Report 7 of the Council on Scientific Affairs (A-02)
Full Text

Resolution 505, introduced by the American College of Physicians - American Society of Internal Medicine and referred to the AMA Board of Trustees at the 2000 Interim Meeting, asked: "That the AMA, through the Council on Scientific Affairs (CSA), study the process for endorsing or supporting coverage of screening standards and recommend a process for dealing with resolutions submitted to the House of Delegates (HOD) that pertain to evaluations of screening standards."

Methods

Literature searches were conducted in the MEDLINE database for English-language articles published between 1985 and February 2002 using the search terms mass screening or preventive medicine in combination with methods, standards, predictive value of tests, sensitivity and specificity, and human. Articles were selected that provide information on generally accepted principles of screening, the criteria of effective screening tests, and variables that may confound assessment of such tests. In addition, the AMA Policy Database was searched using the terms screen, screening, or reimbursement to identify relevant policies related to AMA endorsement of screening tests or their reimbursement.

Introduction

In past years, resolutions have been submitted to the AMA House of Delegates seeking either AMA adoption or endorsement* of an organization’s or medical specialty society’s recommendations on the use of specific screening tests (eg, mammography for breast cancer, screening for colorectal or prostate cancer). (*According to Webster’s II 1999 New College Dictionary, adopt means to "take up and make one’s own, as an idea"; endorse means to "to give approval of or support to.") Additionally, resolutions have asked the AMA to advocate for reimbursement of specific screening tests or procedures. Usually, such resolutions have been referred, generating an evidence-based review by the CSA or the Council on Medical Service (CMS).

Delivery of clinical preventive services, including screening tests, has increased in the United States during the past decade.¹ However, many groups and individuals still fail to receive effective preventive services.^2-4 Conceptually, AMA endorsement or public acknowledgement of screening standards may improve recognition and delivery of effective preventive services. However, on occasion the HOD has voted on the merits of a particular screening test in the absence of an independent review or an evidence base that supports clear benefit and improved outcomes in screened individuals. Premature endorsement of screening tests may increase demand and potentially harm some patients.⁵

Through the Division of Clinical Quality Improvement, the AMA develops Quality Care Alerts. These documents alert physicians to important gaps between medical knowledge and practice that may adversely affect the quality of patient care. Alerts are developed in collaboration with relevant national medical specialty societies through their involvement with the AMA in the Physician Consortium for Performance Improvement (The Consortium). Recently, an evidence-based Quality Care Alert entitled "Colorectal Cancer- Screening and Surveillance" was released.

Topics for Alerts are selected in concert with The Consortium, based on input from the specialty societies. Topics may be developed if the disease or condition in question is characterized by a significant burden of suffering (high prevalence and severity of disease), the intervention decreases morbidity and mortality, and a gap exists between evidence and practice.

In responding to Resolution 505 (I-00), this report provides a brief overview of criteria used to establish the validity of screening tests, reviews current AMA policy on the subject, and offers recommendations on how such resolutions should be viewed and acted on by the AMA.

Principles of Screening

Screening constitutes the use of laboratory tests, physical examination, or imaging modalities performed on asymptomatic patients with the intent of identifying subclinical disease. As such, screening differs from clinical investigation, in which tests are ordered after disease is suspected.⁶ The AMA defines screening as health care services or products provided to an individual without apparent signs or symptoms of an illness, injury, or disease for the purpose of identifying or excluding an undiagnosed illness, disease, or condition (Policy H-320.953, AMA Policy Database).

A screening test is the initial action in a screening program, which also includes follow-up tests to confirm the initial test results (eg, biopsy) and treatment given for abnormal findings (eg, surgery, radiation, chemotherapy, pharmacotherapy). Early detection represents discovery of a condition or disease before obvious signs or symptoms have appeared. Screening can be further subdivided into mass screening or individualized screening.⁷ The former is conducted with little regard for the risk profile of the individual patient. Most physicians are involved in the latter, which consists of recommending screening tests in the context of an ongoing patient-physician relationship.

