Sensitivity, Specificity, PPV & NPV - Biostatistics

1. A laboratory runs 500 tests daily. The test has 92% sensitivity and 88% specificity. If disease prevalence is 4%, approximately how many false negatives occur per day?
   Answer: 1.6 When you encounter diagnostic test questions involving sensitivity, specificity, and prevalence, you need to work through the 2x2 contingency table systematically to find the specific cell values. Start by calculating the population breakdown from 500 daily tests with 4% prevalence: 20 people have the disease (500 × 0.04) and 480 don't. With 92% sensitivity, the test correctly identifies 18.4 diseased individuals (20 × 0.92), meaning it misses 1.6 cases. These missed cases are your false negatives. The calculation is straightforward: False negatives = Total diseased × (1 - Sensitivity) = 20 × (1 - 0.92) = 20 × 0.08 = 1.6. This confirms answer A is correct. Looking at the distractors: B (2.4) might result from incorrectly using 12% instead of 8% as the miss rate, or from calculation errors. C (20) represents the total number of diseased individuals, not the false negatives—this is a common trap where students identify the wrong cell in the 2x2 table. D (58) appears to come from false positive calculations (480 non-diseased × 12% false positive rate ≈ 58), showing confusion between false negatives and false positives. Remember this key pattern: false negatives always come from the diseased population that the test fails to detect. Always multiply the total diseased cases by (1 - sensitivity) to find false negatives. Don't confuse this with false positives, which come from the non-diseased population and involve specificity calculations.
2. A clinician wants to use a test with 95% sensitivity and 80% specificity as a screening tool. She is concerned about missing cases (false negatives) but also about unnecessary follow-up procedures from false positives. In which population would this test perform most optimally for her dual concerns?
   Answer: A moderate-risk population with 15% disease prevalence balancing reasonable PPV with high NPV At 15% prevalence: PPV ≈ 61% and NPV ≈ 98%, providing reasonable balance. At 40% prevalence (A): PPV is high (86%) but many false positives still occur. At 2% prevalence (B): PPV is very low (10%), creating excessive false positives. At 0.5% prevalence (D): PPV is extremely low (2.4%), making false positives overwhelming despite excellent NPV. Choice C optimally balances the dual concerns of minimizing false negatives while maintaining acceptable false positive rates.
3. In a population of 2000 patients, 120 have the disease. A test identifies 100 of the diseased patients and incorrectly identifies 150 healthy patients as positive. What is the ratio of PPV to NPV?
   Answer: 0.40 When you encounter diagnostic test problems, you need to construct a 2×2 contingency table to organize the data and calculate predictive values systematically. Start by setting up your table: 2000 total patients, 120 diseased, 1880 healthy. The test correctly identifies 100 diseased patients (true positives) and incorrectly identifies 150 healthy patients as positive (false positives). This means 20 diseased patients test negative (false negatives) and 1730 healthy patients test negative (true negatives). Positive predictive value (PPV) = True positives / All positive tests = $\frac{100}{100 + 150} = \frac{100}{250} = 0.40$ Negative predictive value (NPV) = True negatives / All negative tests = $\frac{1730}{1730 + 20} = \frac{1730}{1750} = 0.989$ The ratio PPV/NPV = $\frac{0.40}{0.989} = 0.40$, confirming answer B. Choice A (0.33) likely comes from miscalculating PPV as 100/300 by adding an extra 50 somewhere. Choice C (0.67) might result from calculating PPV as 100/150, using only false positives in the denominator instead of all positive tests. Choice D (0.85) could come from confusing sensitivity (100/120 = 0.83) with PPV. Always draw the 2×2 table when solving diagnostic test problems. PPV answers "If the test is positive, what's the probability of disease?" while NPV answers "If the test is negative, what's the probability of no disease?" Remember that predictive values depend on disease prevalence in the population being tested.
4. In a study of 1000 patients, a diagnostic test yielded 95 true positives, 855 true negatives, 45 false positives, and 5 false negatives. What is the difference between the positive predictive value and the negative predictive value?
