AI Prompt Library for Nursing Students

Explain [topic, e.g. why DKA causes Kussmaul respirations] in 3 passes: first in one sentence a patient would understand, then at nursing-school depth, then walk the full mechanism step by step. After each pass, ask me one question to check I followed before continuing.

Forces layered depth instead of one wall of text, and the check questions make it retrieval, not reading.

I think I understand [topic]. Here is my explanation in my own words: [your explanation]. Find the weakest or wrong link in my reasoning, tell me why it’s wrong, and re-explain just that link.

Self-explanation plus targeted correction — you fix the one gap instead of re-learning the whole topic.

Build me a comparison table of [e.g. DKA vs HHS] with ONLY the rows that discriminate between them on an exam — skip everything they share. For each row, add one sentence on how a test question would try to trick me with it.

Confusable pairs are lost on the differences, not the definitions.

I got this practice question wrong. Question: [paste stem + options]. I chose [X]; the answer is [Y]. Don’t just explain why Y is right — explain what made X attractive, what cue in the stem I should have weighted more, and give me a rule for next time.

The learning lives in why YOUR distractor was tempting, not in the correct answer’s rationale.

Quiz me on [topic] one question at a time, but never tell me the answer. When I answer, respond only with follow-up questions that lead me to find my own errors. Keep going until I get it fully right, then summarize the reasoning path I should have taken.

Turns the assistant from an answer machine into a clinical instructor.

Give me 4 patients (one line each: age, diagnosis, one key finding). Ask me who I see first and why. Then critique my ranking using ABC, acute-vs-expected, and stable-vs-unstable logic — and tell me which finding I under-weighted.

Prioritization logic is conversational by nature — a genuinely great AI use case.

Here is my exam study guide: [paste]. Sort every topic into 3 buckets: (1) high-yield and commonly tested, (2) know the one key fact, (3) low-yield if time runs out. For bucket 1, list the specific thing exams usually ask about each topic.

Turns a flat topic list into a priority order — the decision students find hardest.

From these notes [paste], build a one-page night-before review: the 10 facts I’m most likely to blank on under pressure, each as a question on one line with the answer on the next. No paragraphs.

Question format even here — recognition reads, retrieval sticks.

Here are the last 10 practice questions I missed, with topics: [list]. What patterns do you see — content gaps, question-type traps, or reading errors? Give me the single highest-leverage fix for next week.

AI is good at spotting patterns — if you bring it the data. (Your practice tool should be keeping this list for you.)

Teach me the [drug class, e.g. beta-blocker] class using the prototype approach: pick one prototype drug, give me its mechanism in plain language, the 3 side effects that follow logically from that mechanism, and the nursing assessments that follow from those side effects. Then quiz me on a second drug in the class and make me predict everything from the mechanism.

Forces the mechanism → effect → nursing action chain instead of a flat fact list.

Quiz me on drug name suffixes one at a time (-olol, -pril, -sartan, -statin, -prazole, -dipine, and so on). Give me a made-up drug name using the suffix, and I have to tell you the class, the big side effect, and one thing the nurse checks before giving it. Correct me as we go.

Suffixes are the highest-leverage 20 minutes in all of pharm.

Give me one medication at a time with a short patient scenario (vitals, one or two labs). I answer: give it, or hold it and why. Push back on my reasoning before revealing the answer.

Hold-or-give is the exact decision pharm questions test.

Teach me acid–base balance for nursing school as ONE relationship, not four disorders. Start with what pH actually is (the ratio of bicarbonate to CO₂), which organ controls each side and how fast, and only then show me how the four disorders are four ways of tipping that ratio. Check my understanding with a question after each part before moving on.

Forces the assistant to start from the mechanism instead of the chart. The "check before moving on" clause turns a lecture into a tutorial.

Explain acid–base compensation to a nursing student, but organize it around TIME: what the lungs can do in minutes, what the kidneys do over 3–5 days, and why that means the same PaCO₂ of 60 can be an emergency in one patient and normal in another. Then give me two ABGs with that PaCO₂ and ask me which is which.