The sensitivity and specificity of a particular test help establish its effectiveness.⁸ The four possible outcomes related to use of screening tests are: (1) true-positive¾ the patient tests positive and has the disease or condition; (2) false-positive¾ the patient tests positive but does not have the disease or condition; (3) false-negative¾ the patient tests negative but actually has the disease or condition; and (4) true-negative¾ the patient tests negative and does not have the disease or condition. Sensitivity equals the proportion of the disease population who have a positive test (true-positives/[true-positives + false-negatives]). The specificity of a test equals the proportion of healthy patients who have a negative test (true-negatives/[true-negatives + false-positives]). Patients and clinicians are often more interested in the positive predictive value of a test, which equals the proportion of patients with a positive test result who actually have the disease (true-positives/[true-positive + false-positives]). This proportion can be used to inform patients of the probability that they will be found to be free of disease after undergoing additional tests or procedures. The relative value of screening tests (particularly for cancer) is also dependent on the predictive value of a negative test result (true-negatives/[true-negatives + false-negatives]), which informs patients about the chances that they have actual disease that will be missed by the test.

Criteria of Effective Screening Tests. For a screening test to be considered effective, certain criteria should be fulfilled, including the following:^7,9,10

• The disease or condition must constitute a significant public health problem (common, with significant morbidity and mortality).

• The screening test must have appropriate sensitivity, specificity, and positive predictive value to detect a sufficiently high proportion of the disease or condition in the preclinical phase.

• Improved health outcomes must be related to screening. Therefore, the natural history of the disease or condition should be known, an acceptable treatment should be readily available, and the potential for cure must be greater among screen-detected patients.

Additionally, to enhance clinical application: (1) the screening test should have sufficient accuracy and reproducibility in the field to ensure transferability across test sites; (2) it should be acceptable to the patient and society; (3) the screening procedure should have a reasonable cost, and the costs of the screening program should justify the benefits; and (4) adequate resources and health services should be available to accomplish the screening and to provide the necessary interventions triggered by a positive test result.

At a minimum, adequate evidence-based reviews evaluating the effectiveness of screening tests should grade the quality of the evidence and directly link recommendations of appropriateness to that evidence. An appropriate benchmark is the methodology used by the United States Preventive Services Task Force for determining the quality of evidence and grading the strength of recommendations for or against preventive interventions, including screening tests.¹⁰ These are listed in the Appendix.

Variables That Confound Assessment of Screening Tests

Several biases are inherent in the conduct of screening tests that can impact on apparent survival measures, thus affecting valid assessment of screening test effectiveness.⁷ In particular, lead time bias makes the assessment of mortality improvements difficult. For example, early detection of cancer or other conditions associated with increased mortality creates a backward shift in the starting point for measuring survival (earlier diagnosis), which may artificially increase incidence and lengthen apparent survival. Length bias can result from overdetection of incidental cancers (eg, prostate). Individuals with more slowly progressive disease will tend to be detected with screening tests. Length bias increases the incidence of early-stage disease and lengthens apparent survival, but has no effect on mortality rates or advanced-stage disease. Both biases make the observation of increased survival an inadequate basis for inferring an effect on mortality.

With mass screening (and certain individualized screening tests), individuals who seek screening are a self-selected group who may be more aware of the disease in question and more health conscious. Selection bias can occur whenever the group actually screened differs from the potential population of individuals to be screened. This bias also can cause apparent increases in survival of individuals with screen-detected disease. Inadvertent misclassification of the cause of death or attribution bias can occur when a screen-detected abnormality is labeled as such on the patient’s chart when in fact this abnormality (eg, prostate cancer) would never have been clinically diagnosed in the absence of screen detection.

Because of these biases, case survival cannot be used to assess the effect of screening on mortality. Rather, prospectively determined mortality from the disease over a follow-up period beginning with randomization should be used. Also, one generally cannot make valid comparisons between screened individuals and those who were unscreened in the past. The only way to determine the degree of benefit without bias is by comparing people offered screening with a group of truly comparable people who are not offered screening.

