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, Reference Hand2020), 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., Reference McCowan, McSkimming, Papworth, Kotzur, McConnachie, MacDonald, Wyke, Crighton, Campbell, Weller, Steele and Robb2019; Ontario Health, 2021). Many patient-level barriers have been described, including low awareness of program existence (Ferdous et al., Reference Ferdous, Lee, Goopy, Yang, Rumana, Abedin and Turin2018; Suwankhong and Liamputtong, Reference Suwankhong and Liamputtong2018), worry about the procedure or outcome (Ferdous et al., Reference Ferdous, Lee, Goopy, Yang, Rumana, Abedin and Turin2018; Suwankhong and Liamputtong, Reference Suwankhong and Liamputtong2018; Muthukrishnan et al., Reference Muthukrishnan, Arnold and James2019), and time and transport required to attend screening (Suwankhong and Liamputtong, Reference Suwankhong and Liamputtong2018; Muthukrishnan et al., Reference Muthukrishnan, Arnold and James2019). People can be encouraged to engage with screening programs through invitations (Radde et al., Reference Radde, Gottschalk, Bussas, Schulein, Schriefer, Seifert, Neumann, Kaiser, Blettner and Klug2016), small and mass media campaigns (Durkin et al., Reference Durkin, Broun, Spittal and Wakefield2019; Schliemann et al., Reference Schliemann, Su, Paramasivam, Treanor, Dahlui, Loh and Donnelly2019), 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., Reference Boudioni, Mossman, Jones, Leydon and McPherson1998; Chapman et al., Reference Chapman, McLeod, Wakefield and Holding2005). Importantly, endorsement of screening by a primary healthcare worker (PHCW) is an important facilitator in a patient’s decision to screen (Duffy et al., Reference Duffy, Myles, Maroni and Mohammad2017). However, PHCWs often experience challenges in engaging in screening programs, with barriers including the financial structure of primary care, the structure of screening programs, time, and screening knowledge (Wender, Reference Wender1993; Yarnall et al., Reference Yarnall, Pollak, Ostbye, Krause and Michener2003; Verbunt et al., Reference Verbunt, Boyd, Creagh, Milley, Emery, Nightingale and Kelaher2022).
The role of PHCWs, including family physicians/general practitioners (GPs)Footnote 1 and practice nurses, in cancer screening programs differs between countries and programs. Accredited PHCWs are responsible for facilitating screening tests in most cervical screening programs (Fontham et al., Reference Fontham, Wolf, Church, Etzioni, Flowers, Herzig, Guerra, Oeffinger, Shih, Walter, Kim, Andrews, Desantis, Fedewa, Manassaram-Baptiste, Saslow, Wender and Smith2020; Australian Government Department of Health, 2021b). As mammography screening mostly occurs outside of primary care, PHCWs play a less direct but important role in promoting participation and facilitating follow-up care (Klarenbach et al., Reference Klarenbach, Sims-Jones, Lewin, Singh, Theriault, Tonelli, Doull, Courage, Garcia and Thombs2018; The Royal Australian College of General Practitioners, 2018). For bowel cancer screening, the role of PHCWs differs according to the program structure. In the US and Canada, family physicians are responsible for recommending, performing, or referring patients to different screening modalities, including colonoscopy and fecal occult blood test (FOBT) (Canadian Task Force on Preventive Health, 2001; Force et al., Reference Force, Bibbins-Domingo, Grossman, Curry, Davidson, Epling, Garcia, Gillman, Harper, Kemper, Krist, Kurth, Landefeld, Mangione, Owens, Phillips, Phipps, Pignone and Siu2016). However, in countries such as the United Kingdom and Australia, eligible participants receive a fecal immunochemical test or FOBT screening kit via mail and have the option to nominate their GP to receive results and provide follow-up care (National Health Service, 2021; Australian Government Department of Health, 2021a).
