The role of imaging in health screening: overview, rationale of screening, and screening economics

Abstract

Imaging screening examinations are growing in their indications and volume to identify conditions at an early, treatable stage. The Radiology Research Alliance’s ‘Role of Imaging in Health Screening’ Task Force provides a review of imaging-based screening rationale, economics, and describes established guidelines by various organizations. Various imaging modalities can be employed in screening, and are often chosen based on the specific pathology and patient characteristics. Prevalent disease processes with identifiable progression patterns that benefit from early potentially curative interventions are ideal for screening. Two such examples include colonic precancerous polyp progression to adenocarcinoma in colon cancer formation and atypical ductal hyperplasia progression to ductal carcinoma in situ and invasive ductal carcinoma in breast cancer. Economic factors in imaging-based screening are reviewed, including in the context of value-based reimbursements. Global differences in screening are outlined, along with the role of various organizational guidelines, including the American Cancer Society, the US Preventive Services Task Force, and the American College of Radiology.

Introduction

One of the predominant applications for imaging examinations is early screening for treatable conditions. The indications for imaging screening for disease are expanding, precipitating a concomitant increase in the volume of imaging examinations requested and performed. In this article, we explore the comprehensive role of imaging in health screening. Following an introduction to the basics of screening, the first section will summarize the benefits and limitations of screening. The second section highlights the economics of screening. Finally, the article concludes with a discussion of international screening guidelines and organizations.

Screening Basics

Early Detection

The primary goal of screening is to identify pathology at an earlier stage than would become apparent via patient symptoms or physical exam (1). This rationale is predicated on the assumption that identification of pathology at an earlier stage results in treatment options that alter the course of the disease and improve the clinical outcomes. Several of the most well-known and performed screening tests involve conditions with discrete stages of progression that can be correlated to specific clinical management options. Colorectal adenocarcinoma, for instance, begins as a dysplastic process that then develops first into a polyp, then carcinoma in situ, and finally adenocarcinoma (2, 3). While the removal of polypoid lesions can often result in curative treatment, true adenocarcinoma is associated with worse outcomes (4, 5). As such, the fundamental purpose of screening colonoscopy is to identify colorectal cancer at an earlier stage of progression, where treatment options are more abundant, and clinical outcomes are more favorable. Similarly, identifying early manifestations of breast cancer such as ductal carcinoma in situ offers a better chance of a curative treatment (6). These principles of screening for precancerous lesions or low-grade cancers hold well in cancers with known progression patterns and a high incidence (Figure 1). This key principle of screening can also be seen in other conditions, such as abdominal aortic aneurysm (AAA). Earlier identification of smaller AAAs may result in lesions that can be treated via endovascular coiling or open surgical repair on an elective basis, while large, ruptured AAAs may result in more limited treatment options and suboptimal outcomes, including high rates of mortality (7, 8).

Figure 1.

Progression of colon cancer and breast cancer.

Top row. The adenoma-carcinoma progression of colon cancer (top) is illustrated. Screening metrics with optical colonoscopy or CT colonography are aimed at identifying and removing precancerous polyps that are easily accessible and their resection prevents the progression to adenocarcinoma. Bottom row. The progression from ductal hyperplasia without atypia to atypia then ductal carcinoma is illustrated. Here, the aim is to identify small early stage cancers through annual screening mammography that are amenable to cure by surgical resection.

Screening Modalities and Patient Outcomes

Various imaging modalities can be employed in screening and are often chosen based on the specific pathology and patient characteristics. Ultrasound, for example, which is often available and does not result in radiation exposure, is frequently employed to screen for aortic disease. Other modalities, such as magnetic resonance imaging (MRI), computed tomography (CT), and mammography offer detailed discrimination of anatomic features and are used in screening for many potentially malignant as well as benign diseases. In addition, several of these techniques can be combined in order to effectively screen for conditions such as breast cancer, depending on family history and other patient-specific factors that may favor a multimodality approach. The calcium score, which is generated from CT imaging of the heart, has emerged as an important tool in the risk stratification of patients with cardiovascular disease. Other non-imaging devices, including colonoscopy, flexible sigmoidoscopy and stool testing are used to effectively screen for various gastrointestinal conditions (2, 6, 7).

