Chapter 2-1: How do I determine which screening tests to order?

It seems intuitive that it is best to prevent a disease from occurring at all and next best to diagnose and treat it early. However, there are risks and benefits to every intervention, and it is especially important to make sure an intervention is not going to harm a healthy individual. This chapter focuses on understanding the reasoning behind current screening practices.

1. Screening can be used to identify an unrecognized disease or risk factor in a seemingly well person.

2. Screening can be accomplished by collecting a thorough history, performing a physical examination, or obtaining laboratory tests.

3. Examples of screening include mammography and cholesterol testing.

   1. Mammography can detect unrecognized, asymptomatic breast cancer.

   2. Cholesterol testing can be used

      1. To identify high-risk individuals who do not yet have coronary disease (called primary prevention by clinicians).

      2. To prevent complications in patients with known coronary disease (called secondary prevention by clinicians, not actually screening).

4. The following criteria are helpful in determining whether screening for a disease is worthwhile:

   1. The burden of disease must be sufficient to warrant screening.

      1. Screen only for conditions that cause severe disease, disability, or death.

      2. Consider prevalence of target disease and ability to identify high-risk group, since the yield of screening is higher in high-risk groups.

   2. The test used for screening must be of high quality.

      1. Screening tests should accurately detect the target disease when it is asymptomatic.

      2. Screening tests should have high sensitivity and specificity.

      3. Test results should be reproducible in a variety of settings.

      4. Screening tests must be safe and acceptable to patients.

      5. Ideally, screening tests should be simple and shown to be cost effective.

   3. There should be evidence that screening reduces morbidity or mortality.

      1. There must be effective treatment for the target disease.

      2. Early detection followed by treatment must improve survival compared with detection and treatment at the usual time of presentation; in other words, people in whom the condition was diagnosed by screening should have better health outcomes than those in whom the condition was diagnosed clinically.

      3. The benefits of screening must outweigh any adverse effects of the screening test, treatment, or impact of early diagnosis.

      4. Ideally, benefits and harms are evaluated through a randomized trial of screening (Figure 2-1).

         1. The best outcome to measure is either all-cause mortality or disease-specific mortality, such as breast cancer or prostate cancer mortality.

         2. Outcomes such as cancer stage distribution (ie, whether there are more or fewer early-stage cancers found) and length of survival after diagnosis can be misleading because of lead time and length time biases.

            1. Lead time bias: If early treatment is not more effective than later treatment, the duration of time the individual lives with the disease is longer, but the mortality rate is the same (Figure 2-2).

            2. Length time bias: Cancers that progress rapidly from onset to symptoms are less likely to be detected by ...