Reviews, although not systematic, exist on the effect of interventions addressing practitioner-level barriers to engaging in cancer screening. A 2012 review concluded that the engagement of PHCWs and screening participation could be improved by using audit and feedback systems and office-system prompts, such as reminders for the clinician to discuss or order cancer screening tests (Emery et al., Reference Emery, Trevena, Mazza, Fallon-Ferguson, Shaw, Williams and Varlow2012). A more recent review found interactive and multi-faceted continuous medical education, training with audit and feedback, enablement through IT-based systems, and collaborative team-based interventions can modify PHCW’s practice and improve patient outcomes (Chauhan et al., Reference Chauhan, Jeyaraman, Mann, Lys, Skidmore, Sibley, Abou-Setta and Zarychanski2017). However, this review was not specific to cancer screening, nor did it discuss any magnitude of effect on cancer screening participation. Further, contextual factors, defined as the ‘features of the circumstances in which an intervention is implemented that may interact with the intervention to produce variation in outcomes’, were not outlined (Craig et al., Reference Craig, Di Ruggiero, Frohlich, MyKhalovskiy and White2018). Contextual factors, such as workload and leadership, are an important consideration for understanding how, and under what circumstances interventions create change (Moore et al., Reference Moore, Audrey, Barker, Bond, Bonell, Hardeman, Moore, O’Cathain, Tinati, Wight and Baird2015; Skivington et al., Reference Skivington, Matthews, Simpson, Craig, Baird, Blazeby, Boyd, Craig, French, McIntosh, Petticrew, Rycroft-Malone, White and Moore2021).
Thus, it is unclear how to best optimize the role of PHCWs to increase cancer screening participation. This review aims to provide a systematic synthesis of primary care practice-based interventions and their effect on participation in population-based cancer screening programs. Contextual factors of effective interventions will be summarized. Findings from this review can be used to guide the development of interventions using PHCWs to facilitate greater participation in bowel, breast, and cervical screening programs.
Methods
The review is structured in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement (Page et al., Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann, Mulrow, Shamseer, Tetzlaff, Akl, Brennan, Chou, Glanville, Grimshaw, Hrobjartsson, Lalu, Li, Loder, Mayo-Wilson, McDonald, McGuinness, Stewart, Thomas, Tricco, Welch, Whiting and Moher2021). 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.
Study selection
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 practice-based 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 practice-based 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, Reference Babineau2014) 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., Reference Popay, Roberts, Sowden, Petticrew, Arai, Rodgers, Britten, Roen and Duffy2006). Studies were categorized by whether they assessed the effect of a single-component or multi-component intervention and study type.
Table 1. Practice-focused and patient-focused intervention components
Quality assessment
To assess the quality of randomized and non-randomized studies, the Downs and Black Checklist was used (Downs and Black, Reference Downs and Black1998). 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., Reference Hooper, Jutai, Strong and Russell-Minda2008).
Results
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.
Figure 1. Prisma diagram of study selection
Critical appraisal
Overall, RCTs had a mean score of 22 (good) and non-randomized trials 17 (fair). RCTs had a higher mean score than non-randomized trials in internal validity/confounding (4 versus 2), with similar mean scores in the remaining methodological categories – reporting (9 versus 8), external validity (3 versus 3), and internal validity/bias (4 versus 4) (Appendix 2).
Study characteristics
Characteristics of studies
Studies originated from the US (n = 36), Canada (n = 5), Australia (n = 3), France (n = 2), Norway (n = 1), Netherlands (n = 1), and Spain (n = 1). Study designs included RCTs (n = 19) and non-randomized trials (n = 28). Interventions focused on increasing participation in bowel screening (n = 27), six targeted cervical screening, two breast screening, and 14 studies focused on a combination of all three screening programs. Fifteen studies were investigations of single-component interventions, including seven RCTs, and eight non-randomized trials. Thirty-four studies included multi-component interventions, of which 12 were RCTs, and 22 were non-randomized trials (Fig. 2).
Figure 2. Characteristics of studies.
Single-component – RCTs
Seven studies from the US (n = 3), France (n = 2), Spain (n = 1), and Canada (n = 1) were single-component RCTs (Dignan et al., Reference Dignan, Shelton, Slone, Tolle, Mohammad, Schoenberg, Pearce, Van Meter and Ely2014; Aubin-Auger et al., Reference Aubin-Auger, Laouenan, Le Bel, Mercier, Baruch, Lebeau, Youssefian, Le Trung, Peremans and Van Royen2016; Guiriguet and Castells, Reference Guiriguet and Castells2016; Rat et al., Reference Rat, Pogu, Le Donne, Latour, Bianco, Nanin, Cowppli-Bony, Gaultier and Nguyen2017; Wang et al., Reference Wang, Ma, Liang, Tan, Makambi, Dong, Vernon, Tu and Mandelblatt2018; Hwang et al., Reference Hwang, Harding, Chang, O’Keefe, Horn and Clark2019; Vaisson et al., Reference Vaisson, Witteman, Chipenda-Dansokho, Saragosa, Bouck, Bravo, Desveaux, Llovet, Presseau, Taljaard, Umar, Grimshaw, Tinmouth and Ivers2019) (Table 2).
