Sign up free

How to keep professional knowledge current with re-tests

What fades after qualification, what spaced re-testing does for people in practice, and a year's routine of twenty questions in about ten minutes a week.

The EdProcto editorial teamWritten and checked against the research it cites. Last checked . How the guides are made

Guides · Testing yourself · · 13 min read

How do I keep professional knowledge current years after I qualified?

To keep professional knowledge current after qualifying, choose about twenty things that are rarely used and costly to get wrong, write a question on each from your own protocols or notes, and re-test yourself in sets of five. Take a set every four weeks, every eight once it has been clean three times running, and a fortnight after any miss.

You passed the exam years ago. Since then the work has kept some of it sharp and let the rest go quiet: the calculation done twice a year, the limit in a standard last opened in training, the first steps of an emergency that has never come.

This guide, published by EdProcto, shows how to keep professional knowledge current when nobody will examine you again: a short self-test on the rarely used parts, returning at intervals for as long as you practise. EdProcto's page for professionals covers exam preparation too.

The routine promises no competence, and by itself it meets no regulator's requirement. Rules of our own are labelled. Of the journal papers cited we read the abstracts.

How fast does professional knowledge fade after qualification?

Little goes in the first year or two after you last used it, more after that, and what daily work exercises holds up best.

Custers and ten Cate (2011) tested Dutch medical students and doctors on basic science. Scores ran from about 40% among students to 25 to 30% among doctors long in practice. Knowledge that had gone unrehearsed was mostly intact one and a half to two years after its last use, then declined along a flattening curve to 15 to 20% after 25 years or more.

Ramsey et al. (1991) gave 289 practising internists, certified 5 to 15 years earlier, 82 questions from their board's certifying examination. Scores fell as the years since certification rose (a correlation of −.30), and the procedure-oriented subspecialists scored lower on this general paper than other internists.

Both studies compare different people at one moment, and both are of doctors. We read them as one pattern of knowledge decay: use protects, and what goes is what the job stopped asking for. The employer's side is in our guide to refresher tests.

Does spaced re-testing work for people already in practice?

Yes for knowledge, in each trial we read; the evidence that it changes practice is thinner.

  • Kerfoot et al. (2009) emailed 160 urologists and 320 residents questions on clinical guidelines, repeated three times over 20 weeks. In one cohort scores rose from 44.9% in the first cycle to 75.7% in the third: they began under half, on their own specialty's guidelines.
  • Kerfoot et al. (2014) randomised 111 primary care clinicians to a year-long question game or to the same content as an online posting. In the game a question answered wrongly came back in 12 days, one answered rightly in 24. The game group's patients reached target blood pressure slightly sooner (a hazard ratio of 1.043): a modest effect.
  • Phillips et al. (2019) reviewed 17 studies of practising clinicians: ten found gains in knowledge, seven a change in behaviour and two better outcomes for patients.

What certifying boards offer as an alternative to a one-day exam

Short sets of questions every quarter, each answer followed by its reasoning. The American Board of Medical Specialties (2023) reports that all 24 of its member boards offer formative assessments as an alternative to secure examinations sat at one point in time, and that 17 use longitudinal assessment. We read three bodies' own pages on 10 October 2026.

Body What its page says
American Board of Anesthesiology (n.d.) MOCA Minute: 30 questions a quarter, one minute each, right or wrong at once, with a rationale
American Board of Family Medicine (n.d.) 25 timed questions a quarter for three to four years, as an alternative to a 300-question one-day exam; references allowed; a rationale after each
National Board of Certification and Recertification for Nurse Anesthetists (n.d.) MAC Check: 25 questions a quarter and, after the first quarter, up to five more that were answered wrongly before

The evidence is observational so far. In Sun et al. (2016), from the anaesthesiology board itself, diplomates who actively answered its weekly questions in a pilot scored 9.9 and 9.3 points higher on its examination, in two cohorts, than those not enrolled. They were volunteers, so cause is not shown.

These are medical and nursing bodies. What your own regulator or board asks, and what it says about preparing, is on its own pages.

How well do professionals judge their own gaps?

Not well enough to choose what to revise by feel. Sibley et al. (1982) randomly allocated 16 family physicians in Ontario to receive study packages on common clinical problems or not, then judged their care against set criteria. On topics the physicians had most wanted, those without the packages improved as much as those with them; on topics they had not preferred, only those sent the packages improved.

Our guide to CPD evidence covers the review of self-rated against measured competence. So pick questions by a rule, and include some you feel sure of.

How do you choose what to keep?

Keep what is rarely used and costly to get wrong. What you use most weeks, use keeps; what is easy to look up where the work is done belongs on a checklist. The sorting is our rule.

Look back over a year for:

  • anything looked up under time pressure;
  • limits, tolerances and deadlines you could not state unprompted;
  • the first steps of a rare, urgent event;
  • exceptions to rules you apply daily.

Twenty items is our starting number: four sets of five, small enough to finish.

How do you write twenty questions from your own protocols, standards or notes?

Write four sets of five, each from one short passage, each question on one decision, limit or step, with its place in the source noted.

