Primary care practice-based interventions and their effect on participation in population-based cancer screening programs: a systematic narrative review

Aim: To provide a systematic synthesis of primary care practice-based interventions and their effect on participation in population-based cancer screening programs. Background: Globally, population-based cancer screening programs (bowel, breast, and cervical) have sub-optimal participation rates. Primary healthcare workers (PHCWs) have an important role in facilitating a patient’s decision to screen; however, barriers exist to their engagement. It remains unclear how to best optimize the role of PHCWs to increase screening participation. Methods: A comprehensive search was conducted from January 2010 until November 2023 in the following databases: Medline (OVID), EMBASE, and CINAHL. Data extraction, quality assessment, and synthesis were conducted. Studies were separated by whether they assessed the effect of a single-component or multi-component intervention and study type. Findings: Forty-nine studies were identified, of which 36 originated from the USA. Fifteen studies were investigations of single-component interventions, and 34 studies were of multi-component interventions. Interventions with a positive effect on screening participation were predominantly multi-component, and most included combinations of audit and feedback, provider reminders, practice-facilitated assessment and improvement, and patient education across all screening programs. Regarding bowel screening, provision of screening kits at point-of-care was an effective strategy to increase participation. Taking a ‘whole-of-practice approach’ and identifying a ‘practice champion’ were found to be contextual factors of effective interventions. The findings suggest that complex interventions comprised of practitioner-focused and patient-focused components are required to increase cancer screening participation in primary care settings. This study provides novel understanding as to what components and contextual factors should be included in primary care practice-based interventions.


Introduction
The success of population-based cancer screening programs in reducing cancer mortality is often limited by sub-optimal participation in the community.Fifty percent of the eligible population participate in Germany's Mammography Screening Program (Hand, 2020), while less than half (44%) participate in Australia's National Bowel Cancer Screening Program (Australian Instititute of Health and Welfare, 2020).Scotland and Canada have participation rates between 60% and 81% for their breast and cervical screening programs.However, lower participation rates are observed in these countries' bowel screening programs (McCowan et al., 2019;Ontario Health, 2021).Many patient-level barriers have been described, including low awareness of program existence (Ferdous et al., 2018;Suwankhong and Liamputtong, 2018), worry about the procedure or outcome (Ferdous et al., 2018;Suwankhong and Liamputtong, 2018;Muthukrishnan et al., 2019), and time and transport required to attend screening (Suwankhong and Liamputtong, 2018;Muthukrishnan et al., 2019).People can be encouraged to engage with screening programs through invitations (Radde et al., 2016), small and mass media campaigns (Durkin et al., 2019;Schliemann et al., 2019), and through other prompts, such as celebrity cancer diagnosis, which have been shown to result in increased screening appointments and call to helplines (Boudioni et al., 1998;Chapman et al., 2005).Importantly, endorsement of screening by a primary healthcare worker (PHCW) is an important facilitator in a patient's decision to screen (Duffy et al., 2017).However, PHCWs often experience challenges in engaging in screening programs, with barriers including the financial structure of primary

Methods
The review is structured in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (Page et al., 2021).The review protocol was registered with the International Prospective Register for Ongoing Systematic Reviews; CRD42020201118.All components of the protocol were adhered to; however, for clarity purposes, the outcome of interest was limited to screening uptake.Search terms were developed with the assistance of a research librarian (Appendix 1).Medline (OVID), EMBASE, and CINAHL were searched for articles published from 1st January 2010 to 23rd November 2023.Reference lists of included articles were searched for additional studies.