Table 2. Overview of single-component primary care practice-based interventions and effect on cancer screening participation
* No effect refers to no statistically significant effect of the intervention.
Training and/or education
Three cluster RCTs assessed whether training and/or education for PHCWs increased bowel cancer screening participation. Aubin-Auger et al. (Reference Aubin-Auger, Laouenan, Le Bel, Mercier, Baruch, Lebeau, Youssefian, Le Trung, Peremans and Van Royen2016) assessed the effect of training focused on improving GPs’ communication skills, Dignan et al. (Reference Dignan, Shelton, Slone, Tolle, Mohammad, Schoenberg, Pearce, Van Meter and Ely2014) investigated whether more comprehensive PHCW-focused training and education on the following topics – screening efficacy, clinical performance measures, patient counseling, and creating a screening-friendly environment increased participation, and Wang et al. (Reference Wang, Ma, Liang, Tan, Makambi, Dong, Vernon, Tu and Mandelblatt2018) tested whether training family physicians on patient-centered communication had an impact. None of these trials increased screening participation compared to usual care.
Provider reminders
Two cluster RCTs assessed whether provider reminders could increase bowel screening participation. Guiriguet and Castells (Reference Guiriguet and Castells2016) evaluated the effect of an alert in electronic medical records for first-time screeners, finding no effect on screening rate when compared to the control group. Conversely, Rat et al. (Reference Rat, Pogu, Le Donne, Latour, Bianco, Nanin, Cowppli-Bony, Gaultier and Nguyen2017) found 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. (Reference Hwang, Harding, Chang, O’Keefe, Horn and Clark2019) found an audit and feedback tool to have a positive effect on bowel and cervical screening rates compared to baseline. Vaisson et al. (Reference Vaisson, Witteman, Chipenda-Dansokho, Saragosa, Bouck, Bravo, Desveaux, Llovet, Presseau, Taljaard, Umar, Grimshaw, Tinmouth and Ivers2019) 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.
Single-component – non-randomized trials
Eight studies from the US (n = 5), Australia (n = 1), Netherlands (n = 1), and Canada (n = 1) were single-component non-randomized trials (Gavagan et al., Reference Gavagan, Du, Saver, Adams, Graham, McCray and Goodrick2010; Curry et al., Reference Curry, Lengerich, Kluhsman, Graybill, Liao, Schaefer, Spleen and Dignan2011; Greene, Reference Greene2013; Kirschner et al., Reference Kirschner, Braspenning, Akkermans, Jacobs and Grol2013; Kiran et al., Reference Kiran, Wilton, Moineddin, Paszat and Glazier2014; Jonah et al., Reference Jonah, Pefoyo, Lee, Hader, Strasberg, Kupets, Chiarelli and Tinmouth2017; Jung et al., Reference Jung, Unruh, Vest, Casalino, Kern, Grinspan, Bao, Kaushal and Investigators2017; Hsiang et al., Reference Hsiang, Mehta, Small, Rareshide, Snider, Day and Patel2019) (Table 2).
Financial incentives
Five studies assessed whether financial incentives to providers would increase screening participation of which three focused on pay-for-performance (P4P) programs. Greene (Reference Greene2013) examined the effect of paying GPs to conduct cervical screening for overdue women, Kiran et al. (Reference Kiran, Wilton, Moineddin, Paszat and Glazier2014) evaluated a P4P program with high uptake among family physicians, which offered substantial financial incentives to increase bowel, breast, and cervical screening, and Kirschner et al. (Reference Kirschner, Braspenning, Akkermans, Jacobs and Grol2013) assessed whether general practice receiving a bonus based on a score of patient care increased screening. None of these P4P programs had an impact on screening rates compared to baseline. Gavagan et al. (Reference Gavagan, Du, Saver, Adams, Graham, McCray and Goodrick2010) found that doubling financial incentives related to achieving group targets on breast or cervical screening rates had no effect. In Jung et al. (Reference Jung, Unruh, Vest, Casalino, Kern, Grinspan, Bao, Kaushal and Investigators2017), family physicians participated in an initiative where payments were received for adopting and using electronic health records (EHRs). Unlike the previous studies, a positive effect on bowel screening rates was found.
