The short answer
These CPHQ exam questions and answers are original items built to the shape of the real paper: a scenario, four defensible options, one right for a stated reason. Shape matters, because only 23% of the 125 scored items test recall. Application is 57% and analysis 20%, so an item answerable from a definition is rare.
How CPHQ questions are built
Every item has a stem, a qualifier and four options. The qualifier decides the answer: first, next, best, most likely. Two or three options are things a competent quality professional would do at some point; only one is right where the stem stops.
The fifteen below are proportioned to the item counts NAHQ publishes, so the practice matches the exam. The full weighting is in the CPHQ exam blueprint.
| Domain | Scored items | Share | Questions here |
|---|---|---|---|
| Performance and Process Improvement | 27 | 21.6% | 3 |
| Health Data Analytics | 26 | 20.8% | 3 |
| Quality Leadership and Integration | 19 | 15.2% | 2 |
| Patient Safety | 18 | 14.4% | 2 |
| Quality Review and Accountability | 16 | 12.8% | 2 |
| Population Health and Care Transitions | 11 | 8.8% | 2 |
| Regulatory and Accreditation | 8 | 6.4% | 1 |
Fifteen original CPHQ exam questions and answers
Question 1. Health data analytics
A team plots monthly falls per 1,000 patient days for two years. One month sits far above the rest: the month a new electronic reporting form went live. What most likely explains it?
- A genuine deterioration in falls prevention that month
- Special cause variation from a change in how events were captured
- Common cause variation within the expected range
- A calculation error in the denominator
Answer: B. An easier reporting route changes ascertainment, not the rate, and that is a special cause acting on the measurement system. A is unlikely when the point coincides with a data-collection change. C is wrong: one extreme point is not within-range variation. D has no support in the stem.
Question 2. Health data analytics
Your crude mortality rate exceeds the regional average and the board asks whether care is worse. What should you do before answering?
- Compare case-mix adjusted rates instead
- Repeat the calculation for the last three years to confirm the trend
- Audit ten deceased patients' notes for avoidability
- Present the crude figure with a note that it is unadjusted
Answer: A. Crude rates cannot be compared between populations with different risk profiles. B tests stability, not comparability. C cannot answer a comparative question from ten records. D hands an uninterpretable number to a board that will act on it.
Question 3. Health data analytics
A unit reports 14 pressure injuries in a month, with 700 admissions and 3,500 patient days. Which measure best supports comparison with units of different size and length of stay?
- The count of 14 injuries
- Injuries per 100 admissions
- Injuries per 1,000 patient days
- The percentage of patients with any injury
Answer: C. Risk accumulates with time in bed, so a patient-day denominator is the fair comparator here. A has no denominator. B ignores length of stay, the confounder here. D shares that weakness and discards patients with more than one injury.
Question 4. Performance and process improvement
A team wants to test a new discharge summary template on a 30-bed ward. What is the most appropriate first Plan-Do-Study-Act cycle?
- Use the template for every discharge on the ward for one month
- Use it with one doctor for the next five discharges
- Circulate the template for comment and revise it before any use
- Write a policy mandating the template, then audit compliance
Answer: B. A first cycle should be small enough that failure costs nothing and fast enough to teach you something this week. A is full implementation dressed as a test. C produces opinions, not data. D implements before any test, the commonest way improvement work dies.
Question 5. Performance and process improvement
On-time theatre starts have risen from 40% to 85% over six months and the project team is standing down. What most reliably sustains the gain?
- A celebration event and a report to the executive team
- Embed the start checklist in the standard theatre process, with an owner and a monthly measure
- Repeating the project on a second site
- Continuing the weekly project meetings indefinitely
Answer: B. Sustainability comes from the change becoming ordinary work, with a named owner and a measure. A recognises the work without protecting it. C is spread, a different objective. D keeps the scaffolding up and proves the process cannot stand.
Question 6. Performance and process improvement
Emergency department waiting times are rising and staff offer six competing explanations. You have process data but have never watched the process. What is your next step?