Inclusion and exclusion criteria
Articles were included if they met the following criteria: (1) Randomized controlled trials (RCTs), non-randomized trials; (2) Intervention conducted in a high-income country (The World Bank, 2022); (3) Focused on the effect of a primary care practicebased intervention(s) to optimize the role of PHCWs in population-based cancer screening program (bowel, breast, and cervical); (4) Measures screening participation as an outcome; and (5) Published in English.Articles were excluded if they met any of the following criteria: (1) Reviews, protocols, and conference abstracts; (2) Impact of the intervention measured outside of a primary care setting; (3) Not focused on primary care practicebased intervention(s) to optimize the role of PHCWs in population-based cancer screening programs (bowel, breast, and cervical); (4) Intervention not targeted at a PHCW; and (5) Screening participation not included as an outcome.
Search results were imported into Covidence (Babineau, 2014) and duplicates were excluded.Two authors (EV, NC) independently screened titles and abstracts for relevance and three authors (EV, NC, CN) independently assessed the eligibility of full-text articles.The authors discussed disagreements to reach a consensus.

Data extraction and synthesis
A data extraction tool was developed for the review and piloted independently by two authors (EV, CN) who met several times to reach a consensus on the final information to be extracted.Three authors (EV, GN, CN) independently extracted: authors; year of publication; country; screening program; study type; population; sample size; intervention components; comparison; follow-up time; effect on screening participation; and contextual factors of intervention.Contextual factors that were explicitly outlined in studies surrounding primary care practice-based interventions were extracted.
Intervention components were categorized as practice-focused or patient-focused, with each category divided into sub-categories (Table 1).Because of the heterogeneity of interventions and study designs, we did not undertake meta-analysis, instead using narrative synthesis to summarize data (Popay et al., 2006).Studies were categorized by whether they assessed the effect of a single-component or multi-component intervention and study type.

Quality assessment
To assess the quality of randomized and non-randomized studies, the Downs and Black Checklist was used (Downs and Black, 1998).The checklist has 27 questions within five categories: reporting [10 items], external validity [3 items], internal validity/bias [7 items], internal validity/confounding [6 items], and power [1 item].Item 27 (power) was modified.Studies were rated on whether or not they performed an a priori power calculation, with the maximum score for this item being 1 rather than 5.The highest possible score for the checklist was therefore 28 instead of 32.Two authors (EV, NC) independently assessed each article, meeting to discuss and resolve discrepancies.Articles were rated as being of excellent (26-28); good (20-25); fair (15-19); or poor (≤14) quality (Hooper et al., 2008).

Study selection
We identified 1564 studies, with 133 duplicate studies removed before screening.Following title and abstract screening, we excluded 1335 studies, leaving 96 studies for full-text screening.Following full-text screening, we excluded 47 studies (Fig. 1), with 49 studies therefore included in our systematic synthesis.No additional studies were found when searching the reference lists of included studies.

Provider reminders
Two cluster RCTs assessed whether provider reminders could increase bowel screening participation.Guiriguet and Castells (2016) evaluated the effect of an alert in electronic medical records for first-time screeners, finding no effect on screening rate when providing GPs with a list of patients who were non-adherent to screening (patient-specific reminders) compared to usual care had a positive effect on screening rates.A positive effect on screening rates was also found when comparing patient-specific reminders to generic reminders; however, no effect was found when comparing generic reminders to usual care.

Audit and feedback
Two studies assessed the effect of audit and feedback on screening.Hwang et al. (2019) found an audit and feedback tool to have a positive effect on bowel and cervical screening rates compared to baseline.Vaisson et al. (2019) examined the effect of emails to providers that used behavior change techniques (either problem-solving, anticipated regret, or material incentive) to promote access to an existing audit and feedback tool.Among practitioners receiving problem-solving emails, there was a positive effect on cervical screening, but there was no effect on bowel or breast screening participation.Emails operationalizing anticipated regret and material incentives had no effect.

Education and training, provider reminders or audit and feedback
Three studies assessed the effect of education and training, provider reminders, or audit and feedback.Curry et al. (2011) examined an education and training intervention, finding no effect on bowel screening rates.Hsiang et al. (2019) assessed the effect of provider reminders, finding that although there was a large increase in clinician ordering of tests, there was no effect on screening rates for bowel and breast.Jonah et al. (2017) assessed an audit and feedback tool, finding a positive effect on screening rates among people due and overdue for bowel and cervical screening, and among patients overdue for breast screening.Among women due for breast screening, there was no change in screening rates.