Studies exploring the relationship between screening for disease and subsequent clinical outcomes employ a variety of metrics. Literature involving oncologic processes, for instance, often uses quantitative measures of survival, such as overall survival (OS) and progression-free survival (PFS) (9). Other modified instruments, including quality of life years (QALYs) as well as patient-reported outcome metrics (PROMs), have recently been used to assess clinical outcomes following episodes of care. These tools are often used to assess outcomes on both the individual and the population levels.

Screening Test Characteristics: Strengths and Pitfalls

The multitude of diagnostic tests is frequently assessed based on characteristics describing their predictive capacity for the disease of interest. Sensitivity and specificity, for example, are basic metrics of a screening test’s ability to identify true positive and true negative findings, respectively (10). A screening test with better sensitivity and specificity is thus inherently of greater value. By contrast, tests with inferior specificity may result in increased false-positive results, potentially contributing to additional invasive work-ups and greater costs of care. While screening tests have the potential to identify early lesions that can be amenable to treatment, it is often unknown whether the subsequent therapies provide a benefit compared to patients with identical, undiagnosed lesions, resulting in the described phenomenon of overdetection and overdiagnosis (11). However, the process of identifying pathology via screening may be confounded by a variety of additional factors. Lead-time bias, for instance, results when earlier detection of a disease is perceived as being associated with increased survival, even when the condition’s natural progression was not changed by the screening test (Figure 2). In addition, length-time bias can result in situations where diseases with a long latency period are preferentially detected when compared to those with a shorter one and consequently earlier symptomatology (12) (Figure 3). Although these biases can present challenges to clinicians and investigators, they may be ameliorated using techniques such as back-end survival measurement and randomized study design.

Figure 2.

Lead time bias. Illustration showing different courses of a population that undergoes a screening event (top) versus an unscreened population (bottom). Top. The screened population undergoes a screening event, which, in this example, identifies a treatable disease, receives treatment before a clinical onset of disease, which subsequently prolongs life and survival. Bottom. The unscreened population does not undergo a screening event, which results in a tumor progressing to the point of clinical symptoms. As a consequence, treatments are at later stages and survival is decreased.

Lead time bias (middle black box) is illustrated by accounting for the time the disease was identified by the screening event as accounting for survival in years. The true value is only apparent when comparing to the unscreened population, comparing their time point with those who undergo screening.

Figure 3.

Length time bias. Length time bias in the context of screening (cancer screening in this example) refers to the observation that more indolent tumors with longer latency periods are statistically more likely to be detected compared to more aggressive tumors.

Benefits and Limitations of Screening

The principles of screening were first described by the World Health Organization in 1968 and elementally involve the application of a test or tests to an asymptomatic population in order to find individuals who are unlikely to have or develop a given disease, and those who are likely to have or develop it (13, 14). Screening tests are used to detect clinically occult disease, in contrast to diagnostic tests, which are performed to determine the cause of the underlying symptoms. This is a crucial distinction given that screening programs have limitations in addition to benefits (13).

Benefits of screening

A significant, generic benefit of disease screening is that the test may provide an indication of underlying disease before any symptoms develop, for example, fecal occult blood testing for asymptomatic colorectal cancer—a relatively simple and non-invasive test to detect the presence of fecal blood, which could be present secondary to underlying colorectal malignancy. A consequent benefit of screening is the opportunity to manage the disease with early interventions. The disease may be discovered at an earlier stage, when treatment is more likely to be curative, and which may lead to better outcomes such as decreased morbidity and mortality.