- Build a process map with the staff who do the work, then verify it by watching
- Run a staff survey to rank the six explanations
- Add two triage nurses on the busiest shift as a trial
- Benchmark waiting times against three comparable departments
Answer: A. When explanations conflict, establish the current state first, and a verified process map does that. B ranks opinions rather than finding the constraint. C spends resource on the most popular guess. D tells you that you are slower, not why.
Question 7. Patient safety
A patient receives a tenfold insulin overdose. The nurse followed a workaround most of the unit uses, because the pump library entry is out of date. Under a just culture approach, what is the appropriate response?
- Retrain the nurse and record the event in her file
- Treat it as at-risk behaviour: correct the library and address the drift unit-wide
- Take no action, as the system was at fault
- Suspend the nurse pending investigation, as the harm was serious
Answer: B. At-risk behaviour is drift into an unsafe habit for reasons that made local sense; remove the reason and coach everyone doing it. A punishes one person for a unit-wide condition. C mistakes system fault for no accountability. D lets severity set the response, the error just culture exists to prevent.
Question 8. Patient safety
A hospital is opening a chemotherapy day unit and wants to reduce the risk of error before the first patient arrives. Which method fits?
- Root cause analysis
- Failure mode and effects analysis
- Incident trend analysis of the old unit
- A safety culture survey
Answer: B. FMEA is prospective: it maps a process that has not yet failed and prioritises the failure modes worth designing out. A is retrospective and needs an event. C examines a different process. D measures conditions for safety, not failure points in this workflow.
Question 9. Quality leadership and integration
The board receives a 40-page quality report each quarter and rarely discusses it. What is the best change?
- Shorten the report to ten pages of the same tables
- A one-page set of board-level measures over time against target, with exceptions narrated and decisions asked for
- Move the report to a monthly cycle so less is covered each time
- Add a verbal presentation from each clinical director
Answer: B. A board needs trend, variance from target and a decision to make. That is what turns a report into governance. A cuts volume without changing the format that caused the problem. C repeats the failure monthly. D adds advocacy without comparability.
Question 10. Quality leadership and integration
You lead a hand hygiene improvement across a division where you have no line authority and compliance has stalled at 60%. What is the most effective approach?
- Ask the medical director to issue a directive
- Recruit a respected clinician in the division as visible sponsor, with the division's own unblinded data
- Publish a league table naming the worst-performing units
- Increase audit frequency from monthly to weekly
Answer: B. Influence without authority runs on local credibility and local data, and a peer sponsor carries both. A produces compliance in name only. C generates disputes about the data rather than change. D measures the problem more often without acting on it.
Question 11. Quality review and accountability
A newly appointed surgeon has completed six months of practice and review shows performance within expected limits. What is the appropriate next step?
- Continue the focused review for a further year
- Conclude the focused evaluation and move to ongoing periodic review
- Take no further action until an incident occurs
- Repeat the focused evaluation with a different reviewer
Answer: B. Focused evaluation of a new practitioner is time-limited and ends when the evidence supports competence; periodic evaluation then continues for everyone. A extends a temporary process without a trigger. C ends monitoring altogether. D duplicates completed work.
Question 12. Quality review and accountability
A department asks you to release individual peer review findings to a service improvement group that includes administrators. What is your best response?
- Release the findings, as quality improvement is a legitimate purpose
- Give the group aggregated, de-identified findings and keep case review inside the peer review process
- Refuse any information to the group
- Release the findings with names removed but cases individually described
Answer: B. Improvement needs the pattern, not the practitioner, and aggregation serves the group without breaching the confidentiality peer review depends on. A treats a legitimate purpose as if it removed that requirement. C withholds usable information. D is re-identifiable in a small department.
Question 13. Population health and care transitions
Thirty-day heart failure readmissions are rising, and most readmitted patients had no follow-up contact after discharge. Which intervention addresses the finding?
- Extend the average length of stay by one day
- A telephone call within 48 hours and a clinic appointment booked before discharge
- A revised patient information leaflet on heart failure
- A readmission dashboard for the cardiology directorate
Answer: B. The gap is the transition itself, so the intervention must sit in it, with a timeframe and an owner. A treats time in hospital as a proxy for readiness. C is education, not contact. D reports the problem without changing anything.