   Answer: 31.5% When you encounter diagnostic test questions, you're working with a 2×2 contingency table that requires calculating predictive values - measures of how reliable positive and negative test results are in practice. First, organize your data: 95 true positives, 855 true negatives, 45 false positives, and 5 false negatives. The positive predictive value (PPV) tells you the probability that a positive test result is correct: $PPV = \frac{\text{True Positives}}{\text{True Positives + False Positives}} = \frac{95}{95 + 45} = \frac{95}{140} = 67.9\%$ The negative predictive value (NPV) tells you the probability that a negative test result is correct: $NPV = \frac{\text{True Negatives}}{\text{True Negatives + False Negatives}} = \frac{855}{855 + 5} = \frac{855}{860} = 99.4\%$ The difference is: $99.4\% - 67.9\% = 31.5\%$ Answer A (22.1%) likely represents a calculation error, possibly confusing sensitivity or specificity values. Answer C (67.9%) is actually the PPV itself - a common trap where students provide one component instead of the difference. Answer D (89.4%) might result from incorrectly adding the predictive values rather than finding their difference. Remember that predictive values depend heavily on disease prevalence in your population. In this study, the low disease prevalence (100 positive cases out of 1000) creates a large gap between PPV and NPV. Always double-check that you're calculating the difference between the correct measures, not providing one of the individual values.
5. Prevalence is 10%; sensitivity 90%, specificity 80%. Probability a positive result has the disease?
   Answer: About 33% Out of 100 people, 10 have disease; 9 test positive. 90 without disease; 20% false positive means 18 positive tests. So positive results total 27, of whom 9 truly have disease: 9/27 = 33%. The tempting 90% confuses sensitivity with predictive value; it ignores that most tested people are healthy.
6. A test has sensitivity 90%, specificity 70%, prevalence 20%. What is the PPV?
   Answer: 43% In 1,000 people, 200 have the disease and 90% sensitivity gives 180 true positives. Of 800 without disease, 30% test positive, giving 240 false positives. PPV = 180 divided by (180 + 240) = 43%. The tempting error is using sensitivity 90% directly, which ignores false positives and disease prevalence.
7. A test has been optimized to minimize false negatives. Compared to the original version, the optimized test most likely has:
   Answer: Higher sensitivity and lower PPV When you encounter questions about test optimization, focus on understanding the fundamental trade-offs between sensitivity and specificity. These measures are inversely related - improving one typically comes at the cost of the other. A test optimized to minimize false negatives is designed to catch as many true positive cases as possible. This means increasing sensitivity (the ability to correctly identify positive cases). To achieve this, you typically lower the threshold for calling a test positive, which inevitably increases false positives as well. The correct answer is B because higher sensitivity (fewer false negatives) combined with more false positives leads to lower positive predictive value (PPV). PPV represents the proportion of positive test results that are truly positive - when false positives increase, this proportion decreases. Answer A is incorrect because while sensitivity does increase, PPV decreases rather than increases due to the rise in false positives. Answer C is wrong because sensitivity increases, not decreases, when minimizing false negatives. Answer D is incorrect because optimizing to reduce false negatives actually decreases specificity (more false positives) and doesn't directly optimize for negative predictive value. Remember this key relationship: when a test is made more sensitive (to catch more cases), it becomes less specific (more false alarms), which reduces PPV. Conversely, making a test more specific improves PPV but reduces sensitivity. Understanding these trade-offs will help you tackle any test performance optimization question on biostatistics exams.
8. A diagnostic test has a sensitivity of 85% and specificity of 90%. In a population where the disease prevalence is 2%, what is the positive predictive value (PPV)?
   Answer: 14.5% When you encounter diagnostic test questions involving sensitivity, specificity, and predictive values, you need to construct a 2×2 table and work through the calculations systematically. These questions test your understanding of how disease prevalence affects the interpretation of test results. To find the positive predictive value (PPV), start by assuming a population of 10,000 people. With 2% prevalence, 200 people have the disease and 9,800 don't. Given 85% sensitivity, the test correctly identifies $200 \times 0.85 = 170$ true positives and misses 30 cases. With 90% specificity, the test correctly identifies $9,800 \times 0.90 = 8,820$ true negatives, but incorrectly labels $9,800 \times 0.10 = 980$ as false positives. The total positive tests are $170 + 980 = 1,150$. PPV equals true positives divided by total positives: $\frac{170}{1,150} = 0.148$ or 14.8%, making (A) 14.5% correct. (B) 17.0% might result from calculation errors in the 2×2 table setup. (C) 85.0% confuses PPV with sensitivity—a common mistake since both relate to positive results, but sensitivity asks "what percentage of diseased people test positive?" while PPV asks "what percentage of positive tests are truly diseased?" (D) 90.0% confuses PPV with specificity. Key takeaway: In low-prevalence populations, even highly accurate tests have surprisingly low positive predictive values due to the large number of false positives. Always construct the 2×2 table with actual numbers rather than working with percentages alone.
