Introduction

Despite the existence of strong evidence for lifestyle modifications and for screening and preventive actions to improve health outcomes, an implementation gap exists due to limited physician time [1], conflicting/unclear guidelines, and difficulties inherent to sustained behaviour change [2]. The original BETTER (Building on Existing Tools To Improve Chronic Disease Prevention and Screening in Primary Care) intervention was designed to address this gap by providing an integrated approach to increasing uptake of chronic disease prevention and screening (CDPS) actions using a framework of shared decision-making between patient and practitioner. In a pragmatic cluster randomised control trial (RCT), the BETTER approach improved the uptake of CDPS actions against heart disease, diabetes and several cancers (colorectal, breast and cervical cancers) by 32.5% in urban primary care settings in Alberta and Ontario, Canada [2, 3]. The intervention consisted of an individual prevention-focussed visit between participants aged 40–65 years and a “Prevention Practitioner” (PP), who used principles of motivational interviewing to empower the participant to set achievable prevention and screening goals, based on the harmonization of evidence, which were then recorded in a goals sheet and a personalized ‘prevention prescription’.

There have been subsequent modifications of the BETTER approach with similar positive results. ‘BETTER 2’ targeted the same age group as the original BETTER but modified the approach for different populations due to equity concerns, including individuals from rural, lower income, or historically marginalized backgrounds in Newfoundland and Labrador and the Northwest Territories, Canada [4]. Subsequently, BETTER WISE (Building on Existing Tools to Improve Cancer and Chronic Disease Prevention and Screening in Primary Care for Wellness of Cancer Survivors and Patients) tailored the BETTER approach for cancer survivors (breast, colorectal, prostate) aged 40–65 and also included screening for poverty, as well as an updated literature review to recommend specific prevention and screening actions [5]. Another modified version, BETTER HEALTH: Durham used a public health-led model with public health nurses serving as PPs for 40–64 year-olds living with lower income in Durham, Ontario, and found a 53% increase in completed health actions (immediate intervention, n = 60 vs. wait-listed arm, n = 66) [6, 7]. Although there were positive results for this age group, the community advisory group for BETTER HEALTH: Durham suggested that starting the intervention at 40 years of age was too late for people living with low income, where evidence shows an earlier onset of chronic diseases [8]. We aimed to adapt the BETTER HEALTH: Durham intervention to a new population of adults aged 18–39 years living with low income, a group known to have earlier mortality due to, and higher prevalence of, preventable chronic diseases than their peers with higher income.

To support the adaptation, we conducted a review of guidelines to identify and assess prevention and screening actions for health issues and risk factors amenable to individual change for the 18–39 year age group. This paper describes the methods and results of the literature review.

Methods

Overview of search strategy

First, we assessed the data sources (clinical practice guidelines) from the most recent BETTER WISE study [9], which had entailed a rigorous evidence review process to recommend specific prevention and screening actions, for applicability to adults aged 18–39 years.Then, we used a structured grey literature search of specific repositories and websites to find relevant clinical practice guidelines for new topics suggested by the research team. If guidelines were unavailable for these topics, we performed a systematic literature search in the databases Ovid Medline, CINAHL (Cumulated Index to Nursing and Allied Health Literature), and the Cochrane Database of Systematic Reviews to identify systematic reviews/meta-analyses. Thus, our search and eligibility criteria for new sources was restricted to clinical practice guidelines (i.e. excluding systematic reviews, meta-analyses, and review of reviews when guidelines were found) and expanded to allow systematic reviews and meta-analyses when guidelines were not available (See Fig. 1).

Fig. 1
figure 1

Search strategy for guidelines for BETTER LIFE

Full size image

Search strategy for topics of interest

To create the overall search strategy, we consulted an experienced information specialist (CZ). We used different combinations of key words such as ‘guidelines’, ‘chronic disease prevention’, ‘prevention’, ‘clinical practice guidelines’, and ‘screening’ with terms from topics of interest from previous versions of BETTER (cardiovascular disease, diabetes, cancer, obesity, diet and nutrition, physical activity, smoking/tobacco and alcohol use) and new topics suggested by the wider research team (co-investigators, PPs, Community Advisory Committee (CAC)) due to their importance for our target population (See Supp Table 1).