Some common methodologies used in observational epidemiology, particularly case-control and cohort studies, are sometimes used to evaluate screening. Valid application of these approaches requires that screening has been in place in a community for a sufficient length of time for a benefit to be detectable if it does occur. Case-control studies have limitations because it can be difficult to differentiate a screening test from a diagnostic test, and this imprecision in classification can have a major impact on the results of such studies.¹¹

Potential Harm Associated with Screening and Treatment

The possible harms associated with screening include the psychological consequences of a positive screening result or of an actual diagnosis. These harms assume increased importance when the screen-detected abnormality represents indolent disease (eg, prostate cancer) or a false-positive result. In such cases, patients are needlessly subjected to the increased morbidity and mortality associated with diagnostic procedures and treatments.^12,13 Patients receiving a (false) negative test result also may be harmed because such results delay diagnosis or may generate false beliefs by patients that they have zero risk for disease, thereby affecting future behaviors.^12,13

Most patients understand that risks are associated with diagnostic procedures and treatment interventions. Depending on the test, some do not understand that a screening test is only the first step in a potential cascade of follow-up tests and treatments with various risks and benefits. Screening for asymptomatic or occult disease differs from traditional patient-physician encounters. In the former, physicians seek out individuals (who believe they are healthy) with an offer to optimize their future health. In the latter, patients seek advice and treatment for existing complaints or illness. However, increasingly patients are also making direct requests for screening tests that may have questionable benefits. Because screening has inherent harms, some otherwise healthy individuals experience avoidable, unnecessary injury after screening. Accordingly, physicians should educate patients on the relative benefits and harms of all screening tests.¹⁴

With both benefits and harms attributed to screening tests, a determination of appropriateness ultimately involves a process of weighing tradeoffs. The use of modeling and summary measures (eg, quality-adjusted life years) may be applied in formal attempts to determine whether benefits outweigh harms. Different conclusions may be reached for patients at different levels of risk, where screening might be inappropriate for the average-risk general population, but offer a better tradeoff for patients at high risk.

Relevant AMA Policy  [Editor's Note:  This discussion of AMA policy reflects policy at the time this report was written; ie, Spring 2002.]

Screening. The AMA Policy Database currently has more than 80 items containing the word screen or screening. The AMA recommends the Guide to Clinical Preventive Services¹⁰ as one resource for evaluating the delivery of clinical preventive services (Policy H-410.967). However, the Guide, per se, does not represent AMA policy on screening procedures.

Specific AMA policy statements endorse, support or encourage the use of, or further discussion of:

• General measures to promote or increase awareness of important population-based public health issues such as National Alcohol Screening Day, bone marrow screening, cholesterol screening, mammography screening for breast cancer, and cardiovascular preparticipation screening of student athletes

• Screening and detection of sexually-transmissible diseases (HIV and Chlamydia trachomatis).

• Screening of pregnant women (HIV, hepatitis B virus, genetic counseling)

• Screening (detection of) drug and alcohol use in cases of adolescent injuries, trauma patients, and in employer-based programs

• Clinical use of screening questionnaires or toxicology tests (alcohol abuse, lead poisoning in children)

• Cancer screening (breast, cervical, colorectal, and prostate cancer)

• Screening for hereditary conditions (cystic fibrosis, sickle cell anemia and other genetic disorders)

Despite the fact that several original resolutions have sought specific AMA support or endorsement, the only current policy in which the AMA endorses a group or specialty society’s screening recommendations is Policy H-170.971¾ Management of Disorders of Cholesterol, Triglyceride, and Lipoprotein Metabolism. In this policy the AMA endorses the recommendations of the Report of the National Cholesterol Education Program Expert Panel on Blood Cholesterol Levels in Children and Adolescents and of the American Academy of Pediatrics Committee on Nutrition's Statement on Cholesterol.

Reimbursement of Screening Tests. The AMA Policy Database has more than 200 items containing the word reimbursement. Most of these policies are related to general reimbursement issues. With regard to screening, current AMA policy calls for evaluating on a regular basis the benefits and cost-effectiveness of clinical preventive service guidelines (Policy H-165.880) and encourages the development of policies and mechanisms to assure the continuity, coordination and continuous availability of patient care, including professional preventive care and early-detection screening services, provided the services are cost-effective (Policy H-425.997). The AMA also urges organizations that have developed practice parameters to recognize that these are educational tools, not mechanisms to determine reimbursement or credentialing (Policy H-410.970). Furthermore, any preventive service such as a screening test that is being considered for inclusion in public or private sector insurance products must have evidence-based data to demonstrate improved outcome or quality of life and the cost-effectiveness of the service (Policy H-425.997).