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. (Reference Curry, Lengerich, Kluhsman, Graybill, Liao, Schaefer, Spleen and Dignan2011) examined an education and training intervention, finding no effect on bowel screening rates. Hsiang et al. (Reference Hsiang, Mehta, Small, Rareshide, Snider, Day and Patel2019) 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. (Reference Jonah, Pefoyo, Lee, Hader, Strasberg, Kupets, Chiarelli and Tinmouth2017) 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.
Multi-component – RCTs
Twelve studies from the US (n = 10), Australia (n = 1), and Norway (n = 1) were multi-component RCTs (Aragones et al., Reference Aragones, Schwartz, Shah and Gany2010; Ornstein et al., Reference Ornstein, Nemeth, Jenkins and Nietert2010; Atlas et al., Reference Atlas, Grant, Lester, Ashburner, Chang, Barry and Chueh2011; Shaw et al., Reference Shaw, Ohman-Strickland, Piasecki, Hudson, Ferrante, McDaniel, Nutting and Crabtree2013; Atlas et al., Reference Atlas, Zai, Ashburner, Chang, Percac-Lima, Levy, Chueh and Grant2014; Price-Haywood et al., Reference Price-Haywood, Harden-Barrios and Cooper2014; Basch et al., Reference Basch, Zybert, Wolf, Basch, Ullman, Shmukler, King, Neugut and Shea2015; Sun et al., Reference Sun, Tsoh, Tong, Cheng, Chow, Stewart and Nguyen2018; Dodd et al., Reference Dodd, Carey, Mansfield, Oldmeadow and Evans2019; Cameron et al., Reference Cameron, Ramirez-Zohfeld, Ferreira, Dolan, Radosta, Galanter, Eder, Wolf and Rademaker2020; Moen et al., Reference Moen, Kumar, Igland and Diaz2020; Walsh et al., Reference Walsh, Potter, Salazar, Ozer, Gildengorin, Dass and Green2020) (Table 3).
Table 3. Overview of multi-component primary care practice-based interventions and effect on cancer screening participation
* No effect refers to no statistically significant effect of the intervention.
Practitioner-focused intervention components
Three studies assessed the effect of interventions with multiple practitioner-focused components on screening rates. Price-Haywood et al. (Reference Price-Haywood, Harden-Barrios and Cooper2014) 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. (Reference Shaw, Ohman-Strickland, Piasecki, Hudson, Ferrante, McDaniel, Nutting and Crabtree2013) found no effect on bowel screening rates when comparing practices participating in practice-facilitated assessment and improvement and education and training to control practices. However, Ornstein et al. (Reference Ornstein, Nemeth, Jenkins and Nietert2010) 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. (Reference Aragones, Schwartz, Shah and Gany2010), 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. (Reference Moen, Kumar, Igland and Diaz2020) 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. (Reference Walsh, Potter, Salazar, Ozer, Gildengorin, Dass and Green2020) assessed the effect of promotional material, in the form of a virtual ‘Doctor’ and provider reminders. However, unlike Aragones et al. (Reference Aragones, Schwartz, Shah and Gany2010) and Moen et al. (Reference Moen, Kumar, Igland and Diaz2020), 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. (Reference Atlas, Grant, Lester, Ashburner, Chang, Barry and Chueh2011) 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. (Reference Atlas, Zai, Ashburner, Chang, Percac-Lima, Levy, Chueh and Grant2014) 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 outreach – via 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 patient-focused components, with both studies including patient education and provision of a bowel cancer screening kit at point-of-care. Dodd et al. (Reference Dodd, Carey, Mansfield, Oldmeadow and Evans2019) 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. (Reference Sun, Tsoh, Tong, Cheng, Chow, Stewart and Nguyen2018), 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. (Reference Basch, Zybert, Wolf, Basch, Ullman, Shmukler, King, Neugut and Shea2015) 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. (Reference Cameron, Ramirez-Zohfeld, Ferreira, Dolan, Radosta, Galanter, Eder, Wolf and Rademaker2020) 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.
