Biostatistics Quiz: Sensitivity And Specificity
20 questions · exam conditions
0:00
Sensitivity And SpecificityQuestion 1 of 20

A screening program uses a test with 95% sensitivity and 80% specificity. After one year, the program director reports that the test's 'accuracy' has improved to 88% compared to 82% in the previous year. What is the most likely explanation for this improvement?

The disease prevalence in the screened population decreased over time
The test's analytical precision improved due to better laboratory procedures
The test's sensitivity increased while specificity remained constant over time
More experienced technicians began performing the test with fewer errors
The screening program began testing a higher-risk population with more disease
← Back to quizzes

Biostatistics Quiz

Biostatistics Quiz: Sensitivity And Specificity

Practice Sensitivity And Specificity in Biostatistics with focused quiz questions that help you check what you know, review explanations, and build confidence with test-style prompts.

What this quiz covers

This quiz focuses on Sensitivity And Specificity, giving you a quick way to practice the rules, question types, and explanations that matter most for Biostatistics.

How to use this quiz

Try each quiz question before looking at the correct answer. Use the explanations to review missed ideas, then come back to similar questions until the pattern feels familiar.

All questions

Question 1

A screening program uses a test with 95% sensitivity and 80% specificity. After one year, the program director reports that the test's 'accuracy' has improved to 88% compared to 82% in the previous year. What is the most likely explanation for this improvement?

  1. The disease prevalence in the screened population decreased over time (correct answer)
  2. The test's analytical precision improved due to better laboratory procedures
  3. The test's sensitivity increased while specificity remained constant over time
  4. More experienced technicians began performing the test with fewer errors
  5. The screening program began testing a higher-risk population with more disease
Explanation: When you encounter questions about test accuracy changes, think about how prevalence affects the overall performance metrics of diagnostic tests. Accuracy represents the proportion of all test results (both positive and negative) that are correct. The key insight is that accuracy is heavily influenced by disease prevalence, even when sensitivity and specificity remain constant. In a population with lower disease prevalence, there are more true negatives relative to true positives. Since this test has higher specificity (80%) than the complement of sensitivity (5% false negative rate), the test performs better on healthy individuals than diseased ones. When prevalence decreases, the proportion of healthy people increases, leading to more correct negative results and higher overall accuracy. Let's examine why the other options are incorrect. Option B suggests analytical precision improved - but the question states the test still has 95% sensitivity and 80% specificity, indicating no change in the test's fundamental performance characteristics. Option C claims sensitivity increased while specificity stayed constant, but this contradicts the given information that both values remain at 95% and 80%. Option D proposes fewer technician errors improved performance, but again, this would change the reported sensitivity and specificity values, which remained stable. Option A correctly identifies that decreased disease prevalence would increase accuracy while keeping sensitivity and specificity unchanged, because the test encounters a higher proportion of healthy individuals (whom it identifies correctly 80% of the time). Study tip: Remember that accuracy can change with prevalence even when sensitivity and specificity stay constant - this is a common trap in biostatistics questions about screening programs.

Question 2

A research team develops a new biomarker test and reports sensitivity of 92% and specificity of 87% based on a study of 50 diseased and 50 non-diseased individuals. A clinician wants to use this test in their practice where they see approximately 1,000 patients per year with 8% disease prevalence. What is the most important limitation of applying these test characteristics to the clinical setting?

  1. The study's 50% prevalence overestimates positive predictive value compared to the 8% clinical prevalence
  2. The small study sample size (n=100) provides insufficient precision for the reported sensitivity and specificity estimates
  3. The test characteristics may not generalize from research participants to real clinical patients (spectrum bias) (correct answer)
  4. The 92% sensitivity is inadequate for clinical decision-making in a population with 8% prevalence
  5. The study design should have used a cohort study rather than a case-control design for better validity
Explanation: When evaluating diagnostic tests, you must distinguish between analytical validity (how well the test performs in controlled research) and clinical utility (how well it performs in real-world practice). This question tests your understanding of potential biases that affect generalizability. The most critical concern here is spectrum bias - the possibility that research participants differ systematically from actual clinical patients. In controlled studies, "diseased" patients often have more advanced or clear-cut cases, while "non-diseased" controls are typically healthy volunteers. Real clinical patients present a spectrum of disease severity, comorbidities, and confounding conditions that can dramatically alter test performance. A biomarker might easily distinguish severe cases from healthy controls but perform poorly on early-stage disease or patients with similar conditions. Option A incorrectly focuses on positive predictive value. While prevalence does affect PPV (lower prevalence decreases PPV), the question asks about limitations of applying the test characteristics themselves, not their predictive values. Option B misidentifies sample size as the primary concern. While n=100 isn't large, it's adequate for initial estimates of sensitivity and specificity, and this doesn't address the generalizability issue. Option D makes an unfounded judgment about adequacy. A 92% sensitivity could be clinically useful depending on the context - the prevalence alone doesn't determine whether sensitivity is "adequate." Study tip: When you see questions about applying research findings to clinical practice, always consider whether the study population matches the target population. Spectrum bias is a frequent concern in diagnostic test evaluation.