Question 14. Population health and care transitions
A primary care network has 12,000 patients with diabetes and resource to coordinate care intensively for 400. How should they be selected?
- First come, first served from patient self-referral
- Risk stratification on clinical, utilisation and social factors, to find who benefits most
- All patients with the highest recorded HbA1c
- An equal allocation across the network's practices
Answer: B. Population health management allocates a scarce resource by predicted benefit, and that prediction needs clinical, utilisation and social inputs. A selects for confidence and access, not need. C uses one variable and misses drivers such as isolation or housing. D is fair administratively and arbitrary clinically.
Question 15. Regulatory and accreditation
Your hospital expects an accreditation survey within the year, and the executive team proposes a documentation project in the last month. What do you recommend instead?
- Continuous readiness: routine self-assessment and mock tracers against the standards, with gaps owned
- A rehearsal week immediately before the expected survey window
- External consultants to rewrite the policy manual
- Suspending improvement projects until after the survey
Answer: A. Readiness is a property of the working system, and tracers test practice rather than paperwork. B is a performance that ends when the surveyors leave. C fixes documents while practice stays unchanged. D removes the activity that produces the evidence.
The five distractor patterns that catch people
The wrong options above are not random. They come from five patterns, and naming the pattern is faster than reasoning an option from scratch. When two options survive elimination, ask which of these the weaker one is.
- True but not responsive. A correct statement that does not answer the question asked, as in question 2, option D.
- Right action, wrong moment. It belongs later in the sequence: benchmarking before you understand your own process, spread before the first gain holds.
- Education or policy as the fix. Training, a rewritten policy or a directive, where the stem describes a system condition.
- Resources the stem never gave you. Authority, budget, data or time the scenario has explicitly withheld.
- Severity driving the response. The punishment tracks how bad the outcome was rather than the behaviour that produced it. This is the trap in every just culture item.
Why exam dumps are a trap
Sites offering "real" CPHQ questions are a fixture of these search results. Three reasons not to use them.
The first is that they do not work. A dump teaches the answer to a remembered item, and recall is 23% of the paper. The reasoning that carries the other 77% is what an answer key cannot build. Candidates who prepare this way recognise topics and still cannot choose between options.
The second is quality. Reproduced items circulate with wrong keys, from expired outlines, with no reasoning attached.
The third should settle it. Distributing or using leaked live items breaches the candidate agreement you accept when you apply, is a copyright matter, and puts the credential itself at risk rather than only the sitting. [VERIFY: exact wording of NAHQ's candidate agreement and code of conduct on disclosure or use of exam content] The credential is renewed every two years through continuing education, so a finding against you is not a one-day problem. [VERIFY: current CE credit requirement for CPHQ recertification, and NAHQ's stated sanctions for exam-content violations]
Candidates who arrive with a folder of downloaded questions are not lazy. They are anxious, and "real" questions feel like certainty. The certainty is false, and the fix is not more questions but a different use of them: fewer items, marked immediately, each with a written sentence on why the correct option beat theirs. [INSERT PRACTITIONER STORY: one paragraph, 110 to 140 words, on an anonymised candidate who prepared from dump-style question sets, what their mock performance looked like compared with their real result, and what changed when they switched to worked items with reasoning. No names, no hospital, no city.]
Where to get more practice
Work in sets of 20 to 30 and keep an error log with one line per wrong answer: content you did not know, a misread qualifier, rushing, or a lost close pair. The repair differs in each case, and the method is in the guide to passing the CPHQ exam on a first attempt. If your errors cluster in analytics, as they do for most clinicians, work through the health data analytics domain first.
For a timed set under exam conditions, sit the free CPHQ practice test. For a full bank with reasoning attached to every option, our Complete CPHQ Prep course includes 1,000+ practice questions and full-length mocks for a single payment of $399 with lifetime access. Judge yourself on the explanations you can give, not the score.