Practitioner-focused intervention components
Three studies assessed the effect of interventions with multiple practitioner-focused components on screening rates.Price-Haywood et al. (2014) looked at whether training family physicians on how to engage in cancer risk communication, in addition to audit and feedback, would increase screening.When compared to audit-only, there was a positive effect on breast screening, but no effect on bowel and cervical screening rates.Shaw et al. (2013) found no effect on bowel screening rates when comparing practices participating in practice-facilitated assessment and improvement       Ornstein et al. (2010) found a positive effect on bowel screening rates when assessing the impact of audit and feedback and education and training compared to control practices.

Provider reminders and patient promotional material
Three studies evaluated the effect of interventions involving provider reminders and patient promotional material in the waiting room compared to usual care.In Aragones et al. (2010), patients watched a bowel screening video and received a brochure summarizing the video's messages.They were also given a reminder to hand it to their family physician.A positive effect on bowel screening rates was found.Similarly, Moen et al. (2020) found a positive effect on cervical screening rates following an educational session for GPs, provider reminders, and promotional material in the form of a poster.Walsh et al. (2020) assessed the effect of promotional material, in the form of a virtual 'Doctor' and provider reminders.However, unlike Aragones et al. (2010) and Moen et al. (2020), no effects on bowel, breast, or cervical screening rates were found.

Audit and feedback, patient reminders, and patient navigation
Two studies assessed the effect of audit and feedback, patient reminders, and patient navigation.Atlas et al. (2011) evaluated the use of an informatics system, which connected overdue patients to appropriate providers and presented providers with a list of their overdue patients.Patients received reminders and navigation.A positive effect on breast screening rates was found when compared to usual care.More recently Atlas et al. (2014) studied the impact of involving family physicians in a web-based IT application, with physicians in the intervention arm receiving a list of their overdue patients and providing individualized outreachvia a letter, practice delegate, or practice navigator.However, in comparison to the control, where overdue patients were automatically sent reminder letters and transferred to patient delegate lists, no effect on bowel, breast, or cervical screening rates was found.
Practitioner-focused and patient-focused intervention components Two studies assessed the effect of interventions comprising different combinations of practitioner-focused and patientfocused components, with both studies including patient education and provision of a bowel cancer screening kit at point-of-care.Dodd et al. (2019) assessed the effect of providing GPs with a script to assist them in endorsing screening and providing patients with a screening kit and education.Similarly, in Sun et al. (2018), family physicians received education and patients received education, a screening kit, and a letter of recommendation from their physician.Both studies reported a positive effect on bowel screening rates.
Difference between practitioner-focused and/or patient-focused intervention components Two studies assessed whether there was a difference between intervention groups in screening rates.Basch et al. (2015) compared screening rates between -1.patients receiving educational material, 2. family physicians receiving education, and 3. family physicians receiving education and patients also receiving education.Cameron et al. (2020) compared bowel screening rates between -1.Physicians receiving education and audit and feedback and patients watching an educational video, 2. Physicians receiving education and audit and feedback, and 3. Usual care.No effect on bowel screening rates was found for either study.
The effect of practice-facilitated assessment and improvement, in the form of Quality Improvement (QI) projects and Plan-Do-Study-Act (PDSA) cycles was assessed by sixteen studies.QI projects apply a systematic approach to design, test, and implement interventions (Jones et al., 2019).PDSA cycles are commonly used in QI projects and provide a structured approach to testing interventions and making appropriate adjustments to increase the likelihood of the intervention delivering its' desired effect.