Question 3

A point-of-care glucose test is compared to laboratory reference values in 300 diabetic patients and 200 non-diabetic individuals. The test correctly identifies 270 diabetic patients and incorrectly identifies 30 non-diabetic individuals as having elevated glucose. If a modified version of the test achieves the same sensitivity but improves specificity to 95%, how would this change affect the number of false positive results in the same population?

  1. False positives would decrease from 30 to 10 (correct answer)
  2. False positives would decrease from 30 to 15
  3. False positives would remain at 30 since sensitivity is unchanged
  4. False positives would decrease from 30 to 20
  5. False positives would increase to 35 due to the sensitivity-specificity trade-off
Explanation: When you encounter diagnostic test questions involving sensitivity and specificity changes, focus on understanding how these measures relate to different populations and error types. Let's work through the calculations systematically. Initially, the test correctly identifies 270 of 300 diabetic patients (sensitivity = 270/300 = 90%) and incorrectly identifies 30 of 200 non-diabetic individuals as positive (specificity = 170/200 = 85%, with 30 false positives). The modified test maintains 90% sensitivity but improves specificity to 95%. Since sensitivity relates to correctly identifying true positives (diabetic patients), this population remains unchanged. However, specificity measures the test's ability to correctly identify true negatives (non-diabetic individuals). With 95% specificity, the test now correctly identifies 200×0.95=190200 \times 0.95 = 190 non-diabetic individuals as negative, leaving 200−190=10200 - 190 = 10 false positives. Answer A correctly calculates this decrease from 30 to 10 false positives. Answer B (decrease to 15) represents a miscalculation, perhaps confusing the improvement magnitude. Answer C incorrectly assumes false positives depend on sensitivity rather than specificity - this reflects a fundamental misunderstanding since false positives occur among non-diseased individuals and are determined by specificity. Answer D (decrease to 20) suggests only 90% specificity, not the stated 95%. Remember this key relationship: false positives = (total non-diseased) × (1 - specificity). When specificity improves, false positives always decrease proportionally, regardless of sensitivity changes. Always identify which population (diseased vs. non-diseased) is affected by the parameter being modified.

Question 4

An emergency department implements a rapid cardiac biomarker test with 88% sensitivity and 82% specificity for detecting myocardial infarction. In their patient population, 12% of chest pain presentations are actual MIs. A physician argues that because the test misses 12% of MI cases, it's too dangerous to use. What additional information would best address this concern?

  1. The negative predictive value of the test, which indicates the probability that chest pain patients with negative tests truly don't have MI (correct answer)
  2. The positive predictive value of the test, which indicates the probability that patients with positive tests actually have MI
  3. A comparison of the test's likelihood ratios for positive and negative results to determine diagnostic value
  4. The overall accuracy of the test calculated as the percentage of all correct classifications
  5. The number needed to screen to prevent one missed MI diagnosis in this population
Explanation: When evaluating diagnostic tests, you need to match the clinical question to the right statistical measure. Here, the physician worries about missing MI cases (false negatives), so you need to know: "If the test is negative, how confident can I be that the patient doesn't have MI?" Negative predictive value (NPV) directly answers this concern. NPV tells you the probability that a patient with a negative test truly doesn't have the disease. With 88% sensitivity, 82% specificity, and 12% prevalence, the NPV calculates to about 96%. This means when the test is negative, there's only a 4% chance the patient actually has MI - much more reassuring than the crude "12% miss rate" the physician cited. Option B (positive predictive value) tells you the probability that positive results are true positives, but doesn't address the concern about missed cases. Option C (likelihood ratios) provides useful diagnostic information but requires more complex interpretation and doesn't directly quantify the risk of false negatives in terms physicians typically use for clinical decisions. Option D (overall accuracy) gives a general performance measure but doesn't specifically address what happens when tests are negative. The key insight is that sensitivity alone can be misleading. A test with 88% sensitivity sounds like it misses 12% of cases, but when disease prevalence is low (12%), most negative results are true negatives. NPV incorporates both test performance and disease prevalence to give you the clinically relevant answer. Study tip: Match the clinical concern to the right measure - NPV for "How much do I trust negative results?" and PPV for "How much do I trust positive results?"

Question 5

Two screening tests for the same condition are being compared. Test A has sensitivity 80% and specificity 90%. Test B has sensitivity 95% and specificity 75%. If both tests are used in sequence (positive on Test A followed by Test B), what is the combined sensitivity of this sequential testing approach?