9. A diagnostic test has PPV = 60% and NPV = 95% in a population with 8% disease prevalence. What is the specificity of this test?
   Answer: 96.8% When you encounter diagnostic test questions involving PPV, NPV, and prevalence, you need to work backwards from these summary statistics to find the underlying test characteristics like sensitivity and specificity. Start by setting up what you know. With 8% prevalence, imagine 1000 people: 80 have disease, 920 don't. Since NPV = 95%, among all negative tests, 95% are true negatives. Let's say there are X positive tests, so (1000-X) negative tests. From NPV: true negatives = 0.95(1000-X), and true negatives also equal (920 - false positives). Since PPV = 60%, among X positive tests, 60% are true positives: 0.6X = true positives. But true positives also equal (80 - false negatives). Setting up the equations: 0.95(1000-X) = 920 - (X - 0.6X), which simplifies to find X = 67 positive tests. This gives us 33 false positives (67 - 40 true positives) and 887 true negatives. Specificity = true negatives/(true negatives + false positives) = 887/(887 + 33) = 887/920 = 96.5%, closest to answer D) 96.8%. The other answers represent common calculation errors: A) 88.3% likely confuses sensitivity with specificity calculations, B) 91.7% might result from incorrectly handling the prevalence adjustment, and C) 94.2% could come from using NPV directly without proper conversion. Remember: when working with PPV/NPV problems, always convert to a concrete population (like 1000 people) using the given prevalence. This makes the 2×2 table calculations much clearer than working with formulas alone.
10. Two diagnostic tests for the same condition have identical sensitivity (80%) but different specificities: Test A has 70% specificity and Test B has 90% specificity. If the disease prevalence is 5%, which statement about their positive predictive values is correct?
    Answer: Test B has higher PPV because higher specificity reduces false positives in low-prevalence populations In low-prevalence populations, specificity has a major impact on PPV. Test A PPV = (0.8×0.05)/[(0.8×0.05)+(0.3×0.95)] = 0.04/0.325 ≈ 12%. Test B PPV = (0.8×0.05)/[(0.8×0.05)+(0.1×0.95)] = 0.04/0.135 ≈ 30%. Higher specificity reduces false positives, improving PPV. Choice A misunderstands the relative importance of specificity in low-prevalence settings. Choice C ignores the impact of specificity on PPV. Choice D confuses more positive results with better accuracy.
11. A research team develops a diagnostic algorithm with 92% sensitivity and 85% specificity. They want to minimize false negative results even if it means accepting more false positives. Which modification would best achieve this goal while maintaining the same underlying test characteristics?
    Answer: Lower the decision threshold to increase sensitivity to 96% and decrease specificity to 75% To minimize false negatives, sensitivity must be increased, which typically requires lowering the decision threshold. This trade-off decreases specificity (more false positives) but achieves the stated goal of fewer false negatives. Choice B increases false negatives by lowering sensitivity. Choice C addresses PPV but doesn't change the test's sensitivity. Choice D maximizes NPV but doesn't specifically target reducing false negatives through threshold adjustment.
12. A screening mammogram has 90% sensitivity and 85% specificity for detecting breast cancer. In a population where 2% of women have breast cancer, what percentage of women with negative mammograms actually do NOT have cancer?
    Answer: Approximately 99.8% This asks for NPV. NPV = (Specificity × (1-Prevalence)) / [(1-Sensitivity) × Prevalence + Specificity × (1-Prevalence)] = (0.85 × 0.98) / [(0.10 × 0.02) + (0.85 × 0.98)] = 0.833 / (0.002 + 0.833) = 0.833 / 0.835 ≈ 0.998 or 99.8%. Choice A incorrectly uses (1-Prevalence) directly. Choice C incorrectly uses specificity. Choice D incorrectly uses sensitivity.