Search sources

We conducted a structured search in repositories of guidelines at the provincial level (Ontario, Alberta, Newfoundland & Labrador): Cancer Care Ontario; Cancer Control Alberta; Eastern Health Cancer Care Program; and national level: Health Canada; Public Health Agency of Canada (PHAC); and the Canadian Task Force on Preventive Health Care (CTFPHC). (Details in Fig. 1).

We did not find guidelines for four topics recommended by our study team for our target population (speeding, texting & driving, seat belts, bullying & cyberbullying). Therefore, we then conducted a systematic search on select databases (Ovid Medline, CINAHL, Cochrane Database of Systematic Reviews) for systematic reviews and meta-analyses published from 2008-August 2020 on these topics.

Inclusion & exclusion criteria

When screening abstracts obtained from our searches, we included articles for full-text review if they met the following criteria: clinical practice guidelines in English only; published from 2008–2020; country of publication was Australia, Canada, England, Ireland, New Zealand, Scotland, or US; included at least one of the identified topics in title or abstract; and addressed prevention or screening.

At full-text screening, we excluded articles if they met any of the following: exclusively focused on management or treatment; exclusively targeted ages not 18–39 years old (i.e., under 18, 40 or older); lacked individual-level recommendations (i.e. contained only macro-level data (e.g. legal, policy)); or lacked evidence of synthesis. With the exception of the four topics covered during the systematic search, we also excluded full-texts if they were systematic reviews, review of reviews, or meta-analyses.

During full-text screening, if multiple eligible sources existed, we used a hierarchical approach to determine inclusion: preference for most recent Canadian guideline/ review and if not available, relevant guidelines from any of 6 aforementioned primarily English-speaking countries of interest. If there were discrepancies or disagreements among guidelines, we searched for and extracted information from primary or common references.

All abstracts and full-texts were uploaded and screened using Covidence [10].

Quality assessment

We chose AGREE-II for quality assessment since it was developed specifically for assessing quality of existing practice guidelines, unlike GRADE (Grading of Recommendations, Assessment, Development, and Evaluations), which is most suited for developing guidelines de novo and for rating primary sources of evidence for specific outcomes, which was outside the scope of our study. We used a two-step process to assess guideline quality. For the first step, two trained reviewers (NM and SC) independently used a shorter 2-item AGREE-II [11] rating system to assess the “Rigour of development” (items 7 and 12—‘Systematic methods were used to search for evidence’ and ‘There is an explicit link between the recommendations and the supporting evidence’, respectively) on all references. If methodological details were missing from guidelines, we emailed authors or guideline developers to request more information. Both reviewers had to assign a score of 4 or higher (out of 7) on both AGREE-II items for the article to move to full quality assessment with the 23-item AGREE-II tool.

Specifically, the reviewers examined the ‘Methods’ section of each guideline to assess the details of systematic methods (Item #7) that were used and consulted any methods papers that governed the overall initiative when available [12,13,14,15,16,17,18,19,20]. If the guideline developer did not report any evidence of an independent synthesis as per the first step in the AGREE–II screening process, the guideline was not assessed further. If no Canadian reference met the criteria for the 2-item AGREE-II screening tool on a given topic, the reviewers then assessed the quality of the non-Canadian documents. Disagreement over scores was discussed and a final decision was determined by consensus.

For step 2, two reviewers independently applied the full AGREE-II instrument on all guidelines that passed the 2-item screening. Overall scores of 5 and above (out of 7) by both reviewers were used to move to full data extraction phase. To ensure consistent interpretation of data quality, we pilot tested the full AGREE-II tool on 5 articles that had previously been included in BETTER WISE and that also met the eligibility criteria for BETTER LIFE.

Data extraction

Two reviewers also pilot tested the data extraction form on 5 articles and resolved differences by consensus. Each reviewer independently extracted data from half the included articles and then checked a subset from the other reviewer for consistency, resolving differences by discussion. Extracted data included publication details (issuing body/author, year and country of publication), participant characteristics (target population, age, ethnicity, socioeconomic metrics, identified risk factors, clinical context) and guideline details (individual-level recommendations, quality of supporting evidence, and whether conflict of interest was declared or not).