  1. 76% (correct answer)
  2. 95%
  3. 88%
  4. 99%
  5. 80%
Explanation: When you encounter questions about sequential screening tests, you need to understand that the combined sensitivity depends on how the tests are arranged. Sequential testing means a patient must test positive on the first test to even receive the second test. For sequential testing, the combined sensitivity equals the product of individual sensitivities: Combined Sensitivity=SensitivityA×SensitivityB\text{Combined Sensitivity} = \text{Sensitivity}_A \times \text{Sensitivity}_B This is because both tests must detect the disease for a positive result. If Test A misses 20% of cases (sensitivity = 80%) and Test B misses 5% of cases (sensitivity = 95%), then the sequential approach will miss cases that either test fails to detect. Calculate: 0.80×0.95=0.76=76%0.80 \times 0.95 = 0.76 = 76\% Answer A (76%) is correct because it represents the probability that both tests will successfully detect disease when present. Answer B (95%) incorrectly assumes the combined sensitivity equals the higher individual sensitivity, ignoring that Test A's limitations affect the overall performance. Answer C (88%) might result from incorrectly averaging the two sensitivities (80%+95%)/2=87.5%(80\% + 95\%)/2 = 87.5\%, which doesn't reflect how sequential testing actually works. Answer D (99%) could come from mistakenly adding the sensitivities together, which would be mathematically impossible since sensitivity cannot exceed 100%. Remember: Sequential testing always reduces sensitivity compared to individual tests because you're multiplying probabilities less than 1. However, it typically improves specificity. This trade-off is crucial in understanding diagnostic test combinations.

Question 6

In a screening program, Test X correctly identifies 72 out of 90 diseased individuals and correctly identifies 420 out of 500 non-diseased individuals. Test Y correctly identifies 81 out of 90 diseased individuals but correctly identifies only 350 out of 500 non-diseased individuals. Which statement best describes the trade-off between these tests?

  1. Test Y has 10% higher sensitivity but 14% lower specificity than Test X (correct answer)
  2. Test Y has 9% higher sensitivity but 15% lower specificity than Test X
  3. Test Y has 12.5% higher sensitivity but 16.8% lower specificity than Test X
  4. Test Y has 10% higher sensitivity but 12% lower specificity than Test X
  5. Test Y has 11% higher sensitivity but 14% lower specificity than Test X
Explanation: When evaluating diagnostic tests, you need to calculate sensitivity (ability to correctly identify diseased individuals) and specificity (ability to correctly identify non-diseased individuals), then compare the differences between tests. For Test X: Sensitivity = 72/90 = 80%; Specificity = 420/500 = 84% For Test Y: Sensitivity = 81/90 = 90%; Specificity = 350/500 = 70% Test Y has 90% - 80% = 10% higher sensitivity and 70% - 84% = -14% (or 14% lower) specificity than Test X. This represents the classic diagnostic trade-off: improving sensitivity often comes at the cost of specificity. Choice A correctly states that Test Y has 10% higher sensitivity but 14% lower specificity than Test X. Choice B miscalculates the sensitivity difference as 9% instead of 10% and the specificity difference as 15% instead of 14%. Choice C uses relative changes rather than absolute differences, calculating (81-72)/72 ≈ 12.5% and (420-350)/420 ≈ 16.7%, which isn't the standard way to compare test performance. Choice D correctly identifies the 10% sensitivity difference but incorrectly calculates the specificity difference as 12% instead of 14%. Remember that test comparisons typically use absolute percentage point differences, not relative percentage changes. Always calculate both sensitivity and specificity carefully, as screening questions often test your ability to recognize the fundamental trade-off between these two performance measures.

Question 7

A study evaluates a biomarker test in two different populations. In Population A (high disease prevalence), the test shows sensitivity 85% and specificity 80%. In Population B (low disease prevalence), the same test shows sensitivity 85% and specificity 80%. A researcher claims that because the test characteristics are identical, the clinical utility is the same in both populations. What is the primary flaw in this reasoning?

