Definition and Types of Screening

What is screening? The different elements of a screening process were clearly presented in a World Health Organization (WHO) bulletin published in 1968.¹ These basic principles are listed in Table 35-1. It is obvious that certain criteria have to be met in order to make the screening process a success. Each of these criteria will be discussed when we review the three screening protocols described in this chapter.

TABLE 35-1 Principles of Screening Presented by Wilson and Jungner ^*

1. The condition sought should be an important health problem.

2. There should be an accepted treatment for patients with recognized disease.

3. Facilities for diagnosis and treatment should be available.

4. There should be a recognizable latent or early symptomatic stage.

5. There should be a suitable test or examination.

6. The test should be acceptable to the population.

7. The natural history of the condition, including development from latent to declared disease, should be adequately understood.

8. There should be an agreed policy on whom to treat as patients.

9. The cost of case-finding (including diagnosis and treatment of patients diagnosed) should be economically balanced in relation to possible expenditure on medical care as a whole.

10. Case-finding should be a continuing process and not a “once and for all” project.

* From Wilson JMG, Jungner G: Principles and practice of screening for disease, Geneva, 1968, World Health Organization.

In addition to these basic principles, it is important to remember that the word screening has multiple meanings. Three specific types of screening are described in the WHO bulletin: selective screening, mass public health screening, and surveillance (Table 35-2).

TABLE 35-2 Differences Between Types of Screening

Selective Screening

The selective screening approach is the application of a diagnostic test to a predetermined asymptomatic population that has certain risk factors. These risk factors are identified by other means than the screening tool. For example, a clinical history of cigarette smoking is a risk factor for aneurysm formation. The yield of an abdominal ultrasound will therefore be greater in smokers than in nonsmokers. The problem with this approach is that the individual who has never smoked can still have an aneurysm, although the odds are much lower than for the smoker. He/she will unfortunately not meet the criteria for screening.

Mass Public Screening

The second type of screening is a mass public screening. This approach is to apply the same test to the whole population. Mass screening does not necessarily mean that everybody is screened. Often, a specific demographic criterion such as age, sex, or geographic location is used to focus the process. Testing young children for thyroid hormone levels is such an example. The basic limitation with this approach is cost.

Surveillance Screening

The third type of screening is surveillance screening, and its implications are often ignored. One of the WHO criteria indicates that the disease being screened for has a well-defined natural history and that an intervention is available to treat it. The finding of an ectatic aorta, while not quite an aneurysm, in a patient is such an example. Although the diagnosis of an aneurysm has not been made, the patient is at increased risk for developing an aneurysm. The patient will need monitoring since the risk for aneurysm formation is increased. The appropriate time interval for the surveillance examinations needs to be determined and justified.

Screening for Asymptomatic Carotid Stenosis

The goal of this type of screening is the detection of individuals who are asymptomatic, at risk for stroke, and would benefit from an operative intervention, currently carotid endarterectomy. The Doppler screening test has to be compared to the results of carotid angiography since the benefits of carotid endarterectomy are measured based on the arteriogram and not the Doppler ultrasound examination. This assumes that there is a cut-point above which, by arteriography, a carotid stenosis is considered significant and an intervention is justified. The question then becomes how accurate is a Doppler ultrasound in detecting the stenosis and grading it as compared to arteriography?

Defining Accuracy of Doppler Ultrasound

Accuracy of Doppler ultrasound is measured as the sensitivity and specificity of the Doppler ultrasound examination and shown in Table 35-3. A derived value, the negative predictive value is also very important. One would believe that the positive predictive value of a test would be more important than the negative predictive value; however, this is not the case. A screening Doppler examination has to be as sensitive as possible when compared to the gold standard measurement made by arteriography. While we would like to make sure that the individuals that we classify as negative are without significant stenosis, we need to set a Doppler velocity threshold low enough so that individuals with disease are not being missed. However, by setting a low screening threshold, we detect not only individuals with significant stenosis but also individuals without disease, and the false-positive rate is very high. This leads to a low positive predictive value and a high negative predictive value (Table 35-4). This is the basic dilemma of any screening test: the sensitivity has to be set very high. Therefore, in order to screen a population for asymptomatic carotid artery stenosis, the Doppler velocity threshold for a given degree of stenosis is set lower than normal, and many individuals are misclassified as having a significant stenosis. The overall performance of the process is also affected by disease prevalence.

TABLE 35-3 Key Parameters for a Screening Test

	Disease Present	Disease Absent
Test positive for disease	a	b
Test negative for disease	c	d
Total	a + c	b + d

Sensitivity: a/(a + c)

Specificity: d/(b + d)

Disease prevalence: (a + c)/(a + c + b + d)

Positive predictive value: a/(a + b)

Negative predictive value: d/(c + d)

TABLE 35-4 Estimated Rates of Detection of Disease and False Positives for a High-Accuracy Diagnostic Test^* ^†

Doppler Ultrasound in Large Trials of Asymptomatic Patients

Diagnostic Performance

Two multicenter trials have used Doppler velocity measurements to identify individuals with significant carotid artery stenosis (≥60% diameter narrowing)^2,³ who would then undergo carotid endarterectomy. The Asymptomatic Carotid Atherosclerosis Study (ACAS)² investigators used carotid Doppler ultrasound as a screening test but with the intent of ensuring that almost all individuals who went on to carotid arteriography and surgery would have a lesion of at least 60% in their carotid artery. For this reason, the selected Doppler velocity threshold was set to a high enough value so that 95% of selected individuals would have a lesion causing a diameter narrowing of 60% or more.² This approach yielded a high positive predictive value and also resulted in a low negative predictive value. In essence, many individuals who likely had carotid artery stenoses above 60% were missed.