Some policies specifically advocate for:

• sufficient payment levels for Pap smears (Policy H-165.876)
• flexibility in coverage dates for screening tests (Policy H-185.965)
• expansion of codes for screening examinations where appropriate (Policy H-70.956)
• payment for federally mandated screening examinations for illegal aliens presenting for emergency services (Policy H-130.967)
• payment for medical screening examinations and further examination and treatment needed to stabilize an emergency medical condition (Policy H-130.970)
• coverage for screening bone densitometry in those at high risk (Policy H-425.98)
• regular mammography screening of all women eligible under Medicare (Policy H-525.986)

In the cancer area, several policies support the general concept of screening and education programs (ie, breast and cervical cancer) without endorsing or supporting a particular organization or specialty viewpoint.

Insurance Benefits. A Minimum Benefits Package (Policy H-165.975) was established in 1990 during a period in which the AMA supported a phased-in, employer-based (insurance) requirement. The AMA Minimum Benefits Package contained a recommended set of physician services, outpatient services, hospital inpatient services, and home health care services that should be included in a package of minimum health care benefits applicable to a program of required employer insurance. This policy recommends that "routine screening tests and exams" not be covered; however, such tests and exams are not specifically defined.

In 1993, the AMA’s Standard Benefits Package (Policy H-165.925) was adopted. This included a broader range of services, including additional preventive services, which the CMS believed better represented the average level of coverage provided to those with employer-based insurance at that time. AMA support for an employer mandate, combined with the Standard Benefits Package, was the cornerstone of the AMA’s "Health Access America" proposal during the early 1990s. The Standard Benefits Package lists several screening tests for use as preventive services, stratified according to age and also pregnancy status. Some of the recommendations in the Standard Benefits Package deviate from other policy recommendations (eg, appropriate interval for mammography screening).

Because of the transition in AMA policy away from an employer mandate and towards a more flexible approach to health care system reform, the CMS recommended rescinding Policies H-165.925 and H-165.975 at the 2000 Interim Meeting (CMS Report 8- Modifications to the AMA Standard Benefits Package). However, the HOD rejected this approach and adopted as amended a recommendation that the AMA undertake no modifications to either Package at this time.

Summary and Conclusion

Physicians are placed in an awkward position when organizations or specialty societies announce a new screening guideline for which sufficient supporting evidence is lacking. Frequently, these organizations rely on consensus panels to develop recommendations. Such guidelines may reflect the personal biases of the participants, and may conflict with or not be supported by the available data, rendering them controversial.

Like many medical decisions, the decision to use a screening test requires weighing of quantitative information surrounding the test itself, and consideration of the risks and benefits of screening outcomes, as well as qualitative factors, such as individual patients' values and preferences. The extent to which a particular screening test (and program) satisfies key criteria for validity should govern its place in medical practice. Screening tests should meet the minimal criteria of effectiveness based on an evidence-based framework. Formal AMA positions and statements regarding the appropriate use of screening tests should rely on the same approach. Accordingly, evidence, not practice patterns, should be the final arbiter of clinical use and reimbursement recommendations. Current policy advocates that screening tests that are being considered for inclusion in public or private sector insurance products must have evidence-based data to demonstrate improved outcome or quality of life and the cost-effectiveness of the service (Policy H-425.997).

The use of a screening test can assure neither a disease-free nor a risk-free state.¹⁴ Explaining the concepts of uncertainty and risk reduction as a foundation of testing, rather than attempting to reach zero risk, can help patients understand why a screening test may not be beneficial.

RECOMMENDATIONS

The following statements, recommended by the Council on Scientific Affairs, were adopted by the AMA House of Delegates as AMA policy at the 2002 AMA Annual Meeting.