  1. Positive and negative predictive values depend on disease prevalence, not just sensitivity and specificity (correct answer)
  2. Sensitivity and specificity must change when prevalence changes due to spectrum bias effects
  3. The test's analytical accuracy decreases in populations with different prevalence rates
  4. Different populations require different cutoff values to maintain equivalent test performance
  5. The false positive and false negative rates vary inversely with population prevalence levels
Explanation: When you encounter questions about diagnostic test performance across different populations, the key concept to consider is how prevalence affects predictive values, even when test characteristics remain constant. The researcher's reasoning is flawed because positive predictive value (PPV) and negative predictive value (NPV) are directly influenced by disease prevalence, regardless of identical sensitivity and specificity. PPV represents the probability that a positive test result indicates true disease, calculated as: PPV=Sensitivity × PrevalenceSensitivity × Prevalence + (1-Specificity) × (1-Prevalence)PPV = \frac{\text{Sensitivity × Prevalence}}{\text{Sensitivity × Prevalence + (1-Specificity) × (1-Prevalence)}} In Population A (high prevalence), more positive results will be true positives, yielding higher PPV. In Population B (low prevalence), more positive results will be false positives, dramatically lowering PPV. The clinical utility differs significantly because the same positive result has different meanings in each population. Option A correctly identifies this fundamental principle. Option B incorrectly suggests sensitivity and specificity must change with prevalence - while spectrum bias can occur, it's not inevitable and isn't what the question describes. Option C wrongly implies analytical accuracy (the test's ability to measure what it's designed to measure) varies with prevalence, but analytical performance is independent of population characteristics. Option D incorrectly suggests different cutoff values are needed, but the question states the test characteristics are identical. Remember: sensitivity and specificity are intrinsic test properties, while predictive values depend on both test performance and disease prevalence. Always consider prevalence when evaluating a test's clinical utility, not just its sensitivity and specificity.

Question 8

A new point-of-care test was evaluated against the gold standard in a clinic setting. The results showed: diseased patients testing positive = 38, diseased patients testing negative = 12, non-diseased patients testing positive = 25, non-diseased patients testing negative = 125. If this test were used in a population where disease prevalence is twice as high as in the clinic study, how would this affect the test's sensitivity?

  1. Sensitivity would remain 76% regardless of prevalence changes (correct answer)
  2. Sensitivity would increase proportionally to 152% due to higher prevalence
  3. Sensitivity would decrease to approximately 38% due to spectrum bias
  4. Sensitivity would need to be recalculated based on the new prevalence rate
  5. Sensitivity would improve to 88% because of increased positive predictive value
Explanation: When you encounter diagnostic test questions, remember that sensitivity and specificity are intrinsic properties of the test itself—they measure how the test performs in diseased and non-diseased populations, respectively, regardless of how common the disease is. Let's calculate the sensitivity from the clinic data. Sensitivity measures the test's ability to correctly identify diseased patients: Sensitivity=True PositivesTrue Positives + False Negatives=3838+12=3850=76%\text{Sensitivity} = \frac{\text{True Positives}}{\text{True Positives + False Negatives}} = \frac{38}{38 + 12} = \frac{38}{50} = 76\% This 76% sensitivity reflects the test's inherent ability to detect disease and won't change based on prevalence. Whether you test this in a population with 1% disease prevalence or 50% prevalence, the test will still correctly identify 76% of diseased individuals. Answer A is correct because sensitivity remains constant regardless of prevalence changes. Answer B incorrectly suggests sensitivity can exceed 100% and changes with prevalence—this confuses sensitivity with predictive values. Answer C mentions spectrum bias, which occurs when test performance differs between populations due to disease severity differences, not prevalence alone. The question doesn't indicate different disease severity, just higher prevalence. Answer D incorrectly implies that prevalence affects the calculation of sensitivity, when prevalence only affects predictive values (positive and negative predictive values). Study tip: Remember the key distinction—sensitivity and specificity are test characteristics that don't change with prevalence, while predictive values do change with prevalence. When you see prevalence changes in questions, immediately ask whether they're testing intrinsic test properties or predictive values.

Question 9

A laboratory technician accidentally switches the labels on positive and negative control samples when calibrating a diagnostic test. As a result, the test's output is inverted: truly positive samples read negative and truly negative samples read positive. If the original test had sensitivity 85% and specificity 90%, what are the apparent sensitivity and specificity of the incorrectly calibrated test?

  1. Sensitivity 10%, Specificity 15%
  2. Sensitivity 90%, Specificity 85%
  3. Sensitivity 15%, Specificity 10% (correct answer)
  4. Sensitivity 0%, Specificity 0%
  5. Sensitivity -85%, Specificity -90%
Explanation: When you encounter questions about diagnostic test errors or calibration problems, focus on how the error affects the fundamental measurements of sensitivity and specificity. Let's trace through what happens when controls are switched. The original test correctly identifies 85% of truly positive samples (sensitivity = 85%) and 90% of truly negative samples (specificity = 90%). However, with inverted calibration, the test now calls positive samples "negative" and negative samples "positive." For the miscalibrated test:
  • Apparent sensitivity: Of truly positive samples, only 15% will now read "positive" (since 85% are incorrectly reading "negative"). Apparent sensitivity = 100% - 85% = 15%
  • Apparent specificity: Of truly negative samples, only 10% will now read "negative" (since 90% are incorrectly reading "positive"). Apparent specificity = 100% - 90% = 10%
Answer choice A (10%, 15%) reverses these values—a common mistake when rushing through the logic. Answer choice B (90%, 85%) incorrectly assumes the values simply flip positions without considering the inversion effect. Answer choice D (0%, 0%) would only occur if the test became completely non-functional, which isn't the case here—it's still detecting, just reporting opposite results. Study tip: For diagnostic test inversion problems, remember that apparent sensitivity becomes (100% - original sensitivity) and apparent specificity becomes (100% - original specificity). The better the original test, the worse it appears when inverted.

