Breast cancer screening best practices: a Canadian multidisciplinary consensus-based approach for primary care providers.

Introduction
Professional associations recommend population-based screening for breast cancer (BC); however, they differ in the age of initiation (40 vs 50 years), frequency of screening (annual vs biennial) and estimates related to benefits and harms of screening.14
Most guidelines recommend shared decision-making (SDM) starting from 40 years, although some state this more explicitly than others. The 2024 US Preventive Services Task Force (TF)1 recommendations endorse routine screening at age 40, based on modelling data, increased BC incidence, and disparities in BC outcomes in women of diverse race and ethnicity (though acknowledging the lack of primary data demonstrating that screening impacts on these outcomes). The Draft 2024 Canadian TF on preventive healthcare recommendations2 states that women aged 40–74 be informed of personal risk factors, benefits and harms of screening and that if they want screening, they should receive it; however, they state women aged 40–49 should not be screened systematically. Both TFs acknowledge limitations and uncertainty in the primary data and advocate for further research.
Pathways to access screening differ in each Canadian province; the majority have shifted from a primary care provider (provider)-based referral system for ages 40 to 49 to one in which women are able to self-refer from age 40.3 4 These pathways decrease the barriers to accessing screening (especially in those without providers); however, there is no published data on SDM use and quality prior to screening.
The variability of guidelines causes confusion for providers who are responsible for discussing BC screening and SDM. In Ontario prior to 2024, differences in screening patterns appear based on providers’ predefined practices rather than an individual’s informed decision.57 Guideline clarity and consistency in messaging across multidisciplinary groups are important determinants of guideline-concordant care.8 Further, most guidelines are intended for women at average or moderately increased risk of BC. Therefore, providers need to be able to differentiate women with a high lifetime risk of BC in whom guidelines do not apply. Following these important care principles ensures that women receive individualised care throughout the spectrum of risk.9
There are many aspects of ‘BC-related health’ that experts from different disciplines agree on. If these are performed routinely and comprehensively, this could improve the quality of care for all Canadians. The objective of this study was to reach consensus on specific Canadian-focused ‘best practice behaviours’ related to BC health and screening to serve as a resource for providers.

Methods
The group lead (senior author MBN) and an individual with lived experience of BC convened a multidisciplinary group of BC screening experts (online supplemental table 1). Each expert was selected purposively based on meeting all three of the following criteria: (1) recognised expertise in their respective field for all disciplines related to BC (surgical oncologist, plastic surgeon, radiation, medical oncologist; radiologist; primary care physician), (2) holding secondary expertise in another area (hospital leadership, national level influence, research methodology, knowledge translation/implementation) and (3) publicly documented research and/or public engagement with the topic of BC screening. In addition to meeting all three criteria, individuals were selected across a spectrum of viewpoints related to BC screening, including individuals known to hold differing and in some cases opposing perspectives within public, academic, or policy discourse. We deliberately created a group with individuals holding a balanced spectrum of viewpoints to align with our objectives of demonstrating consensus across heterogeneity and therefore to enhance credibility for the consensus-based approach once it is translated to real-world clinical practice.

Patient and public involvement statement
An individual with lived experience of BC acted as co-lead to help inform project goals, recruit multidisciplinary members, attend and participate in all meetings and act as a non-voting member of the expert panel. Additionally, four patient partners (with and without a history of BC) were among the external reviewers. The final manuscript is available to patient partners on request. The best practices are publicly available on a website for those who create a log-in (https://canadianbreastpractices.ca).

Objective setting
In the first meeting (October 2022), the project lead provided background on the barriers to guideline-concordant screening. Each group member provided ideas for objectives that could help overcome these barriers, which were grouped into themes. The group agreed on two objectives: (1) build consensus on key ‘best practice’ behaviours related to BC-related health and screening and (2) build consensus on specific definitions related to BC screening. No consulting in regard to method took place.