Limitations of screening

While screening for specific diseases has demonstrated clear benefits, there are several important limitations. First, given the inherent performance characteristics of screening tests (maximal sensitivity with as high as possible specificity), there is the inescapable risk of generating either false-positive or false-negative results. False-positive screening results lead to unnecessary patient anxiety, as well as the application of additional, more invasive, and potentially expensive, follow-up investigations. On the contrary, false-negative screening results lead to the patient not being offered further relevant investigations and treatment for the present, underlying disease, or on the provision of false reassurance, the patient may fail to seek medical attention if he or she becomes symptomatic (15). Second, any individual may be reluctant to be screened if the screening test or procedure is painful, (mammography, for example), time-consuming or complex. Fear of a receiving a potential cancer diagnosis is another reason for avoidance or reluctance to undergo screening (16). Complications related to the work-up of incidental findings unrelated to the disease the screening examination is assessing for is another consideration that may deter patient perception of screening (17). Consequently, any of these factors may reduce screening program participation rates (18). Related, recently published randomized controlled trials on CT colonography versus flexible sigmoidoscopy or colonoscopy both demonstrated lower participation rates in the more invasive screening arms (15). Finally, screening tests associated with radiation exposure may cause concern to patients with radiation’s reported stochastic effects (whether actual or theoretical, is beyond the scope of this work) (19).

In theory, screening test benefits and limitations may be estimated in randomized controlled trials; however, these trials are limited by the fact that there is no discrete study endpoint for screening studies since the process of screening is an ongoing one (15).

Economics of screening

Variations in costs

A central rationale and purpose of screening is its overall economic benefit and cost-effectiveness. The rationale of economic feasibility of screening in healthcare is to have the cost of the screening program be in proper relation to the total cost of treating the disease; however, this does not always balance out. Well-established imaging screening examinations validated by prospective studies and supported with guidelines generated by professional societies demonstrate the value of radiology. Value in healthcare may have different interpretations and definitions, however, one of the most common definition of value is outcome divided by cost. Outcome in screening examinations can include detection of early cancer or precancerous lesions, prevention of developing advanced malignancies, decreasing cancer related deaths or even more broadly: improved life expectancy, improved quality of life, and the willingness for insurers and individuals to pay (20). Cost may be attributed to the number needed to screen to detect one cancer (or prevent one cancer death) (21), individual and insurance carrier costs, and, more broadly: direct medical and non-medical costs along with indirect costs (21). Cost is driven by multiple factors both for healthcare systems, patients, and the economy and may have substantial regional variability (22, 23). One study using several national databases estimated that the productivity loss from cancer mortality in the US was over $115 billion in the year 2000 and estimated to increase to over $147 billion for 2020 (24). The value in screening can be assessed in several ways, but the general concept is that early and ongoing participation in screening programs has a positive impact on a potential or actual late outcome (detecting and treating early cancer or precancerous lesions and preventing the development of advanced cancer) (25).

Current insurance structure

Currently, screening mammography, low dose chest CT screening, and screening for hepatocellular carcinoma with various modalities may be covered by private insurers or Medicare beneficiaries whereas CT colonography is only covered by certain private insurers and has been previously denied coverage by Medicare, stemming from a decision letter issued in 2015 (26). All of these tests are typically performed in a fee-for-service arrangement where payments are not bundled and treated as separate episodes for both the technical and professional components of the imaging examinations.