Question 10

A meta-analysis combines results from five studies of the same diagnostic test. The individual study sensitivities are: 82%, 78%, 85%, 80%, and 75%. The individual study specificities are: 88%, 92%, 85%, 90%, and 95%. If a simple average is used to pool the results, what concern should be raised about this approach?

  1. Different study sample sizes and disease prevalences are not accounted for in simple averaging (correct answer)
  2. The average sensitivity of 80% is too low for clinical utility assessment purposes
  3. The range of specificity values (85-95%) indicates the test is unreliable across studies
  4. Simple averaging violates the assumption that sensitivity and specificity are correlated measures
  5. The meta-analysis should only include studies with identical sensitivity and specificity values
Explanation: When conducting meta-analysis of diagnostic test accuracy, you're essentially trying to synthesize evidence from multiple studies to get a more precise estimate of test performance. The key challenge is that different studies have different characteristics that affect how much weight their results should receive in the final pooled estimate. Option A correctly identifies the fundamental flaw with simple averaging in meta-analysis. Studies with larger sample sizes provide more precise estimates and should contribute more to the pooled result than smaller studies. Additionally, studies with different disease prevalences may have systematic differences in their sensitivity and specificity estimates due to spectrum effects or patient selection. Simple averaging treats a study of 50 patients the same as a study of 500 patients, which ignores important information about precision and reliability. Option B misses the point entirely - while 80% sensitivity might seem low, the question isn't asking about clinical utility thresholds. Option C incorrectly interprets the specificity range (85-95%) as indicating unreliability, when this variation is actually expected and normal across different populations and settings. Option D contains a statistical misconception - there's no assumption that sensitivity and specificity should be correlated, and simple averaging doesn't violate any such principle. Study tip: In meta-analysis questions, always look for issues related to study weighting and heterogeneity. Simple averaging is almost never the best approach because it ignores study quality indicators like sample size, precision, and population differences. Proper meta-analysis uses weighted averages and accounts for between-study variation.

Question 11

A study evaluates a new mammography interpretation algorithm by having it analyze 500 mammograms: 125 from women with breast cancer and 375 from women without cancer. The algorithm correctly identifies 100 cancer cases and incorrectly flags 75 non-cancer cases. If the algorithm's decision threshold is adjusted to reduce false positives by 60%, and this change decreases true positives proportionally by 30%, what would be the new specificity?

  1. 88%
  2. 92% (correct answer)
  3. 85%
  4. 90%
  5. 95%
Explanation: When you encounter diagnostic test performance questions involving threshold adjustments, you need to carefully track how changes affect each component of the confusion matrix and then recalculate the desired metric. Start by identifying the initial conditions: 100 true positives (TP), 25 true negatives (TN = 375 - 75), 75 false positives (FP), and 25 false negatives (FN = 125 - 100). The original specificity is TNTN+FP=300375=80%\frac{TN}{TN + FP} = \frac{300}{375} = 80\%. After the threshold adjustment, false positives decrease by 60%: 75×0.40=3075 \times 0.40 = 30 new FP. Since the total number of non-cancer cases stays constant at 375, true negatives increase to 375−30=345375 - 30 = 345. The new specificity becomes 345375=92%\frac{345}{375} = 92\%. Looking at the wrong answers: A) 88% might result from incorrectly calculating the proportional changes or mixing up the percentage reductions. C) 85% could come from applying the wrong percentage change or confusing sensitivity with specificity calculations. D) 90% might result from rounding errors or incorrectly applying the 30% TP reduction (which doesn't affect specificity) to the specificity calculation itself. The correct answer is B) 92%. Remember that specificity only depends on the non-cancer cases (true negatives and false positives). When analyzing threshold changes, carefully track which metrics are directly affected—changes in true positive rates don't impact specificity calculations, only the false positive changes matter here.

Question 12

A genetic test for a hereditary condition shows 99% sensitivity and 98% specificity in initial validation studies. However, when implemented in a clinical genetics practice, the apparent sensitivity appears lower because some patients with positive family histories test negative despite having clinical symptoms. What is the most likely explanation for this discrepancy?

