Introduction

Screening has been defined as “the presumptive identification of unrecognized disease or defect by the application of tests, examinations or other procedures that can be applied rapidly.”[1] Screening for certain types of cancer can reduce death rates and health care costs. Cancer is set to become a major cause of morbidity and mortality in the coming decades in every region of the world; cancer screening coupled with other targeted interventions such as primary prevention of risk factors, vaccination, and effective treatment programs can ameliorate the projected cancer burden.[2]

Most major medical organizations [3] recommend a population approach to screening for cervical, breast and colorectal cancers for women over 50 years of age. Where available, organized screening programs set participation targets for each cancer screening, develop strategies to reach those targets and monitor the impact of the strategies with performance indicators of participation which ultimately reflect program acceptance and effectiveness.

Rates of adherence to cervical, breast and colorectal cancer screening are commonly reported for each individual cancer. For example in Canada, participation in cervical, breast, and colorectal screening was 70.2%,[4] 46%,[5] and 16.1% [6] respectively. However, for groups in which screening for more than one cancer is indicated (such as in women older than 50 years of age) participation in more than one cancer screening could also be measured. Women between 50 and 69 years of age, for example, are advised to undergo screening for cervical, breast and colorectal cancers, and those in the 70 to 74 age group should be screened for breast and colorectal cancers. For these, an integrated indicator of participation would go beyond each cancer to quantify women screened for two or all three cancers; unless specifically measured, this cannot be inferred from participation in each of three cancer screenings (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Hypothetical scenarios of cancer screening participation.

https://doi.org/10.1371/journal.pone.0161187.g001

There are multiple barriers to achieve an integrated measure of participation in breast, cervical and colorectal cancers, such as the availability of an electronic infrastructure to support data linkage of personal information across different databases. Ideally, databases of clients’ participation in cancer screening would cover large geographic areas to avoid missing clients who are screened in a different jurisdiction. Finally, confidentiality rules of sharing private information across providers may deter access to the data for the indicator. Despite the challenges, integrated cancer screening has been measured in at least one jurisdiction in Canada; in Ontario in 2013, participation in the three cancer screening programs among women 50–69 years old was 35%.[7]

A performance measure that integrates participation across breast, cervical and/or colorectal cancer screenings can improve the quality of cancer screening programs by allowing refinement and balancing of strategies for different target groups. On the one hand, individuals who are up-to-date with some screening may be more open and responsive to receiving additional screenings compared to those who haven't had any screening. For example, women in each section of the Venn diagram (except those in the middle, who are already screened for all cancers they are eligible for) may have different needs, and would therefore benefit from different strategies to increase uptake for one or more screenings they are missing. On the other hand, evidence suggests that screened adults are generally enthusiastic about screening [8], and may thus be possibly over-screened. An integrated measure would allow the pinpointing of populations at risk of harm from unnecessary testing.

The purpose of this systematic review was to examine the literature for the use of performance indicators of adherence across multiple cancer screening (cervical, breast and colorectal) at the population level.

Methods

To identify articles within peer-reviewed literature, a comprehensive search strategy was developed in consultation with a librarian. The key concepts of the search were integrated cancer screening and breast, cervical and colorectal cancer screening. Multiple terms were used to capture literature related to each concept. For integrated cancer screening, search terms included multiple screening, integrated cancer prevention, patient-centred/centered screening, and one-stop shop. For breast, cervical and colorectal cancer screening, the cancer synonyms tumour/tumor, neoplasm and carcinoma were used as well as the following screening tests: mammography, mammogram, papanicolaou test, fecal immunochemical test, fecal occult blood test, colonoscopy, FIT, and FOBT (Fig 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Literature search strategy (applied to all databases).

https://doi.org/10.1371/journal.pone.0161187.g002

Articles were screened at the level of title, then abstract, then full text for relevance by one reviewer. Articles retained at full text were assessed for inclusion based on the following pre-established inclusion criteria outlined in Table 1. Although arbitrary, the rationale for the lower limit of publication date from 2000 onwards was chosen to account for the time frame when CRC screening became more prevalent in Canada [9] and the US [10].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Inclusion and exclusion criteria for the selection of articles.

https://doi.org/10.1371/journal.pone.0161187.t001

The following seven databases were searched: MEDLINE, EMBASE, EBM Reviews(Cochrane Database of Systematic Reviews, ACP Journal Club, Database of Abstracts of Reviews of Effects, Cochrane Central Register of Controlled Trials, Cochrane Methodology Register, Health Technology Assessment, NHS Economic Evaluation Database), PubMed, PubMed Central, CINAHL, and Nursing Reference Center. Results were combined to identify literature that pertained to cancer screening and breast, cervical and colorectal cancers. A hand search through the reference lists of included articles was done to capture any additional articles.