Multi-component – non-randomized trials
Twenty-two studies from the US (n = 18), Canada (n = 3) and Australia (n = 1) were multi-component non-randomized trials (Potter et al., Reference Potter, Yu, Gildengorin, Yu, Chan, McPhee, Green and Walsh2011; Kaczorowski et al., Reference Kaczorowski, Hearps, Lohfeld, Goeree, Donald, Burgess and Sebaldt2013; Dorrington et al., Reference Dorrington, Herceg, Douglas, Tongs and Bookallil2015; Harris et al., Reference Harris, Green, Brown, Roberts, Russell, Fournie, Webster-Bogaert, Paquette-Warren, Kotecha, Han, Thind, Stewart, Reichert, Tompkins and Birtwhistle2015; Hills et al., Reference Hills, Kulbok and Clark2015; Mader et al., Reference Mader, Fox, Epling, Noronha, Swanger, Wisniewski, Vitale, Norton and Morley2016; Marx et al., Reference Marx, Tse, Golden and Johnson2016; Wu et al., Reference Wu, Mulder, Zai, Hu, Costa, Tishler, Saltzman, Ellner and Bitton2016; Baxter et al., Reference Baxter, Sutradhar, Li, Daly, Honein-Abouhaidar, Richardson, Del Giudice, Tinmouth, Paszat and Rabeneck2017; Green et al., Reference Green, Harris, Webster-Bogaert, Han, Kotecha, Kopp, Ho, Birtwhistle and Glazier2017; Hountz et al., Reference Hountz, Coddington, Foli and Thorlton2017; Weiner et al., Reference Weiner, Rohweder, Scott, Teal, Slade, Deal, Jihad and Wolf2017; Bakhai et al., Reference Bakhai, Ahluwalia, Nallapeta, Mangat and Reynolds2018; Nguyen et al., Reference Nguyen, Chien, Meyers, Li, Singer and Rosenthal2020; Desai et al., Reference Desai, Mehta and Vega2021; Funes et al., Reference Funes, Flores, Davidow, Friedman, Korenblit and Thomas2021; Frissora et al., Reference Frissora, Biernacki and Walloch2021; Hussain et al., Reference Hussain, Aurand, Onuorah, Krstevski, Markert, Springer and Agrawal2021; Jones et al., Reference Jones, Schott, Leverence and Cancino2022; Ruggeri et al., Reference Ruggeri, Reed, Coyle, Stoltzfus, Fioravanti and Tehrani2020; Walker-Smith and Baldwin, Reference Walker-Smith and Baldwin2020; Willemse et al., Reference Willemse, Jadalla, Conahan, Sarff and Brady2022) (Table 3).
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., Reference Jones, Vaux and Olsson-Brown2019). 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. (Reference Ruggeri, Reed, Coyle, Stoltzfus, Fioravanti and Tehrani2020) assessed the effect of provider education, patient education, provision of screening kit, and patient reminders. Like Ruggeri et al. (Reference Ruggeri, Reed, Coyle, Stoltzfus, Fioravanti and Tehrani2020), Hills et al. (Reference Hills, Kulbok and Clark2015) 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 (Reference Walker-Smith and Baldwin2020) found a positive effect on breast screening following training practice staff to implement screening tools. Frissora et al. (Reference Frissora, Biernacki and Walloch2021) looked at the impact of educating and training providers on screening modalities, educating patients, and audit and feedback, Desai et al. (Reference Desai, Mehta and Vega2021) assessed the effect of provider education and patient education, and Hussain et al. (Reference Hussain, Aurand, Onuorah, Krstevski, Markert, Springer and Agrawal2021) assessed the impact of provider training, promotional material in the form of a poster, and patient reminder letters. Weiner et al. (Reference Weiner, Rohweder, Scott, Teal, Slade, Deal, Jihad and Wolf2017) 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. (Reference Mader, Fox, Epling, Noronha, Swanger, Wisniewski, Vitale, Norton and Morley2016) 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. (Reference Jones, Schott, Leverence and Cancino2022) 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. (Reference Nguyen, Chien, Meyers, Li, Singer and Rosenthal2020) 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. (Reference Harris, Green, Brown, Roberts, Russell, Fournie, Webster-Bogaert, Paquette-Warren, Kotecha, Han, Thind, Stewart, Reichert, Tompkins and Birtwhistle2015) found no effect on bowel screening rates. However, a supplementary population-level study by