Recent changes in Centers for Medicare and Medicaid Services

In the US, recent changes in the current implementation period for Centers for Medicare and Medicaid Services (CMS) may affect screening reimbursement now and in the near future. In 2015, the Medicare Access and Children’s Health Insurance Program Reauthorization Act (MACRA) was passed with a strong bipartisan vote. The purpose of this MACRA legislation was to incorporate surrogates of value into Medicare reimbursements. Specifically, the two main goals of the MACRA legislation were to emphasize quality or value in fee-for-service and alternative payment models from its implementation in 2015 to 2026. MACRA eliminated the previous model for fee-for-service annual adjustment factors. Once MACRA is in full effect, clinicians will receive payments through either Merit-Based Incentive Payment System (MIPS), traditional fee-for-service model with positive or negative adjustments to payments based on a range of performance metrics, or advanced alternative payment models (APMs). MIPS has similar performance metrics to APMs, and bears a more than nominal financial risk through participating in the APM (27). APMs have several models and require use of electronic health records. The first positive or negative adjustments for programs participating in MIPS was calculated in 2019 for services performed in 2017. To stabilize this transition period, current fee-for-service payments were stabilized as part of the 2015 MACRA legislation for 5 years with a 0.5% annual increase. An original goal of MACRA was to transform 50% of fee-for-service to APMs by 2018, a goal that was not met and subsequently was abandoned. The concept of a single global payment for all rendered services for a defined episode of care raises similar concerns for all providers, including radiologists. The challenge our specialty faces is to ensure that our contribution to patient management are appropriately recognized. An area of concern with MIPS and APMs is that some physician income or expected reimbursements may be at risk if they fall on the lower end of the set performance metrics. Some centers have or have proposed to protect against this by withholding a portion of one’s earnings (such as a salary for an academic physician) to protect against penalization (28).

Under the new MACRA legislation, once the programs are fully into place (key years of implementations being 2020 and 2026), providers that see greater than 100 Medicare beneficiaries annually will get reimbursement through APMs or traditional fee-for-service with a performance-based adjustment (reward or penalty) through MIPS. Some radiology relevant MIPS include documenting the percentage of eligible patients that should receive screening for breast or colorectal cancer (reported and calculated as a percentage), not using a probably benign (BI-RADS category 3) for a screening mammogram, and having a reminder system for patients undergoing screening mammography (29).

In preparation for MACRA’s serial implementation, starting in 2017, the American College of Radiology (ACR) has released annual performance measure for diagnostic radiology for CMS quality reporting (30). Although some of the 2017 and 2018 measures deals with follow-up in general for imaging findings (2017 measure of recommended follow-up for imaging findings), none of the described performance measures at the time of this writing (the 2017 and 2018 measures are published) directly refers to screening imaging examinations referred to in the present work (30).

Screening in context of bundled payments

Bundled payments are a specific type of APM in which all aspects of care for a single diagnosis or procedure are reimbursed under a single payment, which is determined in a prospective model at a predetermined fixed sum, as opposed to a retrospective system where benchmarks are used and then reconciled against fee-for-service payments (31). Bundled payment models are a subset of APMs, aimed at incentivizing reducing episode-of-care costs. These are a potential means of reimbursement that some authors have recently explored for use in screening mammography. Hughes et al. (32) created a theoretical model of bundled payments for breast cancer screening. Using CMS 5% Research Identifiable Files, their study examined patients who underwent screening mammography in 2010–2012, designating each of the three years as a separate cohort, and ultimately reporting the 2012 CMS cohort and comparing it to a 2013 cohort of a large US northeast health system with both Medicare and private insurance beneficiaries. The authors proposed three theoretical models that included reimbursements for the technical and professional fees of imaging examinations associated with mammographic screening: the first included only screening and diagnostic mammography, the second included screening and diagnostic mammography as well as breast ultrasound procedures, and the third included screening and diagnostic mammography, breast ultrasound procedures, and breast MRI procedures. In the main theoretical cohort, the 2012 Medicare beneficiaries, 226,423 of an eligible 679,984 women underwent mammographic screening. The three bundled payment models including the most basic one comprising only screening and diagnostic mammography, had a $161.94 global payment, which increased by 4.5% when adding ultrasound ($169.29), and 6.4% ($172.67) when adding ultrasound and MRI. When excluding high-risk women, there was no substantial change in price for the global payments (<0.5%). The authors proposed that this model could be used for non-patient-facing specialties including diagnostic radiology in an effort for participating in MIPS. The authors also postulated that this model could be used for other cancer screening events, not necessarily limited to screening centered around imaging (32).

