care to perinatal individuals who use substances and has an extensive history of working with non-perinatal substance-using populations. We have engaged this author as a knowledge user and leveraged that information through contributions to the development of the scoping review questions and approach. This author will continue to lend their expertise during interpretation and dissemination phases.

Ethics and dissemination

Ethics approval was not sought as this review does not constitute data collection on human subjects (no information or specimens were collected from interaction or intervention with an individual).

We plan to engage in traditional dissemination approaches such as peer-reviewed publication as well as scientific presentations. For example, the first author annually attends a premier conference on behavioural medicine. Additionally, as members of our study team are actively engaged in treating perinatal patients (first author) and specifically substance-using perinatal patients (fourth author), we will also utilise our networks to disseminate these findings to clinical providers and patients. Additionally, at time of publication, spreadsheets and tables resulting from appraisal and synthesis will be deposited in the OSF Repository associated with project.