Nominal group technique and Delphi (October 2022 - February 2024)
To achieve objective 1, the facilitator led the group through two consensus development methods: a modified Nominal Group Technique and Delphi method.10 11 The former aims to structure interactions within a group and is useful when there is expected to be differences in opinions.12
Online supplemental table 2 provides an overview of all meetings and steps.
For the second meeting (Jan 2023), the Nominal Group Technique was used. Prior to the meeting, participants were encouraged to independently record ideas about several ‘bare minimum recommendations related to BC screening’. During the meeting, the facilitator collected one idea at a time until all ideas were shared. No ranking took place as the goal was not to develop a hierarchy. After this meeting, the lead grouped the statements into five categories.
Between January 2023 and November 2023, voting rounds 1 and 2 took place (followed by meetings 2 and 3, respectively). The facilitator used the ‘bare minimum statements’ created by the expert group to create a modified Delphi vote.11 12 For the first vote, all best practice statements (in their respective five categories) were listed. Group members were asked to vote using the Qualtrics platform on each best practice statement using a 5-point Likert scale to rank (a) level of evidence and (b) feasibility. They were not aware of each other’s responses. Responses were collated, and summaries were shared. Statements with mean scores greater than 3/5 for level of evidence and feasibility and low disagreement moved forward to the next round, and all other statements were discussed by the group via meeting 2. A second vote took place with all best practice statements that moved forward (some modified, some grouped) using the same format as the first vote.
In the third and subsequent rounds of voting (Dec 2023–Feb 2024), participants voted dichotomously to ‘recommend’ or ‘not recommend’ each statement. If the latter was chosen, experts were prompted to provide their reasoning. Based on literature and discussion, the group unanimously agreed that a statement would meet a threshold of consensus if at least 80% of the group voted to recommend it.13

Definitions and companion resources
Definitions were taken from the Canadian Partnership Against Cancer,14 an independent, not-for-profit organisation to facilitate action on cancer control in Canada, and supplemented with other resources as needed. Examples of terms that were defined include terms related to screening rates, cancer incidence rates, lead time, length time bias, over-diagnosis and shared-decision making. An environmental scan took place to review tools, links, decision aids and websites available related to BC screening. Group members agreed to include all Canadian-produced BC screening decision aids, commonly used validated risk assessment calculators and referral links (eg, to refer to high-risk screening and genetics) and to create tables to supplement the best practice statements where no other tools existed. Group members provided feedback on the definitions and companion resource items until all were satisfied; no voting took place.

External review (March–May 2024)
Once consensus was reached on all statements, the multidisciplinary group identified and invited multidisciplinary external reviewers to review and provide feedback on the statements and companion resources. Each group member identified other experts in their field or in related fields, and their expertise spanned across the following roles: all types of oncologists, medical imaging experts, geneticists, policy and primary care. All suggested individuals were invited to participate by email and (once they signed a confidentiality agreement) were emailed a draft of the best practices (including preamble and companion resources) and a link to a questionnaire hosted on Qualtrics. Within the questionnaire, external reviewers were asked to vote to ‘agree’ or ‘disagree’ with each best practice statement (and to provide feedback in an open text box if they disagreed) and rank and provide feedback on the clarity, usefulness and accuracy of the information in the preamble and each category and general feedback on the resources. All invited external reviewers (whether or not they participated in external review) could suggest other reviewers, and all suggested reviewers were invited in a similar fashion.

Final best practices (June–September 2024)
Following external review, meetings occurred to review feedback, adjust ‘best practices’ and fine-tune the companion resources. A fifth vote took place (‘recommend’ or ‘not recommend’) after the group reviewed the feedback. Outstanding disagreements were resolved via email discussion until consensus was reached. Most of these situations represented a misunderstanding of the intent of the best practice and were resolved with slight modification of the wording of the statement. In one situation, one expert felt strongly about a recommended practice, but others provided evidence recommending against its uptake. In another situation, agreement was made to indicate within the best practices that one expert disagreed.