Green et al. (Reference Green, Harris, Webster-Bogaert, Han, Kotecha, Kopp, Ho, Birtwhistle and Glazier2017) found a positive effect on bowel screening and cervical screening. Marx et al. (Reference Marx, Tse, Golden and Johnson2016) 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. (Reference Dorrington, Herceg, Douglas, Tongs and Bookallil2015) used rapid PDSA cycles to implement interventions including education to wider practice-team members and promotional material. A positive effect on cervical screening rates was found. Hountz et al. (Reference Hountz, Coddington, Foli and Thorlton2017) implemented reminders to nurses, promotional material, and simplified FOBT ordering processes, and Bakhai et al. (Reference Bakhai, Ahluwalia, Nallapeta, Mangat and Reynolds2018) 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. (Reference Baxter, Sutradhar, Li, Daly, Honein-Abouhaidar, Richardson, Del Giudice, Tinmouth, Paszat and Rabeneck2017) 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. (Reference Potter, Yu, Gildengorin, Yu, Chan, McPhee, Green and Walsh2011) and Funes et al. (Reference Funes, Flores, Davidow, Friedman, Korenblit and Thomas2021) 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. (Reference Wu, Mulder, Zai, Hu, Costa, Tishler, Saltzman, Ellner and Bitton2016) 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. (Reference Willemse, Jadalla, Conahan, Sarff and Brady2022), 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. (Reference Kaczorowski, Hearps, Lohfeld, Goeree, Donald, Burgess and Sebaldt2013) 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., Reference Ornstein, Nemeth, Jenkins and Nietert2010; Atlas et al., Reference Atlas, Grant, Lester, Ashburner, Chang, Barry and Chueh2011; Shaw et al., Reference Shaw, Ohman-Strickland, Piasecki, Hudson, Ferrante, McDaniel, Nutting and Crabtree2013; Atlas et al., Reference Atlas, Zai, Ashburner, Chang, Percac-Lima, Levy, Chueh and Grant2014; Harris et al., Reference Harris, Green, Brown, Roberts, Russell, Fournie, Webster-Bogaert, Paquette-Warren, Kotecha, Han, Thind, Stewart, Reichert, Tompkins and Birtwhistle2015; Hills et al., Reference Hills, Kulbok and Clark2015; Mader et al., Reference Mader, Fox, Epling, Noronha, Swanger, Wisniewski, Vitale, Norton and Morley2016; Wu et al., Reference Wu, Mulder, Zai, Hu, Costa, Tishler, Saltzman, Ellner and Bitton2016; Hountz et al., Reference Hountz, Coddington, Foli and Thorlton2017; Weiner et al., Reference Weiner, Rohweder, Scott, Teal, Slade, Deal, Jihad and Wolf2017; Bakhai et al., Reference Bakhai, Ahluwalia, Nallapeta, Mangat and Reynolds2018; Nguyen et al., Reference Nguyen, Chien, Meyers, Li, Singer and Rosenthal2020; Hussain et al., Reference Hussain, Aurand, Onuorah, Krstevski, Markert, Springer and Agrawal2021; Jones et al., Reference Jones, Schott, Leverence and Cancino2022). Ornstein et al. (Reference Ornstein, Nemeth, Jenkins and Nietert2010) 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. (Reference Mader, Fox, Epling, Noronha, Swanger, Wisniewski, Vitale, Norton and Morley2016) 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., Reference Shaw, Ohman-Strickland, Piasecki, Hudson, Ferrante, McDaniel, Nutting and Crabtree2013; Hills et al., Reference Hills, Kulbok and Clark2015; Mader et al., Reference Mader, Fox, Epling, Noronha, Swanger, Wisniewski, Vitale, Norton and Morley2016; Wu et al., Reference Wu, Mulder, Zai, Hu, Costa, Tishler, Saltzman, Ellner and Bitton2016; Weiner et al., Reference Weiner, Rohweder, Scott, Teal, Slade, Deal, Jihad and Wolf2017; Bakhai et al., Reference Bakhai, Ahluwalia, Nallapeta, Mangat and Reynolds2018; Desai et al., Reference Desai, Mehta and Vega2021; Jones et al., Reference Jones, Schott, Leverence and Cancino2022). Hills et al. (Reference Hills, Kulbok and Clark2015) 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. (Reference Bakhai, Ahluwalia, Nallapeta, Mangat and Reynolds2018) 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-of-practice 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., Reference Emery, Shaw, Williams, Mazza, Fallon-Ferguson, Varlow and Trevena2014; Chauhan et al., Reference Chauhan, Jeyaraman, Mann, Lys, Skidmore, Sibley, Abou-Setta and Zarychanski2017), 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., Reference Thomson O’Brien, Oxman, Davis, Haynes, Freemantle and Harvey2000). 