QI projects
Ten practice-facilitated assessment and improvement studies looked at the effect of a QI project on screening.Ruggeri et al. (2020) assessed the effect of provider education, patient education, provision of screening kit, and patient reminders.Like Ruggeri et al. (2020), Hills et al. (2015) reported a positive effect on cervical screening rates following the implementation of a clinical decision support system, provider education, and patient reminders.Walker-Smith and Baldwin (2020) found a positive effect on breast screening following training practice staff to implement screening tools.Frissora et al. (2021) looked at the impact of educating and training providers on screening modalities, educating patients, and audit and feedback, Desai et al. (2021) assessed the effect of provider education and patient education, and Hussain et al. (2021) assessed the impact of provider training, promotional material in the form of a poster, and patient reminder letters.Weiner et al. (2017) looked at the effect of a practice facilitator supporting the implementation of office systems, such as screening reminders.Participating practices also received a financial incentive and screening kits to disseminate.All studies had a positive effect on bowel screening participation.Mader et al.
(2016) evaluated whether education followed by practice-facilitated assessment and improvement, whereby practices worked with QI professionals to conduct activities such as reminder systems streamlining, would increase screening rates.When compared to baseline, a positive effect on bowel screening and breast screening was found, but no effect was found on cervical screening rates.Jones et al. (2022) evaluated the effect of audit and feedback and patient reminders on screening rates, finding a positive effect for bowel and cervical; however, no effect for breast.Contrary to the previous practice-facilitated and assessment studies, Nguyen et al. (2020) found no effect on bowel, breast, and cervical screening rates when assessing the effect of provider education and financial incentives.

Large-scale QI projects
Three studies provided an overview of large-scale QI projects.The QI program utilized PDSA methodology and included components such as provider education and training, and patient education.On a practitioner level, Harris et al. (2015) found no effect on bowel screening rates.However, a supplementary population-level study by Green et al. (2017) found a positive effect on bowel screening and cervical screening.Marx et al. (2016) assessed the effect of a continuous QI project to increase bowel screening rates at five clinics, with components including audit and feedback, provider education and training, provider financial incentives, patient education, and patient reminders.A continuous positive effect was found in three of the five clinics.

PDSA cycles
Three studies assessed the effect of PDSA cycles, involving practitioner-focused and patient-focused intervention components.Dorrington et al. (2015) used rapid PDSA cycles to implement interventions including education to wider practiceteam members and promotional material.A positive effect on cervical screening rates was found.Hountz et al. (2017) implemented reminders to nurses, promotional material, and simplified FOBT ordering processes, and Bakhai et al. (2018) outlined the effect of eight PDSA cycles, with cycles including interventions such as provider reminders, patient navigation, and patient education.Both studies had a positive effect on bowel screening rates.
Practitioner-focused intervention components Baxter et al. (2017) evaluated the impact of strategies on bowel screening uptake, finding practices that employed 4-5 strategies, such as provider reminders and audits and feedback, had a positive effect on bowel screening rates when compared to practices that employed 0-1 strategies.

Practitioner-focused and patient-focused intervention components
Five studies involved interventions comprising different combinations of practitioner-focused and patient-focused components.Potter et al. (2011) and Funes et al. (2021) adapted the FLU-FOBT program.Nurses received education and training on the program, and patients were provided with a screening kit, received education on bowel screening, or a reminder to screen.Wu et al. (2016) assessed the effect of an intervention whereby doctors reviewed rosters of patients due for bowel screening and chose practice delegate outreach or default reminder letter.Patients who were referred to the delegate received education about bowel screening and if they declined to undergo colonoscopy, they were facilitated with ordering a FOBT screening kit.In Willemse et al. (2022), practice staff received education on the importance of bowel screening and the available options for screening.Patients also received education, along with provider reminders and a screening kit.Kaczorowski et al. (2013) assessed the effect on breast and cervical screening rates when combining P4P incentives, provider reminders, and patient reminders.A positive effect was found for all five interventions.

