Recruitment Strategies for Lung Cancer Screening: An Umbrella Review Protocol

Lung cancer and the role of screening

Lung cancer is the leading cause of cancer-related mortality worldwide, accounting for approximately 1.8 million deaths annually¹. Despite advances in treatment, five‐year survival rates continue to range between 10–20%, largely owing to late-stage diagnosis^2,3. Low-dose computed tomography (LDCT) lung cancer screening (LCS) has emerged as an effective intervention; controlled trials have demonstrated that LDCT screening can reduce lung cancer-specific mortality by 20%^2,3. As a result, multiple international guidelines now recommend risk-stratified LDCT screening for high-risk individuals, typically long-term smokers⁴.

The United States Preventive Services Task Force (USPSTF) recommends annual LDCT screening for adults aged 50 to 80 years with a minimum 20 pack-year smoking history, provided they currently smoke or have quit within the past 15 years⁵. Similarly, the European Commission’s 2022 Council Recommendation prioritises the implementation of LCS pilot programmes across EU member states, recognising the need for targeted recruitment strategies to optimise participation.

Global policy and implementation challenges

Despite robust evidence supporting LDCT screening, its implementation remains fragmented, with substantial disparities in screening availability and participation across different healthcare systems. The United States has been at the forefront of LCS policy development, with structured screening programmes integrated into national guidelines. However, even within well-established screening systems, participation remains low, often below 5% of eligible individuals⁴.

In contrast, European countries have been more cautious; only a handful—such as Poland, Croatia, and Italy—have committed to nationwide LCS initiatives. Most European nations rely on opportunistic recruitment, often through primary care or hospital-based settings, contributing to low participation rates. In Asia and other regions, the lack of organised screening programmes, coupled with epidemiological differences—such as a higher proportion of lung cancer cases among never-smokers—further complicates the adoption of standardised screening criteria⁶.

Beyond technical and policy issues, infrastructural constraints such as limited CT scanner availability, insufficient radiology expertise, and inadequate follow-up systems further impede large-scale implementation. Financial constraints and lack of reimbursement policies further exacerbate disparities, particularly in lower-resource settings⁷. These challenges have been amplified by strains on healthcare systems following the COVID-19 pandemic, delaying progress in LCS programme development⁷.

Disparities in screening participation

Even within countries where LCS guidelines are in place, participation remains critically low. Whereas breast and colorectal cancer screening programmes often achieve participation rates of 60–75%, lung cancer screening rates are typically below 15%⁴. Notably, populations at highest risk—heavy smokers, individuals from lower socioeconomic backgrounds, and ethnic minorities—are also those most likely to be under-screened.

Several structural, psychological, and informational barriers contribute to low LCS participation. Structural factors include limited access to screening facilities, transportation difficulties, and financial constraints, disproportionately affecting underserved populations⁴. Psychological barriers—such as fear of diagnosis, stigma associated with smoking-related diseases, and fatalistic attitudes towards lung cancer—further deter individuals from engaging with screening programmes⁸. Additionally, informational deficits—including lack of awareness of LCS eligibility criteria and misconceptions about its risks and benefits—reduce uptake³.

Studies in the United States and Europe have highlighted that residents of the most deprived neighbourhoods have higher lung cancer incidence and mortality rates yet are less likely to participate in screening⁹. This underlines the urgent need for targeted recruitment strategies that address both practical and psychological barriers to LCS participation

Comparative effectiveness of recruitment strategies

The effect of recruitment interventions on lung cancer screening participation will be examined using a simplified social ecological classification adapted from Bronfenbrenner’s ecological systems theory¹⁰ and McLeroy et al.’s framework for health behaviours¹¹. To enhance practical relevance, interventions will be grouped into three categories:

Research from other cancer screening programmes (e.g., breast and colorectal cancer screening) suggests that combining personalised outreach with patient navigation services is likely to yield the highest participation rates¹². However, direct evidence from LCS recruitment remains limited, with most studies focusing on single recruitment interventions rather than comparative effectiveness across diverse populations.