The ACAS trial ² gives insights into the use of carotid Doppler ultrasound as a screening tool. On the technical side, lessons to be learned from this study are the need for strict adherence to a quality control process, adherence to strict protocols, and selection of an appropriate ultrasound imaging device. On the practical side, ACAS showed the relatively low cost-effectiveness of screening asymptomatic individuals suitable for carotid endarterectomy.

Although the overall sensitivity of carotid Doppler in all ACAS centers was not measured, the overall specificity of carotid ultrasound was above 97%.⁴ Carotid ultrasound was, however, used according to a very strict protocol. Qualifying centers needed to show a strong correlation between Doppler measurements and the results of carotid arteriography. Marked variations between centers and outlier centers with poor accuracy of carotid Doppler⁴ improved with new instrumentation selection and a standardized imaging protocol.⁵ The success of carotid ultrasound in ACAS was based on the use of a standard imaging protocol, a certification program for both sonographers and vascular laboratories, and the implementation of a quality assurance program supervised by a central coordinating center.

Effect of Ultrasound Devices

Differences in imaging devices or imaging protocols likely affected the diagnostic performance of Doppler ultrasound in ACAS ⁶^–⁸ and are therefore likely to affect any screening process that looks at detecting significant carotid artery stenosis in the population. The variability of carotid Doppler ultrasound is measurable at the level of individual laboratories and for different imaging devices,^7,⁹ confirming the observations made in ACAS.⁵ Despite this variability, carotid Doppler ultrasound is likely to perform well as a screening tool. The European Asymptomatic Carotid Surgery Trial (ACST) used locally derived parameters to define percent stenosis by carotid ultrasound and achieved outcome results very similar to those of the ACAS study.³

Defining Thresholds

The effectiveness of carotid artery screening is dependent on the reliability of Doppler velocities to estimate the degree of percent stenosis in the carotid artery. Results are compared to those of arteriography, the gold standard. Since carotid Doppler ultrasound has a sensitivity and specificity of less than 100%, there will always be false positives and false negatives. Falsely elevated Doppler velocity measurement leads to extra costs, as a risk to the patient, when individuals without disease go on to the intervention (false positives). Conversely, individuals with disease can be missed because of falsely low Doppler velocity readings and excluded from the benefits of an intervention (false negative).

The use of carotid Doppler ultrasound to detect the presence of “significant” disease requires the selection of a velocity parameter and of a threshold value for this parameter. Based on ACAS ⁵ and on a consensus recommendation,¹⁰ the internal carotid artery peak systolic velocity is the most reliable of the Doppler velocity parameters and should be used. A strategy that relies on the power Doppler image¹¹ can help triage patients, but this adds an intermediate step and is therefore not likely to be cost-effective. A peak systolic velocity value of 125 cm/sec should suffice for detecting a 50% diameter carotid artery stenosis.^10,¹² Detection of a 70% carotid artery stenosis can be done using a value of 230 cm/sec,^7,^10,¹³ although this value may show greater interlaboratory variability.⁷ A value of 260 cm/sec has been reported for stenosis greater than or equal to 60%,¹⁴ but is higher than that reported for lesions causing 70% or greater stenosis (230 cm/sec).¹³ Of the two, the lower threshold of 230 cm/sec seems more suitable since it has high sensitivity and therefore fewer individuals with disease are likely to be missed (Figure 35-1).

FIGURE 35-1 Incidental finding of a peak systolic velocity of slightly greater than 230 cm/sec in a patient without symptoms who participated in a screening program.

An Overview of Overall Effectiveness

The value of the screening process also depends on the degree of stenosis used to determine the need for a surgical intervention. ACAS results show a benefit of carotid endarterectomy for stenosis of 60% or greater as a relative risk reduction of 55%. This, however, translates to an absolute risk reduction, at 5 years, of 5.9%, or slightly more than 1 morbidity event for slightly less than 100 operated patients each year. Data from a European study also shows an absolute risk reduction of 5.3% over 5 years.³ Based on the data from these studies, 94 to 98 asymptomatic patients with significant (>60% diameter stenosis) need to have carotid endartertectomy performed in order to prevent one stroke. This shows that the yield of the screening process is very low since a large number of patients need to be screened and of the small proportion of identified patients, a large number need to undergo surgery in order to prevent one stroke.

It is the combination of the yield of positive examinations and the yield of the intervention that needs to be factored into an evaluation of the efficacy of the screening process.

Some of the answers to some of these issues can be derived from epidemiologic studies in which carotid ultrasound has been used to estimate the degree of stenosis. In the Cardiovascular Health Study (CHS), a multicenter study of 5888 individuals age 65 years or more, carotid Doppler velocity measurements were used to measure the severity of carotid disease. The prevalence of lesions causing velocities above 250 cm/sec was 1.03%¹⁵ and a still-low 5.5 % for lesions causing Doppler velocities above 150 cm/sec.¹⁵ Figure 35-2 shows on the y-axis the proportion of individuals who are likely to have a Doppler velocity below the value listed on the x-axis. Only a small proportion of the population being screened has values above a selected threshold. Using data from CHS¹⁵ and assuming that individuals with Doppler velocities above 250 cm/sec warrant an intervention, 1030 out of 100,000 individuals age 65 years or more presenting for screening each year would be candidates for endarterectomy. Based on ACAS/ACST, approximately 10.8 strokes would be prevented if the 1030 underwent the intervention. The cost of 100,000 carotid ultrasounds has to be added to the cost of 1030 endarterectomies to give a rough appreciation of the efficacy of using Doppler ultrasound to screen for asymptomatic individuals. Even if we account $100 for the Doppler examination and $2000 for the surgery, the saved “strokes” cost $ 12,060,000. The net cost for saving one stroke each year is $1,117,000. This does not take into account the cost of investigating false-positive Doppler ultrasound studies.