Harmonization and synthesis of extracted data

The extracted data were grouped by topics. Each article was assigned to two reviewers who independently either categorized recommendations for inclusion in BETTER LIFE or excluded them if they were duplicative, out of scope, or not actionable (See Fig. 2).

Fig. 2
figure 2

Harmonization process for BETTER LIFE

Full size image

The reviewers met to discuss and assign a final primary categorization to each recommendation with the overall team meeting to resolve differences if there was no agreement between reviewers. The senior co-authors (AL and MAO) reviewed the categorizations, clarified unclear recommendations and identified specific recommendations for further review from content experts/co-investigators in the BETTER team.

Harmonization and synthesis

We followed a similar harmonization process to Campbell-Sherer et al [9] within an overarching ADAPT-ITT framework [21].

All the co-investigators and PPs in the BETTER team were invited to provide input on topics in which they had expertise and asked to rank the newly included recommendations in an online survey (Qualtrics, Provo, UT), with the goal of reaching consensus on the top ranked (most relevant) recommendations. Recommendations ranked with a mean of 90% or above were included, while those that were that consistently ranked low (mean of less than 75%) were removed. For topics with multiple individual recommendations with mean scores of 80–89%, we combined, summarized and simplified the multiple recommendations where it seemed appropriate to do so and included them.

After the harmonization process, we compiled the final list of recommendations and topics into a table and also grouped all related included topics into existing or new ‘domains’ for data visualization.

Results

There were 864 abstracts, of which 762 were unique. Of these, 435 were moved to the full-text phase and assessed for inclusion. One hundred and eighty-five guidelines met the inclusion criteria for quality assessment (Fig. 3a).

Fig. 3
figure 3

a Summary flow from literature search to full-text review for quality assessment. b Quality assessment of guidelines using the AGREE-II instrument to the data extraction stage

Full size image

From the 150 guidelines included in BETTER WISE that were published in 2008 or later, 40 guidelines were applicable to the 18–39 year age group, of which 14 had been updated since inclusion in BETTER WISE. Newer versions were available for the following 8 topics: cancers (breast, cervical, colorectal), CVD, diabetes, obesity, lifestyle (alcohol; healthy nutrition/physical activity).

From the search for topics for which there were no identified guidelines (speeding, texting and driving, seat belts, bullying and cyberbullying), 213 papers were uploaded into Covidence after removing duplicates. However, all the papers on these topics were excluded at various stages.

Quality assessment

One hundred and eighty-five guidelines were eligible for quality screening (Fig. 3b). After exclusion at various stages, 93 guidelines were rated with the 2-item AGREE-II. Of these, 75 were rated with the full AGREE-II tool and 58 papers (77%, 58/75) were scored 5 or higher by both reviewers.

We extracted data from 85 guidelines (58 were new guidelines and 27 were from previous versions of BETTER). Of the 38 new topics (Supp Table 1), 22 were relevant to the 18–39 year age group (Supp Table 2).

Harmonization and synthesis

Of the 19 colleagues invited, 9 responded, reporting expertise on atleast one of the topics on the list (between 1–8 respondents provided ranking on each of the various new recommendations). At the harmonization stage, the team removed the topic ‘falls/injury prevention’ as the recommendation was deemed not in scope for the 18–39 age group.

Due to low ranking scores from Co-investigators, we removed 6 topics from inclusion in the final BETTER LIFE recommendations (intimate partner violence, sexual health, skin cancer, sleep, violence, vitamins). We also excluded hepatitis Cas only one co-investigator provided a ranking for this recommendation, and the recommendation was to not screen for hepatitis C. On the advice of the research team, we also included screening for Adverse Childhood Experiences (ACE) [22, 23].

Based on the results of the data extraction and harmonization, the final list of topics contained 42 recommendations for 18–39 year-olds across 21 total topics (Table 1). We grouped the final list of topics into existing or new domains (See Supp Fig. 1).

Table 1 List of all* topics and recommendations for BETTER LIFE participants aged 18–39 years
Full size table

The CDPS recommendations for heart disease and colorectal and breast cancers were only targeted to those deemed ‘high-risk’ (based on various clinical criteria such as family history) in the 18–39 age group. For most of the new topics, we also identified specific maneuvers or screening questions/tools that could be incorporated into the BETTER visits or into BETTER tools.