Quality metrics

Several quality metrics can be used to gauge the economic effectiveness of screening examinations. The most common and practical example is number of years of life saved. One study estimated a 1% annual reduction in lung, colorectal, breast, pancreatic, brain cancer, and leukemia would result in lost annual productivity by approximately $814 million per year. In the same study, the annual lost productivity cost ranged from over $115 billion in the model year 2000 to $147 billion projected for 2020. Of those annual costs, lung cancer death was a substantial contributor of all cancer deaths, accounting for 27% of the productivity loss from cancer death in the US (24). A cost-effectiveness analysis of the National Lung Screening Trial (33), which included over 53,000 persons randomized to receive low dose chest CT or chest radiographs annually for 3 years with just over 5 years of follow-up concluding the last year of screening, found low-dose chest CT conferred an additional cost of $1,631 per individuals screened and provided an additional 0.03 life-year per person. The subsequent incremental cost-effectiveness ratio was a cost of $52,000 to $81,000 per life-year or quality-adjusted life-year gained in the overall cohort.

Screening elderly patients

In an aging population, projections estimate that the proportion of adults in the US over 65 years will increase from 13.7% in 2012 to 16.8% and 21.0% in the years 2020 and 2040, respectively (34). Given this, how patients should pursue established screening programs is an important consideration. Some have advocated that at least 10 years life expectancy is an important evolution, along with a discussion with patients on the risks and benefits of continued screening (35). Validation cohorts of screening examinations may not specifically design or perform analyses on older patients, such as the age of Medicare beneficiaries (age 65 and older). Such limitations may affect expert society recommendations and, in some cases, reimbursement. For example, Medicare denied coverage of CT colonography in 2015, and one of their reasons was that the mean age in previous trials of CT colonography was below the average Medicare beneficiary age (26). This subsequently led to specific studies to demonstrate the performance of CT colonography in older patients (36). Many societal guidelines have specific narratives on decisions on screening in older adults. For example, the American Cancer Society (ACS) suggests that women should continue screening mammography if they are in good health and have a life expectancy of greater than 10 years (37). As another example, the ACS recommends that decisions to continue colorectal cancer screening in patients 76 through 85 years be based on patient preference, life expectancy, health status, and prior screening status, and that adults greater than 85 years of age should be discouraged from ongoing screening (38).

International Screening Guideline and Organizations

Over the past half-century, multiple international and national screening programs have been established for malignant and non-malignant conditions, as well as for antenatally-detected conditions (15). Globally, these programs differ with respect to the types of screening tests offered, the demographics of the groups targeted, recommended testing frequency, and screening cost (13).

National and international organizations influence screening program promotion and implementation. For example, in the United Kingdom (UK), the National Screening Committee (UKNSC) assesses evidence for screening programs and subsequently advises the government about possible program implementation (13). The International Cancer Screening Network (ICSN) is a voluntary consortium of 33 countries within Europe, North America, the Middle East and Australasia, all of which have established, population-based screening programs (15). The ICSN not only aims to improve breast, colorectal and cervical cancer screening processes and outcomes, but also aims to promote evidence-based cancer screening implementation and evaluation worldwide, and serve as a resource for countries that are developing national screening programs. Finally, the Medical Screening Society (MSS), established in 2002, provides a forum for discussion for individuals and groups involved in medical screening (39). The MSS also creates working groups to examine problems and controversies in medical screening and publishes scientific reports related to medical screening within their Journal of Medical Screening.

Existing screening guidelines

Numerous organizations are involved in the development and the publication of screening guidelines. Given their ability to non-invasively detect disease, imaging tests have become essential in several screening programs. Consequently, dedicated, imaging-based screening guidelines also exist (22). Some examples of general and imaging screening guidelines are discussed in this section.

American Cancer Society (ACS) Cancer Screening Guidelines

Since 1980, the ACS has introduced and updated imaging and non-imaging screening tests by cancer type and by age, to inform clinicians, the public and health policy (40). Historically, these screening guidelines have involved tests to detect cancers of the breast, colon, rectum, cervix, endometrium, lung and prostate.