  1. Genetic heterogeneity means the test only detects certain mutations, not all causes of the clinical condition (correct answer)
  2. The clinical population has a different disease prevalence than the validation study population
  3. Laboratory processing errors are more common in clinical practice than in research studies
  4. The test's analytical sensitivity decreases when sample throughput increases in clinical labs
  5. Clinical symptoms are not reliable indicators of genetic disease status for validation purposes
Explanation: When you encounter questions about test performance discrepancies between validation and clinical settings, focus on what could cause the same test to miss cases it should theoretically detect. The key insight here is understanding genetic heterogeneity - the principle that identical clinical presentations can result from different underlying genetic causes. The original validation study likely used samples from patients with known mutations that the test was designed to detect, achieving 99% sensitivity for those specific genetic variants. However, in real clinical practice, patients presenting with the same hereditary condition might have different causative mutations, some of which the test cannot detect. This explains why patients with positive family histories and clinical symptoms test negative - they truly have the condition, but caused by genetic variants not covered by this particular test. The test isn't failing; it's simply encountering genetic diversity it wasn't designed to capture. Option B is incorrect because prevalence affects positive predictive value, not sensitivity. A test's ability to detect true cases shouldn't change based on how common the disease is in the population. Option C, while laboratory errors do occur, wouldn't systematically affect sensitivity in the specific pattern described (positive family history + clinical symptoms + negative test). Option D is wrong because analytical sensitivity refers to the test's ability to detect low concentrations of the target, which wouldn't change with throughput and doesn't explain the clinical pattern. Remember: when sensitivity appears lower in clinical practice despite maintained specificity, suspect that the test covers only a subset of the genetic causes of the condition.

Question 13

A rapid diagnostic test shows the following results in a validation study: 45 true positives, 15 false positives, 5 false negatives, and 135 true negatives. If the cutoff value for the test is lowered to increase sensitivity, which of the following changes is most likely to occur?

  1. Specificity will decrease and some true negatives will become false positives (correct answer)
  2. Specificity will increase and some false positives will become true negatives
  3. Sensitivity will decrease and some false negatives will become true negatives
  4. Both sensitivity and specificity will increase proportionally to maintain accuracy
  5. Prevalence will change and affect the positive predictive value calculation
Explanation: When you encounter diagnostic test questions involving sensitivity and specificity, remember that these measures have an inverse relationship - changing the cutoff to improve one typically worsens the other. Let's set up the 2×2 table from the given data: 45 true positives, 15 false positives, 5 false negatives, and 135 true negatives. Current sensitivity is 45/(45+5) = 90%, and specificity is 135/(135+15) = 90%. Lowering the cutoff means the test becomes more liberal in calling results "positive." This captures more true cases (increasing sensitivity) but also incorrectly identifies more healthy individuals as positive. Some people who were previously classified as true negatives (correctly identified as healthy) will now test positive and become false positives. Simultaneously, some false negatives (missed cases) become true positives. While sensitivity increases, specificity decreases because the denominator (true negatives + false positives) shifts - fewer true negatives and more false positives. Answer A correctly describes this trade-off: specificity decreases as true negatives convert to false positives. Answer B incorrectly suggests specificity increases, which contradicts the inverse relationship. Answer C wrongly states sensitivity decreases - lowering the cutoff actually increases sensitivity. Answer D is impossible because both measures cannot increase simultaneously when changing a single cutoff; this represents a fundamental misunderstanding of the sensitivity-specificity trade-off. Study tip: Remember the seesaw principle - in diagnostic testing, improving sensitivity typically decreases specificity and vice versa. Visualize how moving the cutoff shifts cases between the four categories of your 2×2 table.

Question 14

Two diagnostic tests are performed independently on the same patient. Test 1 has sensitivity 80% and specificity 95%. Test 2 has sensitivity 90% and specificity 85%. If both tests give positive results and the tests are assumed to be independent, what can be concluded about the combined specificity of this parallel testing approach?

  1. The combined specificity is 80.8%, which is lower than either individual test's specificity (correct answer)
  2. The combined specificity equals the sum of individual specificities (180%)
  3. The combined specificity equals the higher of the two specificities (95%)
  4. The combined specificity is the average of individual specificities (90%)
  5. The combined specificity cannot be determined without knowing disease prevalence
Explanation: When you encounter questions about combining diagnostic tests, you need to understand how sensitivity and specificity change when tests are used together. In parallel testing (where a positive result from either test is considered positive), specificity behaves differently than you might expect. To find combined specificity, you need to calculate the probability that both tests are negative when disease is absent. Since the tests are independent, combined specificity = Specificity₁ × Specificity₂ = 0.95 × 0.85 = 0.808 or 80.8%. This is indeed lower than either individual test's specificity, which is a key characteristic of parallel testing - you gain sensitivity but lose specificity. Answer A correctly identifies this calculation and the crucial insight that combined specificity decreases in parallel testing. Answer B incorrectly suggests you simply add the specificities, which would give an impossible result over 100% and ignores the fundamental probability rules. Answer C assumes you take the higher specificity, which would be true if you only used the better test, but not when combining results. Answer D suggests averaging the specificities, which has no basis in probability theory and doesn't account for how the tests actually interact. Remember this pattern: parallel testing (positive if either test positive) increases overall sensitivity but decreases specificity, while series testing (positive only if both tests positive) increases specificity but decreases sensitivity. The mathematics always involves multiplication of the relevant probabilities, never simple addition or averaging.