Discussion

We used a structured search of published and grey literature, and a systematic search of specific databases to compile recent evidence from clinical practice guidelines on risk factors and individual prevention and screening actions relevant to adults aged 18–39 years, particularly those living with low income, in Canada. We also obtained input from our co-investigators, a team of experts in primary care, public health, the social determinants of health, and the BETTER program. Through this process, we were able to identify 42 recommendations within 21 total topics that will be applied in the BETTER LIFE approach for younger adults living with low income.

Some topics and health recommendations from previous BETTER versions were updated or included, such as those addressing diabetes, cardiovascular disease, cancer, smoking, alcohol, nutrition, and exercise. Risk assessments for diabetes, cardiovascular disease and most cancers were similar for those aged 18–39 years old as with previous versions of BETTER, though routine screening was only recommended for those deemed high risk (with the exception of cervical cancer screening). We found evidence-based guidelines addressing new topics relevant specifically to 18–39 year olds grouped into the following new domains: healthy relationships and healthy sexuality, immunization, oral health, social determinants of health, and substance use. Some recommendations in BETTER LIFE were similar to those published by others [97,98,99], though the recency, diversity, and sources of our search; our harmonization and implementation process, as well as the definition of our target population were different. For example, Persaud et. al. developed 15 preventive care recommendations and 1 policy recommendation that promote health equity in Canada. Although their work and ours both prioritize health equity in primary care, Persaud et al. did not have any age restrictions on their target population nor a primary focus on uptake of individual-level preventive actions. They also utilized systematic reviews, primary research articles and randomized controlled trials to develop recommendations using a GRADE approach. Because we prioritized recommendations that were individually actionable, supported by evidence that met our criteria, and ranked highly by content experts, topics like vitamins and skin cancer prevention were eventually omitted. Although we ultimately excluded skin cancer, this topic is an important one in many countries such as Australia [100].

Taking specific contexts into account is important when determining how best to implement and support uptake of the recommendations. For some new topics, we found stronger evidence for resources and screening tools for PPs than for specific recommendations (e.g. the National Institute on Drug Abuse (NIDA) Quick Screen or the Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) for substance use). PPs identified local community resources for some new health topics (parenting; substance use; oral health) which could help to support participants achieve recommended actions. They also suggested considering social contexts as opportunities for engagement, e.g. by focusing conversations in BETTER LIFE visits on the concepts of health promotion or meaningful overall health and social well-being rather than explicit chronic disease prevention; by using different media for sharing health information (e.g. mobile apps, social media or online resources); by considering social contexts as barriers or enablers of behaviour change, especially regarding physical activity, alcohol, substance use; or by taking life stage into account (single adult vs. parenting).

Our study had several strengths and limitations. Our strengths include a rigorous critical appraisal of the literature with a two-step quality assessment process and independent review that ensured that only guidelines that met high methodological rigour and transparency were included for data extraction and harmonization; focus on actionable recommendations (e.g. goal-setting, access/referral to community resources); and meaningful collaborations with diverse community, public health, and clinical experts. However, all the guidelines were published prior to the COVID-19 pandemic, so did not take pandemic-related disruptions and health impact into account. COVID-19 has exacerbated health and economic inequities and disproportionately affected racialized and low income groups with a higher risk of exposure due to living and working conditions; higher prevalence of co-morbidities; inequitable access to testing and treatment; and disruption of health services [101, 102]. We also relied on consensus to resolve disagreements during the screening process and to formulate the final recommendations as well as on voluntary responses during harmonization which led to varied numbers of reviewers for each recommendation, and which may be subject to bias. However, we used AGREE-II to ensure transparency and careful documentation, and also consulted a wide and diverse range of experts (in primary care, public health, the social determinants of health, Prevention Practitioners, and the Community Advisory Committee) at many stages of the project. Finally, we may have missed guidelines because we targeted our search to specific criteria, repositories, and databases.

Conclusion

Adults living with low income are at increased risk of chronic disease. Through critical literature review and guideline harmonization, we have curated a list of individual-level actionable recommendations relevant to prevention and screening for people aged 18–39 living with low income in English-speaking countries.