Important Note

Tufts ended funding for its Open Courseware initiative in 2014. We are now planning to retire this site on June 30, 2018. Content will be available for Tufts contributors after that date. If you have any questions about this please write to

Tufts OpenCourseware
  • To enhance students’ understanding of sensitivity, specificity, predictive value positive and predictive value negative
  • To introduce students to emerging concepts of screening

Outside Preparation: Due in Small Group #5


Researchers have developed a new blood test to diagnose pneumocystis pneumonia. The result is expressed in units such that the higher the units, the more likely it is that the patient actually has the disease. Initially, the researchers wanted to set 80 units as the cutoff value such that patients with values less than 80 would be deemed negative while patients with values of 80 or more would be deemed positive. However, the researchers noted that 97.5% of patients without the disease have a value of 80 or higher while 2.5% of patients with the disease have a value of 80 or less.

  1. The researchers believe it is very important to minimize false positive results so they decide to eliminate 80 as the cutoff value. To accomplish this goal, should the researchers select a value higher than 80 or lower than 80 as the cutoff value? Explain your answer.
  2. The researchers decide on a cutoff value that has a sensitivity of 70% and a specificity of 95%. If 200 subjects without the disease have this new test, how many will be told they have a positive test? Show your work.
  3. If 500 subjects with the disease have this new test, how many of them will be told they have a negative test? Show your work.
  4. If a group of 800 subjects with a 20% prevalence of pneumocystis pneumonia had the new test, what percentage of subjects with a negative test would actually have the disease? Show your work.
  5. If a group of 800 subjects with a 30% prevalence of pneumocystis pneumonia had the new test, what percentage of them with a positive test would not have the disease? Show your work.


  1. Review Lecture 11 - Screening notes on screening
  2. Read paper Direct-to-Consumer Marketing of High-Technology Screening Tests; NEJM, 2-14-02, Vol 346, No 7, pages 529-532 ; bring a copy of the article to small group

Approximate Class Schedule:

30 minutes Instructor review of key concepts from Lecture 11
15 minutes Collect Critiques; Review of homework assignment
45 minutes The class will be divided into small groups. Each group will answer the lab’s questions.
30 minutes Class discussion of lab answers

To Be Completed in Small Group Session

The paper comments (page 529) that an expert consensus document from the American College of Cardiology and the American Heart Association published data in 2000 that the sensitivity of electron-beam CT was 80% and its specificity 40% in screening patients for coronary heart disease.

  1. The gold standard in developing these data was coronary angiography.
    1. What is meant by gold standard?
    2. What attributes should a gold standard have?
    3. Is it possible that the sensitivity and specificity of the electron-beam CT could have been different if a different gold standard had been used? Explain your answer.
  2. The authors state (page 529) that the false positive rate was higher in the elderly. Assuming the sensitivity of the electron-beam was the same in the elderly vs. the non-elderly, would the predictive value positive be higher or lower in the elderly vs. the non-elderly? Explain your answer.
    1. Based on what you have read in this paper:
    2. How would you respond to a concerned, asymptomatic middle-aged pt. who would like to have the electron-beam CT to look for heart disease?
    3. Do you think third party payors (insurers) should be required to pay for the electron-beam CT as a screening test for coronary artery disease? Explain your answer.
  3. The authors (page 531) state that, “Trials are underway to evaluate the ability of electron-beam CT and low-dose spiral CT to prevent death due to coronary disease and lung cancer, respectively.” If you were an investigator in such a trial:
    1. how would you define “prevent death” as an outcome? (Think of both statistical and clinical significance in your answer.)
    2. would you consider other secondary outcomes? If no, explain your answer. If yes, what outcome(s) would you select and why?
  4. Assume that electron-beam CT will be used as a coronary artery disease-screening tool on 1,000 patients with a 10% prior probability of having coronary artery disease. Assume the sensitivity of the test is 80% and the specificity is 40%. Assume the test costs $600. Patients who test positive will have the gold standard test, coronary angiography (cardiac catheterization) that costs $3,000.
    1. What is the total cost of the screening program in this group? Show your calculations.
    2. What is the cost of the screening program per case of coronary artery disease diagnosed?
    3. Do you think other factors besides those listed should be included in the cost of the screening program? If yes, what are they?
    4. Assume for purposes of this exercise that for every case of coronary artery disease diagnosed, that one life year is saved (the patient lives one “extra” year because the disease was diagnosed by the screening procedure.) What is the cost of the screening program per life year saved?
  5. (The following question does not pertain to the paper.) Can you think of any potential medical-legal (malpractice) issues that could be raised in the context of screening texts?