Timing is the piece charts leave out and the piece that makes compensation stages make sense.

Act as a nursing instructor. Give me one ABG at a time (pH, PaCO₂, HCO₃⁻, and a one-line patient story). Do not interpret it. Make me answer three questions in order — is the pH acidotic/normal/alkalotic, which value explains it, and is it uncompensated/partially/fully compensated — and only tell me if I was right after each answer, with one sentence of why. Mix in gases with a normal pH. Keep going until I get five in a row right.

Retrieval before feedback, one call at a time. The normal-pH clause is the trap the exam sets.

I got this acid–base NCLEX question wrong. Here is the question, the options, my answer, and the correct answer: [paste]. Do not just explain the right answer. Tell me which of the three ABG calls I got wrong, what I probably believed that made my answer look right, and what single rule would have corrected it.

Diagnoses the belief behind the miss, which is what changes the next answer.

Explain Winter’s formula for metabolic acidosis to a nursing student, then give me three metabolic acidosis gases: one with appropriate compensation, one with an added respiratory alkalosis, one with an added respiratory acidosis. Make me calculate the expected PaCO₂ for each and say which is mixed before you confirm.

Mixed disorders are a size question; the prompt makes you do the arithmetic before the reveal.

For each of these patients, tell me the acid–base disorder you expect, the ABG pattern, the two assessment findings that confirm it, the finding that would mean it is getting worse, and the first nursing action — in a table: DKA, three days of vomiting, opioid overdose, COPD at baseline, panic attack, salicylate overdose, septic shock with hypoventilation.

Converts the four boxes into the patients the exam actually writes about.

A patient with metabolic acidosis from DKA has pH 7.26, PaCO₂ 23, HCO₃⁻ 10. The options are: give sodium bicarbonate, start isotonic fluids, give regular insulin IV push, slow the respiratory rate with sedation. Explain why each wrong option is wrong in one sentence each, and what a student who picked it probably believed.

Wrong-option reasoning is where the marks are. Swap in any question you missed.

Using only the de-identified case below, sort the data into subjective cues, objective cues, abnormal cues, and missing assessment data. Do not create a nursing diagnosis yet. Case: [paste de-identified case]

Prevents a diagnosis from being chosen before the assessment is organized.

Review this de-identified nursing diagnosis statement for structure. Identify the problem, related-to factor, and as-evidenced-by cues. Flag any medical diagnosis used as the related-to cause, and explain what kind of modifiable nursing factor belongs there instead. Statement: [paste statement]

Targets the related-to versus as-evidenced-by distinction without writing the assignment for you.

Compare these de-identified nursing problems using ABCs and Maslow. Show how each framework ranks them, identify any immediate safety threat, and ask me to defend the final priority before you reveal your recommendation. Problems: [paste list]

Turns prioritization into an argument you must retrieve.

Audit this de-identified care-plan section. For each outcome, check whether it is specific, measurable, patient-centered, realistic, and time-bound. For each intervention, ask me for the rationale and how I will evaluate it. Do not invent patient data or citations. [paste section]

Keeps the work yours while checking whether it can actually be evaluated.

Here is my draft care plan: [paste]. Act as a sharp but fair clinical instructor. Question my nursing diagnosis choices and priority order, challenge any intervention that isn’t linked to an assessment finding, and make me defend each expected outcome. One question at a time; don’t rewrite it for me.

You keep authorship; the AI supplies the pushback your instructor will.

Generate a realistic patient for me to care-plan: [unit/topic, e.g. post-op day 1 bowel resection]. Give me history, meds, vitals, labs, and this shift’s assessment findings — some normal, some concerning, without labeling which. I’ll write the priorities and you critique them after.

Unlimited practice patients — the rep that clinical days can’t always provide.

Check the internal logic of this care plan [paste]: does every intervention trace back to an assessment finding, and does every outcome have a measurable target and timeframe? List any orphan interventions or vague outcomes. Do not add new content — just audit the links.

Broken assessment→intervention→outcome chains are the #1 grading comment.