13. A point-of-care test has sensitivity of 75% and specificity of 88%. A physician uses this test in two different populations: Population A (prevalence 30%) and Population B (prevalence 3%). How do the negative predictive values compare between these populations?
    Answer: NPV is approximately 10 percentage points higher in Population B than Population A When you encounter questions about predictive values across different populations, remember that negative predictive value (NPV) depends heavily on disease prevalence, even when test characteristics remain constant. Let's calculate the NPV for both populations using the given sensitivity (75%) and specificity (88%). NPV equals true negatives divided by all negative test results. For Population A (30% prevalence): - In 1000 people: 300 have disease, 700 don't - True negatives: 700 × 0.88 = 616 - False negatives: 300 × 0.25 = 75 - Total negative tests: 616 + 75 = 691 - NPV = 616/691 = 89.1% For Population B (3% prevalence): - In 1000 people: 30 have disease, 970 don't - True negatives: 970 × 0.88 = 853.6 - False negatives: 30 × 0.25 = 7.5 - Total negative tests: 853.6 + 7.5 = 861.1 - NPV = 853.6/861.1 = 99.1% Population B has an NPV approximately 10 percentage points higher than Population A, confirming answer D. Answer A overstates the difference (15 vs. actual 10 percentage points). Answer B incorrectly assumes NPV stays constant—this confuses sensitivity with NPV. Answer C gets the direction wrong, claiming Population A has higher NPV when it's actually lower. Remember: NPV increases as prevalence decreases because there are fewer false negatives relative to true negatives. In low-prevalence populations, negative tests are more reliable for ruling out disease.
14. A test has 80% sensitivity, 95% specificity, and 10% prevalence. What is the NPV?
    Answer: 98% In 1,000 people, 100 have the disease and 900 do not. Sensitivity 80% means 20 infected people test negative, while specificity 95% gives 855 true negatives among healthy people. NPV is 855 divided by 855 + 20, or 855/875, which is 97.7%, about 98%. The tempting 99% is wrong because it overlooks the 20% of diseased patients who falsely test negative.
15. A diagnostic test correctly identifies 180 of 200 patients with disease (true positives) and correctly identifies 760 of 800 patients without disease (true negatives). If this test is applied to a new population where the disease prevalence is 25%, what will be the negative predictive value?
    Answer: Approximately 96.6% First calculate test characteristics: Sensitivity = 180/200 = 90%, Specificity = 760/800 = 95%. Then calculate NPV for 25% prevalence: NPV = (0.95 × 0.75) / [(0.10 × 0.25) + (0.95 × 0.75)] = 0.7125 / (0.025 + 0.7125) = 0.7125 / 0.7375 ≈ 0.966 or 96.6%. Choice A represents a calculation error. Choice B incorrectly uses specificity directly. Choice C incorrectly uses sensitivity directly.
16. A test has sensitivity 75% and PPV 40% in a population with 8% disease prevalence. What is the test's specificity?
    Answer: 91.3% When you encounter diagnostic test questions involving sensitivity, specificity, and predictive values, you need to work systematically through the 2×2 table relationships. These problems test your understanding of how test characteristics interact with disease prevalence. Given the sensitivity (75%), PPV (40%), and prevalence (8%), you can calculate specificity by working through the diagnostic table. Start by assuming 1000 people in the population. With 8% prevalence, 80 people have disease and 920 don't. Since sensitivity is 75%, the test correctly identifies 75% of the 80 diseased people: $0.75 × 80 = 60$ true positives. The remaining 20 diseased people are false negatives. Here's the key insight: PPV tells you that of all positive tests, 40% represent true disease. Since you have 60 true positives and PPV is 40%, you can find total positives: $60 ÷ 0.40 = 150$ total positive tests. This means $150 - 60 = 90$ false positives. Finally, specificity is the percentage of healthy people correctly identified as negative: $\frac{920 - 90}{920} = \frac{830}{920} = 0.913 = 91.3\%$ Answer choice A (86.7%) likely comes from calculation errors in the table setup. Answer B (89.1%) and D (93.5%) represent common computational mistakes when working through the PPV relationship or confusing the sensitivity/specificity formulas. Remember: always set up the complete 2×2 table systematically. Use the given information to fill in what you know, then work step-by-step to find the missing values. PPV problems often require working backward from predictive values to find the test characteristics.