Several grey literature resources were consulted, including Google, Google Scholar, OpenDOAR (www.opendoar.org), Health Sciences Online [HSO] (www.hso.info), OAISter (http://oaister.worldcat.org), Health Canada (http://www.hc-sc.gc.ca/index-eng.php), Public Health Agency of Canada [PHAC] (http://www.phac-aspc.gc.ca/index-eng.php), and the National Cancer Institute [NCI] (http://www.cancer.gov).

Results

The search strategy is outlined in Fig 2. The strategy identified 78 potential articles, supplemented with an additional five records located through grey literature. Upon removal of duplicates and screening via title/abstract and inclusion/exclusion criteria, eight papers were read in full-text, seven of which are included in this review (Fig 3 and S1 PRISMA Checklist).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. PRISMA Flow Diagram [11] of Integrated Screening: Breast, Cervical, and Colorectal Cancer.

https://doi.org/10.1371/journal.pone.0161187.g003

Seven studies reported an integrated measure of participation across breast, cervical and colorectal cancer screening, three cross-sectional studies, [12–14] two randomized controlled trials, [15–16] an evaluation of a triple screening pilot program, [17] and a group pre-post intervention study.[18] The indicators of integrated screening for the three cancers were 51.4% in the pilot evaluation [17], and ranged from 16% to over 8% in the three cross sectional studies.[12–14] In the two randomized trials, the tested intervention almost doubled the number of women screened for all three cancers to 43% [16] and 19% [15]. None of the studies included uptake of combinations of 2 specific cancer screenings (such as breast and cervical, but no colorectal). Only, one study [16] reported uptake measures for “two or more cancer screenings” in the intervention and usual care groups, but details about which specific two cancer screenings were not provided. The seven studies included are summarized in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Summary of included studies.

https://doi.org/10.1371/journal.pone.0161187.t002

Six of seven studies [12–16, 18] were conducted in the US, and one was from France.[17] Four of the US studies addressed disadvantaged groups including Latino, [12,18] Appalachian, [13] and minority and low income women. [16] Clients’ screening histories were collected from electronic medical records, [12,16] administrative data, [15] self- report, [14,18] or a combination of self-report followed by a review of administrative data. [13,17] Only the French study [17] extracted data from a cancer screening database.

The main purpose of the studies were to evaluate interventions to increase cancer screening behavior, [15,16,18] document inequities in cancer screening, [12,13] report the results of a pilot program, [17] and test whether cervical and breast screening adherence are predictors of compliance with colorectal cancer screening.[14] The largest sample sizes were used in the pilot evaluation (around 100,000 participants) [17] and in one of the cross sectional studies (n = 43,000), while for the remaining five studies, sample sizes ranged from around 77 to 3,000 women.

One of the cross sectional studies [12] noted that adherence to multiple screening dropped as the number of tests that women were eligible for increased. Pap test uptake was 65% among young women for which cervical cancer screening is the only screening recommended, but only 16% of women 50–70 years old (eligible for cervical, breast and colorectal cancer screening) and 24% of women >70 years of age (eligible for colorectal and breast cancer screening) regularly underwent all screening tests they were eligible for. In one cross sectional study, [13] 31.9% of the surveyed women reported being current with all three cancer screening tests but only 8.6% were confirmed up-to-date on the subsequent medical record review. Another survey [14] noted that women who adhered to both breast and cervical cancer screening were more likely to adhere to colorectal cancer compared to women who adhered to cervical cancer screening alone; self-reported participation in the three cancer screenings was 8.8%.