The AMA believes that:

1. Delegates, state, or specialty societies submitting a resolution seeking endorsement or AMA adoption of specific screening tests must also submit an evidence-based review that determines the strength or quality of the evidence supporting their request, and that evaluates the degree to which the test satisfies the minimal criteria for validating the appropriateness of the screening test, which are: (a) the test must be able to detect the target condition earlier than without screening and with sufficient accuracy to avoid producing large numbers of false-positive and false-negative results; and (b) screening for and treating persons with early disease should improve the likelihood of favorable health outcomes compared with treating patients when they present with signs or symptoms of disease.
2. This review will be made available to the reference committee, which will either recommend to the House of Delegates that the resolution be referred to the Council on Scientific Affairs for further study and report back, or not be adopted.

References

1. Coffield AB, Maciosek MV,McGinnis M, et al. Priorities among recommended clinical preventive services. Am J Prev Med. 2001;21:1-9.
2. National Center for Health Statistics. Healthy People 2000 review, 1998-99. Hyattsville, MD: Public Health Service; 1999.
3. National Committee for Quality Assurance. The state of managed care quality 2000. Washington, DC: National Committee for Quality Assurance; 2000.
4. Solberg LI, Kottke TE, Brekke ML. Variation in clinical preventive services. Eff Clin Pract. 2001;4:121-126.
5. Moore W. Screening. more harm than good? Health Serv J. 2000;110(5702):Suppl 3.
6. Ewart RM. Primum non nocere and the quality of evidence: rethinking the ethics of screening. J Am Board Fam Pract. 2000;13:188-196.
7. Nielsen C, Lang RS. Principles of screening. Med Clin North Am. 1999;83:1323-1338. (9)
8. Meyer DL. Screening for disease, cost-effectiveness, and guidelines. Primary Care. 1995;22:591-599.
9. Cochrane A, Holland A. Validation of screening procedures. Br Med Bull. 1969;27:3-8.
10. US Preventive Services Task Force. Guide to Clinical Preventive Services: Report of the US Preventive Services Task F. 2^nd ed. Baltimore, MD: Williams & Wilkins; 1996.
11. Concato J: What is a screening test? misclassification bias in observational studies of screening for cancer. J Gen Intern Med. 1997;12:607-612.
12. Flood AB, Wennberg JE, Nease RF, et al. The importance of patient preference in the decision to screen for prostate cancer. J Gen Intern Med. 1996;11:342-349.
13. Wolf AM, Nasser JF, Schorling JB. The impact of informed consent on patient interest in prostate-specific antigen screening. Arch Intern Med. 1996;156:1333-1336.
14. Doukas D, Fetters M, Ruffin MT, McCullough LB. Ethical consideration in the provision of controversial screening tests. Arch Fam Med. 1997;6:486-490.

Grade Definitions

Good: Evidence includes consistent results from well-designed, well-conducted studies in representative populations that directly assess effects on health outcomes.

Fair: Evidence is sufficient to determine effects on health outcomes, but the strength of the evidence is limited by the number, quality, or consistency of the individual studies.

Poor: Evidence is insufficient to assess the effects on health outcomes because of limited number or power of studies, important flaws in their design or conduct, gaps in the chain of evidence, or lack of information on important health outcomes.

Grades and Recommendations

Grade Recommendation

1. The USPSTF strongly recommends that clinicians routinely provide [the service] to eligible patients. The USPSTF found good evidence that [the service] improves important health outcomes and concludes that benefits substantially outweigh harms.
2. The USPSTF recommends that clinicians routinely provide the service to eligible patients. The USPSTF found at least fair evidence that [the service] improves important health outcomes and concludes that benefits outweigh harms.
3. The USPSTF makes no recommendation for or against routine provision of [the service]. The USPSTF found at least fair evidence that [the service] can improve health outcomes and concludes that the balance of benefits and harms is too close to justify a general recommendation.
4. The USPSTF recommends against routinely providing [the service] to asymptomatic patients. The USPSTF found at least fair evidence that [the service] is ineffective or that harms outweigh benefits.
5. The USPSTF concludes that the evidence is insufficient to recommend for or against routinely providing [the service]. Evidence that [the service] is effective is lacking, of poor quality, or conflicting, and the balance of benefits and harms cannot be determined.
   
   Back to Text