Results

Nominal group technique and Delphi
Best practice behaviours were grouped into five categories (prior to considering screening, risk and risk factors, discussion about BC screening, decisions and referrals for BC screening, screening outcomes). See table 1 for details on voting results and online supplemental appendix for the full Best Practices and Companion Resources. After the first round, 8/35 statements (23%) met the threshold to move to the next round. Within category 1, roundtable members agreed that it was important to provide clarity on diagnostic imaging (as compared with screening) as individuals may ask ‘Do I need a mammogram?’ as a way of alerting their provider about a breast symptom. If these individuals had instead indicated the presence of a symptom, it is likely that assessment and/or diagnostic work-up would have been timelier and more targeted.15 The group agreed to recommend the Canadian Association of Radiologists statement section related to any practices for breast disease diagnostic work-up.16 If no level 1 evidence existed for a statement but was a standard medical practice (eg, assess for a diagnostic concern prior to proceeding with screening), the group agreed to consider this ‘high’ for level of evidence.
For the second vote, 9 participants voted on 34 statements. Within category 2, the group agreed that risk factor assessment is required to understand if the guidelines apply to the individual (guidelines apply to those at ‘above or slightly above average risk’); however, to decrease the burden on providers, a formal risk calculation was not recoommended. Within category 3, the group unanimously agreed that SDM was important but disagreed on who should initiate this and what data should inform the conversation. The term ‘SDM’ was chosen (over ‘informed decision-making’) based on its use and definition from the literature; the term ‘sharing’ refers to the ‘sharing of information’—the decision to screen or not lies with the individual. The group agreed to include any available Canadian-based validated decision aid and to create a table of benefits and harms of screening using both absolute and relative numbers but not modelling data as this is not helpful to primary care. A definition of ‘over-diagnosis’ was included for the harms of screening as this term is not well understood.
After the second round of voting, two members left due to conflicts occurring outside the group. These members held opposing interpretations of the screening literature, so balance remained within the group after they left. Both were offered to participate in the external review. Meeting 3 took place to discuss areas of disagreement. The strength of evidence was balanced with clinical feasibility through the group discussions and modification of the statements.
Seven participants completed the third Delphi vote. Overall, 46/48 (96%) best practice statements met the 80% threshold to recommend. Group discussion occurred on the statements that did not pass and those without unanimous agreement. In the fourth Delphi, votes took place on statements that did not pass and statements that were changed based on discussion. Consensus was reached on all statements. By the end of the fourth vote, consensus was reached on a total of 48 statements.
The updated 2024 US guidelines3 were published during the Delphi process. The group agreed that the process should continue since the ‘best practices’ cover important breast health and screening-related issues that are not covered in guidelines. Data used in informing both the updated US and Canadian guidelines were reviewed. The group discussed studies showing an increased incidence of BC at younger ages and in different ethnic groups. The group agreed that this is a very important consideration, but that it does not necessarily imply that there is a benefit from screening. As such, the group agreed to include more information in the resources about differences in BC incidence by ethnicity, where Canadian data were available.

External review
Fifty-five external reviewers were invited, and 29 agreed to participate. The reviewers included four patient partners (with and without a BC history), one medical oncologist, one radiologist, one physicist, one scientist (medical genetics), two BC surgeons, two plastic surgeons, one radiation oncologist and 16 primary care physicians (mix of academic/community and four who currently or previously have held policy roles). See table 1 for results. Overall, 43/48 (90%) of statements met the threshold of agreement. Feedback themes were notable for clarification of the eligible and target populations, best practices related to equity, diversity and inclusion, titles of the categories, imaging for men and trans-populations, age/interval for risk assessments, questions/comments about breast density, risk factors/risk calculators, screening interval and use of supplemental screening. Particular attention was paid to the burden on primary care providers with issues raised to concern about lack of primary care providers, time constraints and responsibilities. In these situations, either the language was softened and/or a larger range was placed on timeline or frequency to allow primary care providers more flexibility.

Final vote
Following external review, 18 statements were added for a total of 66 statements, of which 34 were voted on due to being new or modified. Of these, three statements did not meet the 80% threshold. One statement was removed from category 3. Consensus was reached through individualised communication. The final document consists of 65 statements.