In a pilot program [17] of combined triple screening, about 90% of women living in an administrative district of France were invited to participate. Data from the cancer screening database showed that around 40% of the invited women attended, of whom 55% received the triple screening. Therefore, overall adherence with the three cancer screenings was around 22% of the eligible population.

The two randomized controlled trials were conducted by the same group of researchers.[15,16] They tested a Prevention Care Management (PCM) intervention to increase participation in breast, cervical and colorectal cancer screening among a minority, low-income population of women in New York City. Participation in the first (intensive setting) and second (practical) trials showed higher but modest participation among women in the PCM versus usual care groups.

Finally, Allen et al, 2014 [18] assessed the feasibility, acceptability and impact of a church-based educational program to promote cancer screening among Latinas. There was substantial attrition and a non-significant self-reported increase in compliance with breast cancer screening (n = 5) and all tests recommended for age (n = 3). This latter category did not provide details on women’s age or participation in multiple cancer screening tests.

Discussion

The purpose of this study was to review the literature for the use of performance indicators of adherence across cervical, breast and colorectal cancer screening. The number of studies identified was very limited, perhaps due to the fact that integrated indicators are used as internal performance evaluation programs and may not get published. However, reports of adherence to single cancer screening abound in the literature. Among studies that reported integrated screening, levels of participation in all three cancers ranged from slightly over 50% to 8%. Overall, the purpose of the studies included in this review were to document inequities among low income groups, [12–13]^, and to test interventions to increase uptake of cancer screening among vulnerable women, [15–16, 18] or at the population level [17]. Overall, interventions were successful in increasing adherence to multiple screening except for one study with a small sample size, in which the intervention did not reach statistical significance in increasing screening uptake.[18] As expected, results from the studies that used a combination of a self-report followed by a review of administrative data [13] or cancer screening database [17] confirmed that self-reporting may overestimate participation in multiple cancer screening.

The paucity of literature reporting on integrated cancer screening adherence is curious given that population-based surveys (such as the National Health Interview Survey in the US and the Canadian Community Health Survey)[19–21] having the potential to report integrated measures of participation in cancer screening fail to do so. Instead, they report adherence for each cancer screening individually. Only one study included in this literature review used population-based survey responses and concluded that women who adhere to breast and cervical cancer screening were significantly more likely to adhere to colorectal cancer screening, compared to women who adhered to either cervical or breast cancer screening alone.

Integration of cancer screening services is challenging [22] because tests include differing body sites, target groups and testing intervals that vary depending on age, and comfort of the test provider in administering screening and interpreting test results. As such, we were interested in determining whether this literature review would show that adherence with breast and cervical cancer is more common than any other combination of two cancer screenings, as a result of women obtaining screening at “well women” or healthcare visits dedicated to gynecologic health where providers may not be as familiar with colorectal cancer screening. Unfortunately none of the papers included in this review reported on uptake of cervical and breast, or any other combination of two cancer screenings; only uptake of individual cancer screening or the three combined were included.

In the context of a client-centered delivery approach that aims to deliver services in a way that satisfies clients’ needs, [23] (such as a “one-stop-shop” for cancer screening) integrated measures of adherence seek to address all screenings for which an individual is eligible for [24]. An integrated approach to the provision of cancer screening is highly consistent with a patient-centered orientation to medical care and more fruitful than separate visits for each cancer screening test. [12] Integrated cancer screening delivery requires its own targets for performance indicators [25] that reflect uptake of all cancer screenings that clients are eligible for. These indicators can position cancer screening programs as well as individual providers to better understand the needs of their target population and clients. Different messages and strategies could be used if the goal is to recruit screening naïve women into cancer screening compared to women who already participate in one or more cancer screenings, and those at risk of harms from over-screening.

In summary, this review identifies a gap in the literature around integrated adherence to breast, cervical and colorectal cancer screening, which can provide additional information on the screening needs of the population, and have the potential to add value to process improvement initiatives. Integrated participation complements performance measures of individual programs, with implications that can extend beyond cancer screening to other population-based interventions with specific target groups, such as non-cancer screening (e.g. infant hearing screening) and immunizations. Integrated measures of participation may become more widespread in the future with advances in information technology to support linkages of population level data.”

Supporting Information

Acknowledgments

The authors would like to acknowledge Dr. Ignacio Salvati for his critical input into Fig 1.