Contextual factors of interventions
Multi-component studies outlined circumstances under which interventions were more likely to optimize the role of PHCWs and increase participation in screening programs through the use of two strategies: engaging whole teams in a practice setting and the use of champions (Tables 2 and 3).The 'whole-of-practice approach' contextual factor is defined as empowering an array of practice staff (eg, administrative staff, nurses, managers, and family physicians) to be involved in cancer screening interventions (Ornstein et al., 2010;Atlas et al., 2011;Shaw et al., 2013;Atlas et al., 2014;Harris et al., 2015;Hills et al., 2015;Mader et al., 2016;Wu et al., 2016;Hountz et al., 2017;Weiner et al., 2017;Bakhai et al., 2018;Nguyen et al., 2020;Hussain et al., 2021;Jones et al., 2022).Ornstein et al. (2010) concluded that practices that 'meet as a team to plan evidence-based (quality) improvement strategies : : : can achieve much higher levels of screening than typically reported.'Additionally, a multi-site study by Mader et al. (2016) discussed how practices with the greatest change in cancer screening rates 'had fully engaged staff at several levels within the practice.'Another contextual factor commonly reported was having a 'practice champion' to drive activities within screening programs (Shaw et al., 2013;Hills et al., 2015;Mader et al., 2016;Wu et al., 2016;Weiner et al., 2017;Bakhai et al., 2018;Desai et al., 2021;Jones et al., 2022).Hills et al. (2015) reported nearly doubling cervical screening rates (38.1% to 69.7%) following the selection of a practice nurse to facilitate and provide clear direction for a QI project.Similarly, Bakhai et al. (2018) outlined that 'engaged (practice) leadership' implementing a QI project exceeded their aim of increasing bowel screening rates from a baseline of 50% to 70%, reaching 75%.

Discussion
This systematic review of 49 studies targeting primary care practices has several key findings as an avenue to increase participation in cancer screening programs.Firstly, interventions with a positive effect were predominantly multi-component, and most included combinations of strategies such as audit and feedback, provider reminders, practice-facilitated assessment and improvement, and patient education across all screening programs.Regarding bowel screening, the provision of screening kits at point-of-care was an effective strategy to increase participation.Secondly, evidence to support the effectiveness of financial incentives for providers was limited, with the review finding most studies to have no effect on screening rates.Finally, 'whole-ofpractice approaches' and identifying 'practice champions' were found to be contextual factors of effective interventions.This study provides novel understanding of components and contextual factors that should be included in interventions using PHCWs to facilitate greater participation in screening programs.
The findings suggest that complex interventions comprised of practitioner-focused and patient-focused components are required to increase cancer screening participation in primary care settings.Supporting the findings of previous research (Emery et al., 2014;Chauhan et al., 2017), audit and feedback were the only intervention components that had a positive effect on cancer screening rates across single-component and multi-component studies.By drawing attention to gaps in performance, audit and feedback likely act as a motivator for PHCWs to change the way they engage with screening programs (Thomson O'Brien et al., 2000).Of the 17 multi-component interventions assessing the effect of practice-facilitated assessments and improvements, fourteen studies reported a positive effect on screening participation rates.Practice-facilitated assessment and improvement may be effective due to being tailored to the needs of the practice, and therefore being more acceptable and appropriate.Existing research suggests provider reminders can increase PHCW's engagement in cancer screening programs (Emery et al., 2012).Whilst we found Primary Health Care Research & Development provider reminders were often a component of multi-component interventions with a positive effect on screening rates, provider reminders as a single-component intervention did not always have a positive effect on screening rates.
Patient-focused components of interventions were identified as improving the ability of PHCWs to facilitate participation in cancer screening.The benefit of this approach is that it targets different barriers to change (Grimshaw et al., 2001), possibly explaining why provider reminders as a single-component intervention were less effective at increasing screening participation than when incorporated within a multi-component intervention.This is exemplified by Hsiang et al. (2019) who found that although provider reminders resulted in a large increase in clinician ordering of tests, there was no effect on bowel and breast screening rates.
Regarding bowel screening, patient education and the provision of a screening kit at point-of-care were common components of effective multi-component interventions.Educating patients before their appointment likely mitigates patient-level barriers to screening, including worry about the procedure or outcome (Ferdous et al., 2018;Suwankhong and Liamputtong, 2018;Muthukrishnan et al., 2019).Additionally, a practitioner providing a bowel cancer screening kit at point-of-care may mitigate structural barriers, such as time (Suwankhong and Liamputtong, 2018;Muthukrishnan et al., 2019), and demonstrates direct endorsement by a PHCW, an important facilitator in a patient's decision to screen (Duffy et al., 2017).The importance of mitigating structural barriers to screening is highlighted by Potter et al. (2011) and Funes et al. (2021) who reported a positive effect on bowel screening rates for patients participating in the FLU-FOBT program when compared to non-recipients.A follow-up study by Potter et al. (2011) found components of the program and screening rates were maintained one year later (Walsh et al., 2012).Further supporting the importance of making it easier for patients to screen, mailing self-sampling cervical screening kits is more effective at reaching under-screened patients than sending an invitation or reminder letters for clinician sampling (Arbyn et al., 2018;Yeh et al., 2019).
Practitioner-focused financial incentives did not optimize the role of PHCWs, with single-component non-randomized trials reporting no effect on screening participation.However, Jung et al. (2017) did find that financial incentives for adopting and using EHRs had a positive effect on screening rates.The difference in these results may be due to financial incentives for engaging patients failing to consider the demanding setting in which PHCWs work, where they are faced with a plethora of competing health issues (Wender, 1993).EHRs facilitate the delivery of care for PHCWs by improving the organization and accessibility of clinical information.
Lastly, our review identified circumstances under which interventions were more likely to optimize the role of PHCWs and increase participation in screening programs.A whole-ofpractice approach is likely an effective contextual factor due to spreading practice workload, with 'time' a known barrier to PHCWs engaging in cancer screening (Wender, 1993;Yarnall et al., 2003;Verbunt et al., 2022).A whole-of-practice approach is supported by a previous review (Chauhan et al., 2017), which outlined the importance of collaborative team-based interventions to effectively modify PHCW practice and patient outcomes.Having a practice champion to drive cancer screening interventions may also be an important contextual factor due to their role in promoting positive practice culture toward cancer screening programs (Verbunt et al., 2022).

