All articles

Medical school predictive-value workflow

Anki for Predictive Values: Stop Confusing PPV with Sensitivity

Confuse PPV with sensitivity on medical school questions? Use worked prevalence examples, focused Anki cards, and checked AnkiQuiz practice to fix the mix-up.

You can finish the Anki cloze “sensitivity = TP / (TP + FN)” instantly. Then a practice question asks what a positive result means in a low-prevalence population, and you choose the sensitivity as the answer. The formula survived review; the direction of the question did not.

To study predictive values with Anki, name the group in the denominator, reconstruct a complete two-by-two table, and compare what changes when prevalence changes. Make cards for the specific confusion, then use a small, checked AnkiQuiz set to test the same distinction in different wording.

Disclosure: We build AnkiQuiz and are not affiliated with Anki, AnKing, UWorld, NBME, or USMLE. This is an editorial study workflow. The numbers below are invented teaching data, not performance estimates for a clinical test or guidance for patient care.

Use the medical-school Anki workflow for your daily routine and the biostatistics Anki guide for broader calculation and interpretation problems. This article focuses on the positive-result question and its dependence on the population.

Start with the denominator, then retrieve the formula

Before calculating, finish the sentence “Among people who __, what fraction __?” This reveals which group the question starts with.

MeasureStarting groupOutcome being countedFormula
SensitivityHave the conditionTest positiveTP / (TP + FN)
SpecificityDo not have the conditionTest negativeTN / (TN + FP)
Positive predictive value (PPV)Test positiveHave the conditionTP / (TP + FP)
Negative predictive value (NPV)Test negativeDo not have the conditionTN / (TN + FN)

TP means true positive, FP false positive, FN false negative, and TN true negative. The FDA diagnostic-test guidance defines these measures in relation to an appropriate reference standard and intended population. Its statistical definitions support this distinction; it does not evaluate this Anki workflow.

Sensitivity and PPV share a numerator, but their denominators answer different questions. A card asking only for the letters can hide that confusion. Ask yourself to name the starting group before revealing the answer.

Worked example: the same test assumptions, different PPV

Assume a fictional test has 80% sensitivity and 90% specificity in each of two hypothetical populations. Each population contains 1,000 people. For this exercise, condition status is known from the reference standard, everyone receives one test, and results are either positive or negative.

Population A: 10% prevalence

There are 100 people with the condition and 900 without it. Applying the assumed rates gives:

Test resultCondition presentCondition absentTotal
Positive80 true positives90 false positives170
Negative20 false negatives810 true negatives830
Total1009001,000

PPV = 80 / (80 + 90) = 47.1%, rounded. NPV = 810 / (810 + 20) = 97.6%, rounded. Sensitivity remains 80 / 100 = 80%; specificity remains 810 / 900 = 90%.

State the interpretation: among positive results in this hypothetical population, about 47% come from people who have the condition. “80% sensitive” answers a different question.

Population B: 1% prevalence

Now there are 10 people with the condition and 990 without it. Keep the same sensitivity and specificity assumptions:

Test resultCondition presentCondition absentTotal
Positive8 true positives99 false positives107
Negative2 false negatives891 true negatives893
Total109901,000

PPV = 8 / (8 + 99) = 7.5%, rounded. NPV = 891 / (891 + 2) = 99.8%, rounded. Sensitivity is still 8 / 10 = 80%; specificity is still 891 / 990 = 90%.

The larger group without the condition contributes more false positives, while the smaller group with the condition contributes fewer true positives. That changes the composition of positive results. The low PPV does not mean the test's specificity became 7.5%.

Keep the assumption attached to the lesson

In this model, lower prevalence reduces PPV and increases NPV while sensitivity and specificity are held fixed. We imposed those fixed rates. Real populations may differ in disease spectrum, test use, and other characteristics; do not assume measured performance transfers unchanged. The FDA guidance emphasizes specifying the intended population and reference standard when interpreting estimates.

A population-level teaching calculation also does not by itself establish an individual patient's probability or a treatment decision. Keep your cards scoped to what the supplied data support.

Repair the card that caused the mistake

The Anki manual's effective-learning guidance recommends manageable prompts and understanding the material before memorization. Our application is to give each card one clear retrieval target, with enough context to make its answer meaningful.