Reviewed and updated in 1997, the ACS process now involves a more formal process which includes nine steps for the development or revision of guidelines, which include: oversight by a multidisciplinary panel; defined objectives per cancer type and target groups; outcome modeling; documentation of rationales for each recommendation; review of specific types of scientific evidence; peer review; and periodic evaluation and review of guidelines (40).

US Preventive Services Task Force (USPSTF) Recommendations for Primary Care Practice

Created in 1984, the USPSTF is a volunteer panel of experts in evidence-based medicine and preventative medicine and includes members from the fields of internal medicine, family medicine, pediatrics, behavioral health, nursing, and obstetrics and gynecology. Their clinical scope of screening recommendations includes lung cancer, abdominal aortic aneurysms, BRCA-related cancers, human immunodeficiency virus, and others (41).

A key philosophy of the USPSTF is that the task force does not consider the costs of a screening service when determining recommendations. Further, the recommendations address services only offered in, or referred within, the primary care setting. Annually, the USPSTF makes a report to the US Congress that identifies critical gaps in research and recommends priority areas of focus for further examination. Topics currently in the developmental stage of research plan development include screening for: hearing loss in older adults, hepatitis B virus infection in non-pregnant adolescents and adults, and vitamin D deficiency in adults (41).

National Institute for Health and Care Excellence (NICE) Clinical Guidelines

Based in the UK since 1999, NICE is a non-departmental public body the United Kingdom Department of Health. NICE was established to reduce variation in the quality and availability of health care services within England, Wales, Scotland and Northern Ireland, and consequently provide national guidance, quality standards and information services for health care organizations, including screening recommendations (42).

The NICE Centre for Guidelines involves stakeholders from patient organizations, professional bodies and royal medical colleges, who develop and maintain clinical, public health and social care guidelines. The Clinical Guideline portfolio includes screening recommendations for cervical cancer, antenatal screening, bowel, and breast cancer, and abdominal aortic aneurysms, among others (42).

American College of Radiology (ACR) Appropriateness Criteria

On recognizing a need for national guidelines to assist healthcare providers in making the most appropriate imaging or image-guided treatment decisions, the ACR established the ACR Task Force on Appropriateness Criteria in 1993. The specific aim was to create published, consensus- driven, evidence-based, nationally accepted guidelines for imaging resources, keeping in mind finite health care resource availability (43).

The ACR Task Force currently consists of 11 panels covering topics within breast, cardiac, gastrointestinal, interventional radiology, musculoskeletal, neurological, pediatric, thoracic, urological, vascular and women’s imaging. Each panel has published no less than nine sets of recommendations, each, on a variety of sub-topics. New topics are added as required, (for example, guidelines for breast imaging of pregnant and lactating women, the management of shoulder pain or suspected new-onset or non-acute heart failure), and many are revised annually (43).

Conclusion

Imaging-based screening allows for the detection of pathologic conditions at earlier stages compared to detection due to symptoms or physical examination findings. Radiologists play an integral role in performing and interpreting imaging-based screening tests for treatable conditions with the goal to decrease false-positive findings, better differentiate aggressive malignancies from more indolent ones, minimize overdiagnosis and overtreatment, decrease the radiation dose associated with screening modalities and establish best practices for managing incidental findings.

The economics of screening determine how screening programs are implemented in different countries and aim to reallocate resources necessary to maximize health gains from limited health care resources. If the screening tests are cost-prohibitive, adherence to screening recommendations by some societies and patient participation rates can be adversely affected. Introduction of fee for performance in medicine represents an opportunity for radiologists to adhere to high quality standards with pay for performance providing financial incentives for the investment of practices meeting defined benchmarks.

There are a number of organizations that publish guidelines on screening examinations. It is important to know different organizations recommendations and how they differ or when they agree. In general, an optimal imaging test used in population-based screening should be reproducible and inexpensive. The major costs associated with screening are technology costs and personnel costs. In the future, it is thought that screening may be streamlined due to advances in artificial intelligence and machine learning, especially in examinations with large data sets.