Discussion
We provide 65 best practice statements and companion resources related to breast screening for primary care providers. Strengths of this study include using rigorous expert consensus methods (nominal group and Delphi technique), prespecified agreement of 80%, multidisciplinary group and extensive external review. To our knowledge, this is the first study to create a comprehensive document distilling the latest evidence to provide practical Canadian-focused consensus-based advice for providers.
Several guidelines on BC screening programmes exist across the world, with considerable overlap between the Canadian and US TFs in their recommendations.17 Specifically, SDM between an individual and their provider is emphasised in preventative and screening practices in North America, Europe, Japan and other countries.1822 As such, although our focus is Canadian, we believe these best practices can be a useful resource for a more complete and individualised approach to breast health and BC screening globally and especially in resource-limited settings.
Areas of debate included the definition of SDM, frequency of BC screening, and the role of supplemental screening. SDM is an integral component of the Canadian guidelines and the best practices, so it is essential to have clarity about its meaning. Our group proposed a definition taken from the literature as ‘a process in which both the patient and healthcare professional share information to inform a patient’s decision’. Healthcare professionals share the best available evidence about the benefits and harms of the available intervention, and patients share their values and preferences. The final decision made by the patient is respected. A recent Delphi study identified core elements of SDM in BC screening, through consensus among women aged 40–49, providers and health decision scientists.23 Similar to these best practices, the authors emphasise the importance of transparent discussion of benefits and harms and that the individual makes the final decision. Related to the frequency of screening and use of supplemental screening, the group agreed that both more frequent BC screening and supplemental screening lead to a decrease in interval cancers; however, to date, studies and modelling demonstrate an increase in cases of overdiagnosis and no studies describing impact on survival.24 While we recommend an approximate biennial screening interval for individuals not at high risk, we summarise the available evidence and limitations.
We included a diverse group of experts throughout the process to demonstrate multidisciplinary consensus, as recommendations can vary among healthcare specialists of different backgrounds.25 To help bridge the gap from specialist recommendations to primary care and ensure practical, feasible recommendations, we included two primary care physicians in the roundtable and 16 in the external review. Barriers to accessing primary care and time in a primary care encounter were noted and discussed several times; the roundtable tried to decrease the burden on primary care where possible and included some best practices targeting policy makers to try to improve these gaps. We also included some patient-facing resources for those without access to a provider.
The best practices were reviewed by experts with experience in breast health in transgender populations, including language and recommendation review. Attention was paid to address issues faced by equity-deserving groups, such as indigenous populations, those with low socioeconomic status, immigrants, and racialised groups. These groups have unique barriers to accessing BC screening, limited evidence base and different clinical presentation, including earlier age of onset and more aggressive disease at diagnosis. Because of this, a tailored approach may be more appropriate for individuals belonging to these populations.2628
Our study has limitations. First, our expert group was selected and primarily Ontario-based without a rigorous application process, and after the second vote, two members left the expert panel (22%). Attrition is an expected limitation of such studies, particularly when several rounds of voting take place.29 We expect that the limitations of the multidisciplinary group (selection, smaller number) were made up for by having a large diversity of opinions and expertise both within the group and across a large number of external reviewers across Canada. Second, the best practice statements were often limited by limitations in the primary data itself and/or disagreements on how to interpret the primary data. In these situations, if the group could not agree on a best practice, this (and the reason why) was noted and different guidelines were listed for reference. Third, we have not tested implementation, acceptability or usefulness of these best practices; this remains as a next step in our work. We have planned a pilot mixed methods implementation study across one academic and one community primary healthcare clinic to allow refinements prior to wider dissemination. The best practices are publicly available online at https://canadianbreastpractices.ca.

Conclusion
Our expert consensus group was able to successfully generate a collection of statements and companion resources which comprehensively address all relevant Canadian-focused breast health and screening practices. This document should serve as a resource and summary document for providers to determine who is eligible for average/high-risk BC screening and how to proceed with SDM and screening referrals for these individuals. Implementation and evaluation studies are ongoing and may provide further information to increase uptake and use of this work.

Supplementary material
10.1136/bmjopen-2025-107264online supplemental table 110.1136/bmjopen-2025-107264online supplemental file 1