Epidemiological and economic considerations

The global burden of lung cancer reinforces the need for effective recruitment strategies. An estimated 2.5 million new cases are diagnosed annually, with a disproportionate number of patients presenting at advanced stages¹³. Early detection through LDCT screening has the potential to significantly improve survival outcomes by shifting diagnoses to earlier, more treatable stages¹¹.

Economically, lung cancer represents a major healthcare expenditure. Cost-effectiveness analyses indicate that well-targeted LCS programmes, particularly those with optimised recruitment strategies, fall within acceptable cost per quality-adjusted life year (QALY) thresholds¹⁴. Early detection reduces treatment costs by identifying cancers before they require expensive therapies¹⁵. Moreover, targeted recruitment of underserved populations could yield additional economic benefits by reducing the financial burden of late-stage lung cancer treatment.

Justification for an umbrella review

The broad range of recruitment strategies available with their respective comprehensive syntheses removes the need for another similar systematic review. Rather, current literature lacks a focus on high-risk subgroups in the context of LCS. Therefore, our plan is to conduct a rigorous review of systematic reviews (i.e., an Umbrella Review) to provide an overview of the existing interventions, and comment on which appear to work best when stratified by population subgroups¹⁶.

Aims and objectives

This umbrella review aims to evaluate the effectiveness of recruitment strategies for lung cancer screening (LCS), particularly among populations with historically low participation rates, via the following specific objectives:

1.   To identify and assess the quality of all systematic reviews examining the efficacy of recruitment strategies for lung cancer screening (LCS).

2.   To classify identified recruitment interventions according to both intervention level (i.e., system-level, provider-level, patient-level), intervention modality (e.g., digital outreach, GP referrals, direct invitations), and healthcare setting (e.g., primary care, hospital-based programmes, regional/national initiatives)

3.   To summarise the evidence on the efficacy of each intervention type, using change in LCS participation rates as the primary outcome.

4.   To stratify the effect of recruitment strategies by population subgroups, such as:

5.   To summarise the impact of recruitment strategies on secondary outcomes, including:

6.   To conduct a sensitivity analysis to explore the influence of methodological assumptions regarding:

7.   To produce a narrative synthesis that integrates key insights from the identified reviews, structured by population subgroup, to highlight effective recruitment strategies for each subgroup.

Protocol and registration

This umbrella review follows Joanna Briggs Institute (JBI) guidelines¹⁷ and adheres to the Preferred Reporting Items for Overviews of Reviews (PRIOR) checklist¹⁶. The proposed methods are pre-registered on the Open Science Framework (osf.io), with further details provided in this a priori publication.

Eligibility criteria

Systematic reviews are included if they meet the following criteria:

Eligible studies must be systematic reviews with or without meta-analysis. Scoping reviews, rapid reviews, and systematic narrative reviews will be considered if they meet predefined systematic review criteria, including a clear research question, a reproducible search strategy, eligibility criteria, risk of bias assessment, and systematic data synthesis.

Studies will be excluded if they:

Information sources and search strategy

A comprehensive literature search will be conducted across MEDLINE (via Ovid), Embase, Scopus, Web of Science, Cochrane Library, CINAHL, and systematic review registries (PROSPERO, Joanna Briggs Institute Database). The search will include studies published from 1 January 2011 up to 31 October 2024 in the English language. Studies dating before 2011 will be excluded as LCS has yet to be established as an evidence-based practice before this time (i.e., before the publication of the National Lung Screening Trial)¹⁸.

The search will be structured around three core concepts: (1) lung cancer, (2) screening, and (3) recruitment strategies. Medical Subject Headings (MeSH) terms and free-text terms will be combined using Boolean operators (AND/OR), with filters applied to identify systematic reviews and meta-analyses.

A tailored search strategy will be developed for each database using the SR-Accelerator Polyglot tool to optimise retrieval. Reference lists of included reviews will be hand-searched to identify additional relevant studies¹⁹.