17. Assuming independent repeats, requiring two positive results instead of one changes sensitivity and specificity how?
    Answer: Sens decreases; Spec increases Requiring two positives makes the overall test harder to call positive. Sensitivity is the chance a diseased person tests positive, so it becomes sensitivity squared and falls. Specificity is the chance a healthy person tests negative; with two tests it rises because both false positives are needed to be wrong. The tempting error is thinking extra testing increases both sensitivity and specificity, but sensitivity suffers under a stricter positive rule.
18. In a study comparing three diagnostic tests, Test X has the highest sensitivity, Test Y has the highest specificity, and Test Z has the highest NPV. In a low-prevalence population (2% disease prevalence), which test is most likely to have the highest PPV?
    Answer: Test Y, because high specificity increases PPV especially at low prevalence When you encounter questions about diagnostic test performance across different disease prevalences, focus on how sensitivity and specificity interact with prevalence to determine predictive values. In low-prevalence settings, positive predictive value (PPV) becomes heavily dependent on specificity because most positive tests will be false positives. The formula $PPV = \frac{TP}{TP + FP}$ shows that as prevalence drops, the number of true positives (TP) decreases dramatically while false positives (FP) can remain substantial if specificity is poor. High specificity minimizes false positives, making any positive result more likely to be truly positive. Test Y is correct because high specificity is crucial for maintaining reasonable PPV in low-prevalence populations. Even modest decreases in specificity can flood the system with false positives, devastating PPV regardless of sensitivity levels. Option A is incorrect because while sensitivity affects the numerator of PPV, it doesn't address the false positive problem that dominates in low-prevalence settings. High sensitivity with poor specificity actually worsens PPV by increasing false positives. Option C reflects a common misconception. NPV and PPV are inversely related through the disease prevalence - when prevalence is low, NPV is naturally high regardless of test performance, but this doesn't predict PPV. Option D ignores the mathematical reality that specificity has disproportionate impact on PPV at low prevalence, creating meaningful differences between tests. Remember: In low-prevalence screening scenarios, specificity is king for PPV. High sensitivity helps find cases, but high specificity prevents your positive results from being overwhelmed by false alarms.
19. A test has 92% sensitivity and 88% specificity. If the NPV in a certain population is 98%, what is the approximate disease prevalence in this population?
    Answer: 5% When you encounter questions linking sensitivity, specificity, and predictive values, remember that these measures are interconnected through disease prevalence. The key insight is that NPV depends heavily on prevalence - as prevalence increases, NPV typically decreases. To solve this, you need the NPV formula: $NPV = \frac{TN}{TN + FN}$. However, it's easier to use the relationship: $NPV = \frac{\text{Specificity} \times (1-\text{Prevalence})}{\text{Specificity} \times (1-\text{Prevalence}) + (1-\text{Sensitivity}) \times \text{Prevalence}}$ Substituting the given values (sensitivity = 0.92, specificity = 0.88, NPV = 0.98): $0.98 = \frac{0.88 \times (1-P)}{0.88 \times (1-P) + 0.08 \times P}$ Solving this equation yields a prevalence of approximately 5%, confirming answer B. Answer A (2%) would produce an NPV higher than 98% given these sensitivity and specificity values. Answer C (8%) would result in an NPV lower than 98%, as higher disease prevalence reduces NPV when sensitivity isn't perfect. Answer D (12%) would produce an even lower NPV, making it clearly incorrect. Study tip: For NPV/PPV problems, remember that NPV decreases as prevalence increases (more disease means more false negatives relative to true negatives), while PPV increases with prevalence. Practice working backward from predictive values to prevalence - it's a common exam pattern that tests your understanding of how these measures interconnect.
20. Disease prevalence rises from 5% to 20% with sensitivity and specificity unchanged. What happens to predictive values?
    Answer: PPV rises; NPV falls When disease becomes more common, a positive result is more likely to be a true positive, so PPV rises. At the same time, with more disease in the population, a negative result leaves more real disease among those who test negative, so NPV falls. The tempting error is thinking rising prevalence makes positives less reliable, but prevalence raises the prior probability of disease before testing.

Start this quiz