Examkin

Question

Difficulty: Very hardDiagnostic Testing Performance

An epidemiology task force is redesigning several disease screening and diagnostic protocols to optimize resource allocation and predictive values. Match each proposed modification in the testing strategy (Left) with its most direct and mathematically certain impact on the diagnostic test's performance metrics (Right).

  • Adjusting a quantitative immunoassay's positive threshold downward from 4.0 ng/mL to 2.5 ng/mL to capture early-stage disease.Decreases the False Omission Rate and increases the Negative Predictive Value (NPV), but causes a reciprocal decline in Positive Predictive Value (PPV).
  • Deploying a rapid point-of-care test (LR+ = 2.5, LR- = 0.8) universally in an asymptomatic cohort with a disease prevalence of 0.5%.Generates a predominantly high False Discovery Rate, resulting in a post-test probability of disease that remains clinically negligible despite a positive result.
  • Modifying a workflow to require a highly specific confirmatory test only after an initial positive sensitive screening test, rather than running both simultaneously.Maximizes overall diagnostic specificity and Positive Predictive Value (PPV), at the expense of an increased False Negative Rate (FNR).
  • Redirecting a diagnostic assay from a general primary care screening pool to a specialized tertiary care symptomatic referral cohort.Substantially elevates the Positive Predictive Value (PPV) and the False Omission Rate, given that intrinsic test sensitivity and specificity are fixed.

Answer

The correct pairings map the adjustment of thresholds to inverse predictive value shifts, the application of weak LRs in low prevalence to high False Discovery Rates, serial testing to maximized specificity/PPV, and increased prevalence to elevated PPV and False Omission Rates.
Each modification directly alters the 2x2 contingency table. Lowering cutoffs shifts the threshold into the healthy distribution, increasing sensitivity and NPV while lowering PPV. Applying a test with an LR+ of 2.5 in a 0.5% prevalence cohort results in a post-test probability of just 1.2%, meaning nearly all positive results are false alarms (high False Discovery Rate). Serial testing mandates passing multiple specific hurdles, elevating overall specificity and PPV but sacrificing sensitivity. Lastly, increasing prevalence in a tertiary setting intrinsically boosts PPV but degrades NPV (increasing the False Omission Rate).

Step-by-Step Solution

1
Analyze the impact of lowering a quantitative cutoff threshold.
Lowering the cutoff increases sensitivity (fewer False Negatives) and decreases specificity (more False Positives). NPV goes up (False Omission Rate goes down) and PPV goes down.
Lowering the threshold captures more true cases but also incorrectly captures more healthy individuals.
2
Calculate the impact of a weak test in a very low prevalence setting.
Pretest odds = 0.005. Post-test odds = 0.005 * 2.5 = 0.0125. Post-test probability = ~1.2%. The False Discovery Rate (1 - PPV) is ~98.8%.
When prevalence is extremely low, the sheer volume of true negatives overwhelms test specificity, creating a high proportion of false positives.
3
Evaluate the net statistical effect of transitioning from parallel to serial testing.
Serial testing requires multiple sequential positive results to define disease presence. It maximizes specificity and PPV, but reduces total sensitivity, increasing the False Negative Rate.
Serial testing minimizes false positives by applying sequential filters, but allows true positives to be falsely excluded at each step.
4
Analyze the impact of moving a test to a higher prevalence (tertiary care) population.
PPV increases significantly. However, as the absolute number of diseased patients rises, the proportion of False Negatives outpaces True Negatives, raising the False Omission Rate (lowering NPV).
Predictive values are highly dependent on disease prevalence, while intrinsic parameters (sensitivity/specificity) remain static.

Key Concept

Impact of Prevalence, Cutoff Thresholds, and Serial/Parallel Testing on Diagnostic Test Metrics
Estimated Time:2m 0s
Rate this question

Topics

Diagnostic Testing Performance

More questions from this topic