Study selection

Search results will be imported into Rayyan for screening and data management²⁰. Duplicate studies will be automatically removed. Two reviewers will independently screen titles and abstracts, with discrepancies resolved through discussion or third-party arbitration if necessary. Full-text articles will be assessed in the same manner, and the study selection process will be documented using a PRISMA flow diagram²¹.

Screening algorithm

A structured four-step screening process will be applied:

Data extraction

A standardised data extraction form has been developed and piloted to ensure consistency and completeness. Two independent reviewers will extract data, with discrepancies resolved through discussion or third-party adjudication. Extracted information will include study characteristics (authors, year of publication, country, type of review, and databases searched), recruitment strategies (description of interventions, mode of delivery, and comparator groups), target populations (general screening populations and high-risk subgroups such as smokers and individuals from lower socioeconomic backgrounds), and reported outcomes. The primary outcome is LCS participation rates. The secondary outcomes are screening adherence, knowledge and awareness of LCS, patient experience, or barriers and facilitators to participation, and cost-effectiveness, where available.

Data synthesis

The umbrella synthesis will be conducted in three phases:

Phase 1 (Summary of Eligible Reviews)

Phase 1 will involve a structured summary of the included reviews, conducted in three steps:

This structured approach will ensure clear documentation of review characteristics, methodological details, and both the scope and timing of primary studies.

Phase 2 (Efficacy of LCS Recruitment Interventions)

Phase 2 will focus on analysing the primary outcome—lung cancer screening participation—through a structured examination of effect estimates. This phase will involve the following steps:

These analyses will enable the identification of intervention types associated with increased participation in lung cancer screening, while attempting to account for differences in review quality, healthcare settings, and population characteristics, thereby guiding more robust recommendations for LCS implementation.

Phase 3 (Summary of Secondary Outcomes)

Phase 3 will consist of cataloguing the secondary outcomes reported across the included reviews. These outcomes will be systematically categorised into the following domains:

No additional risk of bias assessment specific to these secondary outcomes will be conducted, as the eligibility criteria of this umbrella review focused exclusively on reviews assessing the primary outcome (participation rates). Since studies evaluating these secondary outcomes independently of participation were likely excluded, a resource-intensive expansion of the risk of bias assessment to an incomplete set of reviews cannot be justified. Instead, Phase 3 provides a structured summary of these secondary outcomes so readers can appreciate the potential ancillary effects, both positive and negative, of interventions aimed primarily at enhancing participation in lung cancer screening.

Risk of bias assessment

The methodological quality of included systematic reviews will be assessed using AMSTAR 2²², which evaluates the transparency, reliability, and risk of bias in systematic reviews. Reviews will be rated as high, moderate, low, or critically low quality. For systematic reviews that include meta-analyses, the ROB-ME tool¹⁹ will be used to assess the risk of bias due to missing evidence, including selective outcome reporting and small-study effects.

Sensitivity analysis

Sensitivity analyses will assess the robustness of our findings by testing key methodological assumptions. Additional assumptions may emerge during the review process, and corresponding sensitivity analyses will be added as required. This flexible approach will ensure that the influence of subjective judgment on our conclusions is transparent. However, two sensitivity analyses are anticipated:

•   Quality Assessment: Results will be compared when excluding only critically low-quality reviews versus excluding both low- and critically low-quality reviews (i.e. when using AMSTAR 2).

•   Primary Study Overlap: To address primary study duplication, we will explore methods such as weighting reviews by unique studies, excluding reviews with extensive overlap, or prioritising reviews with stronger synthesis methods.

Certainty of evidence

The GRADE framework will be applied to assess the certainty of evidence for key outcomes²³, using the standard classifications of High, Moderate, Low, or Very Low certainty. However, since this is an umbrella review, our GRADE assessments will reflect the quality of the evidence from systematic reviews and meta-analyses rather than direct evaluation of primary studies. The five GRADE domains will be employed:

•   Risk of Bias will be assessed based on the methodological quality of the included systematic reviews, using AMSTAR 2 and ROB-ME as outlined above. We acknowledge that primary study limitations may not always be adequately reported. Consequently, we will rely on reviewers' assessments of primary study risk of bias where reported and highlight cases where such information is absent or inconsistently described.