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., Reference Emery, Trevena, Mazza, Fallon-Ferguson, Shaw, Williams and Varlow2012). Whilst we found 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., Reference Grimshaw, Shirran, Thomas, Mowatt, Fraser, Bero, Grilli, Harvey, Oxman and O’Brien2001), 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. (Reference Hsiang, Mehta, Small, Rareshide, Snider, Day and Patel2019) 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., Reference Ferdous, Lee, Goopy, Yang, Rumana, Abedin and Turin2018; Suwankhong and Liamputtong, Reference Suwankhong and Liamputtong2018; Muthukrishnan et al., Reference Muthukrishnan, Arnold and James2019). Additionally, a practitioner providing a bowel cancer screening kit at point-of-care may mitigate structural barriers, such as time (Suwankhong and Liamputtong, Reference Suwankhong and Liamputtong2018; Muthukrishnan et al., Reference Muthukrishnan, Arnold and James2019), and demonstrates direct endorsement by a PHCW, an important facilitator in a patient’s decision to screen (Duffy et al., Reference Duffy, Myles, Maroni and Mohammad2017). The importance of mitigating structural barriers to screening is highlighted by Potter et al. (Reference Potter, Yu, Gildengorin, Yu, Chan, McPhee, Green and Walsh2011) and Funes et al. (Reference Funes, Flores, Davidow, Friedman, Korenblit and Thomas2021) 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. (Reference Potter, Yu, Gildengorin, Yu, Chan, McPhee, Green and Walsh2011) found components of the program and screening rates were maintained one year later (Walsh et al., Reference Walsh, Gildengorin, Green, Jenkins and Potter2012). 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., Reference Arbyn, Smith, Temin, Sultana and Castle2018; Yeh et al., Reference Yeh, Kennedy, De Vuyst and Narasimhan2019).
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. (Reference Jung, Unruh, Vest, Casalino, Kern, Grinspan, Bao, Kaushal and Investigators2017) 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, Reference Wender1993). 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-of-practice approach is likely an effective contextual factor due to spreading practice workload, with ‘time’ a known barrier to PHCWs engaging in cancer screening (Wender, Reference Wender1993; Yarnall et al., Reference Yarnall, Pollak, Ostbye, Krause and Michener2003; Verbunt et al., Reference Verbunt, Boyd, Creagh, Milley, Emery, Nightingale and Kelaher2022). A whole-of-practice approach is supported by a previous review (Chauhan et al., Reference Chauhan, Jeyaraman, Mann, Lys, Skidmore, Sibley, Abou-Setta and Zarychanski2017), 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., Reference Verbunt, Boyd, Creagh, Milley, Emery, Nightingale and Kelaher2022).
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 population-based cancer screening programs. Future research should explore the effect of combining identified components of effective interventions (audit and feedback, provider reminders, practice-facilitated 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 (whole-of-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
Acknowledgments
We thank Jim Berryman, Liaison Librarian, Brownless Biomedical Library University of Melbourne for his expert input when developing the search strategy. We dedicate this study to the memory of Professor Margaret Kelaher.
Financial support
This work was supported by the Commonwealth Department of Health, Australia. The funder had no role in the design, execution, analyses, interpretation, and decision to publish.
Competing interests
None.
Ethical standards
Ethics committee approval was not required for this review.
Open access