Choose the repair that matches your error:

  • Reversed conditioning: Front: “Among positive test results, what fraction have the condition? Name the measure and denominator.” Back: “PPV; all positive results, TP + FP.”
  • Formula without meaning: Front: “Sensitivity and PPV both use TP. Which starting groups distinguish them?” Back: “All people with the condition versus all people who test positive.”
  • Missing base rate: Front: “In the fictional 1%-prevalence population, TP = 8 and FP = 99. What is PPV, and what does it describe?” Back: “8 / 107, about 7.5%; the fraction of positive results with the condition in that population.”
  • Unqualified prevalence rule: Front: “In the fictional comparison, what was held fixed when prevalence changed?” Back: “Sensitivity at 80% and specificity at 90%, with the same binary test model. The comparison does not prove identical performance in real populations.”

Keep the complete example and source in reference fields. If the card asks you to reconstruct the table, supply the population size, prevalence, sensitivity, specificity, and assumptions on the front. Do not ask for a numerical PPV from sensitivity alone.

You do not need every card above. Repair one observed mistake, then test it with a new example. The medical Anki card-writing guide covers focused prompts and avoiding duplicate reviews.

Use AnkiQuiz for a checked change of wording

AnkiQuiz for medical students can generate supplemental questions from selected card text. Here, its useful role is to vary the wording of a checked definition, interpretation, or comparison.

  1. Open Anki Quiz from Anki's Tools menu and select a relevant existing regular deck. Parent decks include subdecks. Tags and Browse selections do not define input; filtered decks are unavailable, and suspended or buried cards are excluded.
  2. Select text fields containing the full definitions, labeled counts, assumptions, answers, and explanations. Check the time range and sampled coverage. A screenshot of a table or an external source link alone does not supply the required text.
  3. Start with a small single-choice quiz. Independently verify every stem, option, answer, calculation, and explanation. Reject items that omit the prevalence needed for a calculation, confuse sensitivity with PPV, or silently change the assumptions.
  4. When answering, name the denominator before reading the choices. Finish with a fresh authorized course or question-bank problem to see whether the distinction transfers.

If your plan supports custom instructions, try:

Use only the supplied definitions and complete numerical examples. Ask which population each measure conditions on, distinguish sensitivity from PPV, and compare the supplied prevalence scenarios. Preserve all assumptions and labels. Show the numerator, denominator, calculation, rounding, and interpretation in each explanation. Do not invent clinical test-performance estimates or infer an individual's diagnosis.

Instructions do not guarantee correct arithmetic or reasoning. Selected text is sent for generation; use permitted material without identifiable patient information or restricted question-bank content. AnkiQuiz does not interpret table images, validate diagnostic performance, edit your cards, or change Anki scheduling. See the installation guide and card-to-quiz walkthrough for setup.

Check transfer with a fresh denominator

In your next legitimate practice problem, write the starting group, numerator, and denominator before calculating. If the question supplies rates and prevalence, rebuild the counts. If it supplies a table, follow the labels even when the rows and columns are swapped.

Check both the number and its interpretation against the authorized explanation. A correct percentage with the wrong starting population is still a misunderstanding. If you can calculate PPV but cannot explain the prevalence comparison, revisit the two populations rather than adding more formula clozes.

Use fresh representative practice to assess progress. An AnkiQuiz score describes those generated items; it does not establish exam readiness or clinical competence. Protect your question-bank time and keep repairs small.

Frequently asked questions

How do I study positive predictive value with Anki?

Learn the population before the formula: PPV asks what fraction of positive test results are true positives. Use a complete labeled example, explain the denominator, and compare populations with different prevalence under explicitly fixed sensitivity and specificity.

What is the difference between sensitivity and positive predictive value?

Sensitivity starts with people who have the condition and asks how many test positive: TP / (TP + FN). PPV starts with positive test results and asks how many have the condition: TP / (TP + FP). Reversing the conditioning changes the question.

Why does lower prevalence reduce PPV?

When sensitivity and specificity are held fixed and informative in a teaching model, lower prevalence leaves fewer true positives relative to false positives, so PPV falls. Rebuild the counts to explain the change; do not assume measured test performance transfers unchanged between real populations.

Can AnkiQuiz help me practice PPV and NPV?

AnkiQuiz can generate supplemental questions from complete, checked text fields containing definitions, counts, assumptions, and explanations. Verify every generated calculation and interpretation. It does not validate diagnostic tests or establish exam readiness.

Explore the medical school Anki guides.