Question 15

A new screening test for diabetes was evaluated in a population of 1,000 individuals. The test correctly identified 180 of the 200 individuals who actually had diabetes. Among the 800 individuals without diabetes, the test incorrectly indicated diabetes in 80 individuals. If the test result is positive in a randomly selected individual from this population, what is the probability that the individual actually has diabetes?

  1. 0.69 (correct answer)
  2. 0.90
  3. 0.80
  4. 0.10
  5. 0.18
Explanation: When you encounter a screening test question asking for the probability of disease given a positive test, you're dealing with positive predictive value (PPV). This requires understanding the difference between sensitivity/specificity and predictive values. Let's construct a 2×2 table from the given data. Of 200 people with diabetes, 180 tested positive (true positives) and 20 tested negative (false negatives). Of 800 people without diabetes, 80 tested positive (false positives) and 720 tested negative (true negatives). To find PPV, you need: PPV=True PositivesAll Positive Tests=180180+80=180260=0.69PPV = \frac{\text{True Positives}}{\text{All Positive Tests}} = \frac{180}{180 + 80} = \frac{180}{260} = 0.69 This means that when someone tests positive, there's a 69% chance they actually have diabetes. Now for the distractors: Answer B (0.90) represents the test's sensitivity - the probability of testing positive given you have diabetes (180/200). This is a classic trap because sensitivity seems relevant but answers the opposite question. Answer C (0.80) might tempt you as it represents the specificity calculation backward (720/800 = 0.90, so 1-0.90 = 0.10... though this doesn't directly give 0.80). Answer D (0.10) represents the false positive rate among healthy individuals (80/800). Remember: sensitivity and specificity describe test performance, but predictive values tell you what a test result means for an individual patient. Always set up the 2×2 table to avoid mixing up these concepts - it's the most reliable way to get predictive value questions right.

Question 16

A screening test has been optimized to achieve 90% sensitivity. In a population of 10,000 where disease prevalence is 2%, the test produces 200 positive results. What is the specificity of this test?

  1. 89.8% (correct answer)
  2. 90.0%
  3. 80.0%
  4. 95.0%
  5. 85.2%
Explanation: When you encounter screening test problems, you need to work systematically through the 2×2 contingency table using the given information to find unknown values. Given: 90% sensitivity, 2% prevalence in 10,000 people, and 200 total positive results. First, calculate the disease distribution: 10,000 × 0.02 = 200 people have the disease, while 9,800 are disease-free. With 90% sensitivity, the test correctly identifies 200 × 0.90 = 180 diseased individuals as positive (true positives), meaning 20 diseased individuals test negative (false negatives). Since there are 200 total positive results and 180 are true positives, the remaining 20 positive results must be false positives (disease-free individuals testing positive). Therefore, 9,800 - 20 = 9,780 disease-free individuals test negative (true negatives). Specificity = True negatives ÷ Total disease-free = 9,780 ÷ 9,800 = 0.898 = 89.8% Answer A (89.8%) is correct. Answer B (90.0%) represents the sensitivity value, not specificity—a common mistake when students confuse these measures. Answer C (80.0%) and Answer D (95.0%) are arbitrary values that don't correspond to any calculation from the given data. The key strategy here is constructing the complete 2×2 table methodically. Start with what you know (prevalence gives you disease totals, sensitivity gives you the true positive rate), then use the constraint that all positive results must sum to the given total. This systematic approach prevents errors and reveals the specificity calculation.

Question 17

A screening test for HIV has a sensitivity of 99.5% and specificity of 99.8%. In a population where the prevalence of HIV is 0.1%, what is the probability that a person who tests negative actually has HIV?

  1. Approximately 0.05% (correct answer)
  2. Approximately 0.5%
  3. Approximately 5.0%
  4. Approximately 0.005%
Explanation: This asks for 1 - NPV. In a population of 100,000: True positives = 1000 × 0.995 = 995; False negatives = 1000 × 0.005 = 5; True negatives = 99,000 × 0.998 = 98,802; False positives = 99,000 × 0.002 = 198. Total negative tests = 98,802 + 5 = 98,807. Probability of having HIV given negative test = 5/98,807 ≈ 0.05%.

Question 18

A researcher reports that a new biomarker has 'excellent diagnostic accuracy' with sensitivity of 92% and specificity of 94%. However, when implemented in a real-world screening program where disease prevalence is 2%, the positive predictive value is disappointing. What is the most likely reason for this discrepancy?

