A 36-year-old female presents to clinic inquiring about the meaning of a previous negative test result from a new HIV screening test. The efficacy of this new screening test for HIV has been assessed by comparison against existing gold standard detection of HIV RNA via PCR. The study includes 1000 patients, with 850 HIV-negative patients (by PCR) receiving a negative test result, 30 HIV-negative patients receiving a positive test result, 100 HIV positive patients receiving a positive test result, and 20 HIV positive patients receiving a negative test result. Which of the following is most likely to increase the negative predictive value for this test?

Decreased prevalence of HIV in the tested population

Increased prevalence of HIV in the tested population

Increased number of false positive test results

Increased number of false negative test results

Decreased number of false positive test results

A mother presents to the family physician with her 16-year-old son. She explains, "There's something wrong with him doc. His grades are getting worse, he's cutting class, he's gaining weight, and his eyes are often bloodshot." Upon interviewing the patient apart from his mother, he seems withdrawn and angry at times when probed about his social history. The patient denies abuse and sexual history. What initial test should be sent to rule out the most likely culprit of this patient's behavior?

Sexually transmitted infection (STI) testing

Study X examined the relationship between coffee consumption and lung cancer. The authors of Study X retrospectively reviewed patients' reported coffee consumption and found that drinking greater than 6 cups of coffee per day was associated with an increased risk of developing lung cancer. However, Study X was criticized by the authors of Study Y. Study Y showed that increased coffee consumption was associated with smoking. What type of bias affected Study X, and what study design is geared to reduce the chance of that bias?

Observer bias; double blind analysis

Two studies are reviewed for submission to an oncology journal. In Study A, a novel MRI technology is evaluated as a screening tool for ovarian cancer. The authors find that the mean survival time is 4 years in the control group and 10 years in the MRI-screened group. In Study B, cognitive behavioral therapy (CBT) and a novel antidepressant are used to treat patients with comorbid pancreatic cancer and major depression. Patients receiving the new drug are told that they are expected to have quick resolution of their depression, while those who do not receive the drug are not told anything about their prognosis. Which of the following describes the likely type of bias in Study A and Study B?

Lead time bias; Pygmalion effect

A student health coordinator plans on leading a campus-wide HIV screening program that will be free for the entire undergraduate student body. The goal is to capture as many correct HIV diagnoses as possible with the fewest false positives. The coordinator consults with the hospital to see which tests are available to use for this program. Test A has a sensitivity of 0.92 and a specificity of 0.99. Test B has a sensitivity of 0.95 and a specificity of 0.96. Test C has a sensitivity of 0.98 and a specificity of 0.93. Which of the following testing schemes should the coordinator pursue?

Test C on the entire student body followed by Test A on those who are positive

Test A on the entire student body followed by Test B on those who are positive

Test A on the entire student body followed by Test C on those who are positive

Test C on the entire student body followed by Test B on those who are positive

Test B on the entire student body followed by Test A on those who are positive

A pharmaceutical company develops a sequential testing protocol for a rare genetic disorder (prevalence 0.01%). Initial screening test has sensitivity 95% and specificity 90%. Positive results undergo confirmatory testing with sensitivity 99% and specificity 99.5%. The company claims this approach achieves PPV >80% for the final positive result. Evaluate this claim and the rationale for sequential testing in this context.

The claim is true; sequential testing increases PPV by enriching the population tested in the second step

The claim is false; sensitivity decreases with sequential testing, reducing PPV

Sequential testing is unnecessary; the first test alone achieves adequate PPV

The claim is false; sequential testing cannot achieve PPV >80% with such low prevalence

The claim is true; the high specificity of the confirmatory test ensures high PPV regardless of prevalence

Screening test selection criteria for USMLE Step 3: High-Yield Behavioral Science Lessons

Screening Fundamentals - The First Net

Disease Criteria:
- Significant public health issue.
- Well-understood natural history.
- Detectable preclinical phase (long enough for intervention).
Test Criteria:
- Safe, acceptable to patients, and inexpensive.
- High sensitivity (to not miss cases) & high specificity (to not over-diagnose).
System Criteria:
- Effective and available treatment for the disease.
- Screening program must be cost-effective.

Lead-time bias in cancer screening

⭐ Lead-time bias: Screening may detect a disease earlier without improving the outcome. This can make it seem like survival is longer, even if the patient dies at the same point in time. It's an apparent ↑ in survival, not a real one.

Sensitivity & Specificity - The Accuracy Twins

Measures a test's ability to correctly identify individuals with or without a disease. They are intrinsic properties of a test.

Metric	Formula	Clinical Use
Sensitivity	$TP / (TP + FN)$	Identifies true positives. High sensitivity is crucial for screening tests.
Specificity	$TN / (TN + FP)$	Identifies true negatives. High specificity is vital for confirmatory tests.

A highly Sensitive test, when Negative, rules OUT disease.
A highly Specific test, when Positive, rules IN disease.

⭐ Sensitivity and specificity are fixed test characteristics, unaffected by the prevalence of the disease in the population. However, Positive and Negative Predictive Values (PPV/NPV) are heavily dependent on prevalence.

Predictive Values & Prevalence - The Real-World Impact

Positive Predictive Value (PPV): Probability that a subject with a positive test truly has the disease.
Negative Predictive Value (NPV): Probability that a subject with a negative test is truly disease-free.
Prevalence Dependency: Predictive values are critically dependent on disease prevalence in the tested population.
- PPV is directly proportional to prevalence.
  - High Prevalence (testing high-risk group) → ↑ PPV
  - Low Prevalence (testing general population) → ↓ PPV
- NPV is inversely proportional to prevalence.
  - High Prevalence → ↓ NPV
  - Low Prevalence → ↑ NPV

⭐ Screening the general population for a rare disease (low prevalence) yields a very low PPV. This means most positive results will be false positives, leading to unnecessary follow-up tests and patient anxiety.

Screening Biases - The Data Traps

Lead-Time Bias
- Apparent increase in survival time due to earlier detection, not because patients actually live longer.
- The disease's natural history remains unchanged.
Length-Time Bias
- Screening is more likely to detect slow-growing, less aggressive diseases with a better prognosis.
- Aggressive diseases often become symptomatic between screening intervals.
Selection (Volunteer) Bias
- Screened populations are often healthier and more health-conscious, leading to better outcomes independent of the screening.

⭐ Both lead-time and length-time biases can cause an overestimation of the benefits of a screening test.

Lead-time bias in cancer screening

Screening tests target asymptomatic, at-risk individuals for early detection.

For screening, prioritize high sensitivity to minimize false negatives (SNOUT); you don't want to miss a case.

For confirmatory tests, prioritize high specificity to minimize false positives (SPIN) and avoid unnecessary procedures.

The ideal screening test is also safe, low-cost, and acceptable to patients.

Prevalence is key; it directly influences a test's Positive and Negative Predictive Values.

Continue reading on Oncourse

CONTINUE READING — FREE

or get the app

App Store|Google Play