•   Inconsistency will be assessed by evaluating the variability in effect estimates across included reviews. Where meta-analyses are available, heterogeneity metrics such as I² will inform our judgment. For reviews without meta-analysis, inconsistency will be identified based on narrative descriptions of conflicting findings or discordant conclusions across reviews.

•   Indirectness will be assessed by comparing the characteristics of the included evidence with the population, intervention, and outcome criteria for this umbrella review. Evidence drawn from incomparable screening settings, population subgroups, or outcome definitions may result in a downgrade for indirectness.

•   Imprecision will be assessed using pooled effect estimates and their confidence intervals where available. Wide intervals that span both meaningful benefit and harm, or intervals that cross clinically important thresholds, will result in downgrading. For reviews without meta-analyses, imprecision will be judged based on the range and variability of reported effect sizes across reviews.

•   Publication Bias will be considered based on assessments provided in the included reviews. In addition, we will evaluate broader patterns in the data, including unexplained discrepancies between expected and reported findings or trends that suggest unpublished negative results.

Strengths and limitations

Our proposed umbrella review will follow a transparent, systematic approach, adhering to Joanna Briggs Institute (JBI) methodology, the PRIOR reporting framework ensuring methodological rigour. The use of AMSTAR 2 for quality appraisal will strengthen the critical evaluation of included systematic reviews, while citation mapping will help mitigate primary study overlap, reducing the risk of over-representing frequently cited evidence.

A key strength is the comprehensive evaluation of recruitment strategies across diverse healthcare settings, examining system-level (e.g., risk-stratified invitations, automated GP referral mechanisms), provider-level (e.g., clinician training, GP-driven recruitment), and patient-level approaches (e.g., patient navigation, community-based outreach). The integration of the McLeroy 1988 framework¹¹ will also enhance understanding of which strategies are most effective in improving LCS participation, particularly among historically under-screened populations.

However, several limitations must be acknowledged. First, restricting inclusion to English-language publications may limit the generalisability of findings, particularly for LCS programmes in non-English-speaking countries, where recruitment strategies and healthcare systems may differ. Second, excluding grey literature and unpublished studies may result in the omission of emerging recruitment interventions, particularly pilot programmes, implementation studies, and government-led initiatives that may not be formally published. Third, umbrella reviews depend on the quality and scope of existing systematic reviews. If included reviews have methodological weaknesses, inconsistent outcome reporting, or insufficient subgroup analyses, this may limit the strength of the conclusions drawn.

Additionally, heterogeneity in study designs, population characteristics, and healthcare settings may complicate direct comparisons of effectiveness. The application of subgroup analyses and stratification by AMSTAR 2 quality ratings will help address these variations, but residual heterogeneity may remain.

Implications for future research, policy and clinical practice

This review will identify and synthesise the literature on the efficacy of LCS recruitment interventions in high-risk subgroups. We aim to highlight underexplored recruitment strategies, such as digital opt-in models, automated risk-stratified invitations, and community-led patient navigation interventions within the 1988 McLeroy framework¹¹.

Our umbrella review of efficacy remains limited in the sense that it does not help us understand “how, in whom and in what specific contexts” the interventions work. A qualitative umbrella synthesis (e.g., utilising a realist approach) could build upon an existing theory. For instance, the I-SAM model could be integrated in this context to propose design of interventions for understudied subpopulations based on a theoretical framework¹⁵.

For policymakers, findings will support the development of targeted recruitment frameworks that can be adapted across different healthcare systems. The review will inform evidence-based recommendations for national lung cancer screening programmes, particularly in Europe, where implementation remains fragmented and opportunistic. Insights into cost-effectiveness and system-level barriers will be particularly relevant for countries considering national screening implementation.

For clinicians, particularly those in primary care, findings will offer practical guidance on effective referral pathways and patient engagement strategies. Given that GPs play a key role in recruitment, this review will help identify clinically feasible, scalable approaches to improve uptake. Strategies such as personalised risk communication, automated eligibility identification, and GP-endorsed invitations may emerge as high-impact interventions