  1. High specificity becomes less important than sensitivity when prevalence is very low, leading to poor PPV despite good individual test characteristics
  2. The researcher failed to account for spectrum bias in the original validation study population compared to the screening population
  3. Even with good sensitivity and specificity, low prevalence means false positives greatly outnumber true positives, resulting in poor PPV (correct answer)
  4. The biomarker performs differently in asymptomatic screening populations compared to symptomatic diagnostic populations due to lead-time bias
Explanation: With 2% prevalence, sensitivity 92%, specificity 94%: In 10,000 people, TP=184, FP=588, so PPV=184/772≈24%. The fundamental issue is that in low-prevalence populations, even small false positive rates generate many false positives relative to true positives. Option A incorrectly suggests specificity is less important (it's actually more critical). Option B mentions spectrum bias but doesn't explain the PPV issue. Option D incorrectly invokes lead-time bias.

Question 19

Two radiologists independently interpret the same set of mammograms for breast cancer detection. Radiologist A has a sensitivity of 85% and specificity of 92%. Radiologist B has a sensitivity of 90% and specificity of 88%. If a patient has breast cancer, what is the probability that at least one radiologist will detect it, assuming their interpretations are independent?

  1. 76.5%
  2. 87.5%
  3. 95.0%
  4. 98.5% (correct answer)
Explanation: When you encounter probability questions about multiple independent tests or observers, think about whether you need the probability that "at least one" will succeed. This typically requires using the complement rule: P(at least one success) = 1 - P(all fail). Since the radiologists interpret independently, you can calculate the probability that both miss the cancer, then subtract from 1. Radiologist A's sensitivity is 85%, so she misses cancer 15% of the time. Radiologist B's sensitivity is 90%, so he misses cancer 10% of the time. The probability both miss the cancer is: 0.15×0.10=0.015=1.5%0.15 \times 0.10 = 0.015 = 1.5\% Therefore, the probability at least one detects it is: 1−0.015=0.985=98.5%1 - 0.015 = 0.985 = 98.5\% Looking at the wrong answers: A) 76.5% appears to multiply the sensitivities incorrectly (possibly 0.85 × 0.90 = 76.5%), but this doesn't represent any meaningful probability in this context. B) 87.5% might result from averaging the two sensitivities, which ignores the independence and "at least one" requirement. C) 95.0% could come from adding the miss rates (15% + 10% = 25%, then 100% - 25% = 75%) or other flawed reasoning, but doesn't follow proper probability rules for independent events. Remember: for "at least one" probability questions with independent events, always use the complement rule. Calculate the probability that all events fail, then subtract from 1. This approach is much simpler than trying to enumerate all success scenarios.

Question 20

In a meta-analysis of diagnostic studies for a cardiac biomarker, researchers found significant heterogeneity in reported sensitivities (range: 74%-96%) but consistent specificities (range: 89%-92%). What is the most likely explanation for this pattern?

  1. Different analytical platforms and cutoff values were used across studies, primarily affecting the test's ability to detect true positives
  2. Spectrum bias affected studies differently, with some including more severe cases that are easier to detect, inflating sensitivity estimates (correct answer)
  3. Reference standard bias was present, where some studies used less accurate gold standards that missed borderline positive cases
  4. Verification bias occurred when some studies only confirmed negative results in a subset of patients, artificially lowering sensitivity
Explanation: When you encounter meta-analysis questions showing heterogeneity in diagnostic test performance, focus on what factors could selectively affect sensitivity versus specificity. Sensitivity measures the test's ability to correctly identify diseased patients, while specificity measures its ability to correctly identify healthy patients. The pattern here—variable sensitivity with consistent specificity—points to spectrum bias. This occurs when studies include different populations with varying disease severity. More severe cardiac cases produce higher biomarker levels that are easier to detect, inflating sensitivity estimates. Meanwhile, the healthy control populations remain similar across studies, keeping specificity stable. This creates the exact pattern described: sensitivity ranging widely (74%-96%) while specificity stays consistent (89%-92%). Option A is incorrect because different platforms and cutoffs would affect both sensitivity and specificity, not just sensitivity alone. If analytical differences were the primary issue, you'd expect to see heterogeneity in both measures. Option C misidentifies the mechanism. Reference standard bias would typically affect both sensitivity and specificity measurements, as an inaccurate gold standard would misclassify both positive and negative cases. Option D describes verification bias incorrectly. The scenario states that negative results were selectively confirmed, but verification bias affecting sensitivity usually involves selective confirmation of positive results, not negative ones. Study tip: In diagnostic test questions, remember that spectrum bias specifically affects sensitivity because disease severity influences test performance, while the "normal" population remains relatively constant across studies. Look for this sensitivity-only heterogeneity pattern as a key clue.