Limitations
Interventions and study designs were heterogeneous, precluding meta-analysis.It was not possible to present data in forest plots due to the methodological heterogeneity, as well as some studies not reporting P-values or confidence intervals.However, our broad inclusion criteria for primary care practice-based interventions and study type (RCTs, non-randomized trials) ensured we captured the range of studies relevant to the topic, which in turn supports the relevance of findings for policy and planning.It is possible that contextual factors relevant to the outcomes were not all described in the original study reports, and have therefore not been acknowledged in this review.Our review only reports on contextual factors that were explicitly outlined in studies surrounding primary care practice-based interventions, with the potential that more implicit contextual factors were missed.Studies were predominantly from the US, and the majority were focused on bowel cancer screening, with findings potentially not generalizable to countries where the structure of the screening program differs.

Conclusions
Multi-component interventions that are tailored to the needs of a primary care setting, and the patients they serve, may improve the ability of PHCWs to facilitate greater participation in populationbased cancer screening programs.Future research should explore the effect of combining identified components of effective interventions (audit and feedback, provider reminders, practicefacilitated assessment and improvement, and patient education across all screening programs and the provision of screening kits at point-of-care for bowel screening) with contextual factors (wholeof-practice approach, practice champion) to maximize screening participation Supplementary material.To view supplementary material for this article, please visit https://doi.org/10.1017/S1463423623000713

.
Training and/or education Three cluster RCTs assessed whether training and/or education for PHCWs increased bowel cancer screening participation.Aubin-Auger et al. (2016) assessed the effect of training focused on improving GPs' communication skills, Dignan et al. (2014) investigated whether more comprehensive PHCW-focused training and education on the following topicsscreening efficacy, clinical performance measures, patient counseling, and creating a screening-friendly environment increased participation, and Wang et al. (2018) tested whether training family physicians on patient-centered communication had an impact.None of these trials increased screening participation compared to usual care.

Figure 1 .
Figure 1.Prisma diagram of study selection

Table 1 .
Practice-focused and patient-focused intervention components

Table 2 .
Overview of single-component primary care practice-based interventions and effect on cancer screening participation

Table 2 .
(Continued ) *No effect refers to no statistically significant effect of the intervention.

Table 3 .
Overview of multi-component primary care practice-based interventions and effect on cancer screening participation

Table 3 .
(Continued ) *No effect refers to no statistically significant effect of the intervention.