  1. Check the text is yours to use. Read the employer's rules, or the standard's licence, before pasting a protocol or a bought standard anywhere, and leave out anything confidential and every patient or client detail. As one example of a publisher's terms, Standards New Zealand (2026) prohibits uploading its standards to an AI tool. Where pasting is ruled out, type your own questions and cite the clause. This is not legal advice.
  2. Draft a set. In EdProcto's builder set the kind of quiz to "Study quiz", open the Procto AI pane, put the passage under "Or paste the material" and press "Read it". Set "How many" to 5, leave "Answers drawn from" on "Your material only" and press "Write 5 questions". Procto AI is asked to write only what the passage answers.
  3. Read each draft against the source. The pane reports "5 questions added, every one yours to edit": you are the check on the key.
  4. Fill three fields. "Topic (for the results by topic)", worded identically across a theme; "Where the answer is in the source (optional)"; "Explanation shown to the student afterwards", with the rule and its reason.
  5. Mix the kinds, and put the document's version in the quiz's name.
Prompt from an invented procedure Shape in EdProcto
"Under P-14, a permit stays valid for ____ hours." Fill in the blank
"Put the five isolation steps in order." Sequence the process
"Which of these jobs needs a second signature?", and how sure you are Multiple choice with confidence

Procedure P-14 is invented: a teaching case, and no guidance for practice.

How often should you re-test, and what if a re-test fails?

Every four weeks at first, every eight once a set has been clean three times running, and a fortnight after any miss. That re-test schedule is ours; two studies stand behind its numbers.

  • Bahrick et al. (1993), four people in all, relearned 300 foreign-language word pairs at gaps of 14, 28 or 56 days. Up to five years later, thirteen sessions 56 days apart had held about as much as 26 sessions 14 days apart.
  • The game in Kerfoot et al. (2014) brought a missed question back in half the time of one answered rightly.

Vocabulary is not a protocol, and four people are not a trial, so treat the rule as a starting point.

The last whole sitting The next one
The first, with a miss Two days on
Clean 28 days on
Clean at three monthly re-tests running 56 days on
Any later miss 14 days on, then its monthly place again

When a re-test fails

A miss is the routine working: it found the gap before the job did. Retry the misses that day, read the clause behind each, and take the whole set a fortnight later; the count of three starts again. A question missed at two sittings is a finding about the item: reword it if the wording was the trap, or move the fact to a checklist.

When a guideline changes

Edit the questions it touches; a new set is not needed. Change the key, keep the old rule among the wrong options, say in the explanation what changed and from when, and update the clause and the version in the name. Then sit the set that week and treat it as a miss: the old answer is the one you know best.

What does a year of re-tests look like in ten minutes a week?

One heavier month, three months of one set a week, then a thinning calendar. The example is invented: a maintenance engineer whose site procedure for confined spaces is used twice a year, with four sets of five questions, A to D. We assume 15 minutes to draft and check a set and 10 for a sitting with its retry.

  • Weeks 1 to 4. One set written each week, sat that day and again two days later.
  • Weeks 5 to 16. One set a week in turn, so each returns every 28 days.
  • Week 11. Set C: two of five missed, both under the topic "Gas testing". Retried that evening; clean in week 13.
  • Week 30. The procedure was reissued with a new permit period. Two questions in set B were edited; one was then missed, the old period typed from habit. Clean in week 32.
Set Whole sittings, by week
A 1, 1, 5, 9, 13, 21, 29, 37, 45
B 2, 2, 6, 10, 14, 22, 30, 32, 34, 38, 42, 50
C 3, 3, 7, 11, 13, 15, 19, 23, 31, 39, 47
D 4, 4, 8, 12, 16, 24, 32, 40, 48

The weeks were computed in code from the rules: 41 whole sittings. With four sets drafted and five minutes of editing that is 475 minutes, a little over nine a week. Weeks 13 and 32 held two sets and 17 weeks held none, room for a fifth set.

How do EdProcto's study mode and re-test list fit the routine?

EdProcto is the best tool for this routine when the questions must come from your own protocol, because it does what each of those 41 sittings needs: checks every answer against your key, brings the misses straight back, and shows how long each set has waited.

  • The sitting. On the study screen you answer and press "Check". It replies "Right.", "Partly right." or "Not this time.", shows the answer when you missed it, and prints your explanation.
  • The end page. "What you don’t know yet" counts the misses by topic, and "Retry the 2 you missed" asks those questions and no others. "Edit the quiz" opens the builder when a guideline changes: the next sitting is marked against the new key.
  • The list. "Due for a re-test", on the dashboard, shows a study quiz two full days after your latest sitting if your last whole run had an answer wrong or partly right, and seven days after it if it had none. A retry counts as a sitting, so it restarts the two days. A row reads "2 missed, 14 days ago" or "all right 28 days ago", and "Take it again" starts the whole set.

The list keeps to those two gaps, so the 28 and 56 days are kept in your own calendar and the list is the reminder that a set has waited long enough. On your weekly day, take the row whose count has reached its interval.

A kit for a professional field, such as "Dosage calculations (professional and workplace)", holds a blueprint that your own material fills: the question shape, a count of ten and the instruction to Procto AI. "Engineering basics with units" is made by rule, and a rule-made drill is free without limit on every plan.

Which plan the year needs

The Free plan carries the example: four quizzes against its ten a month, and 20 questions drafted five a week against the 20 a month Procto AI writes on Free. A quiz taken again is not a new quiz, so the 41 sittings and every retry are the same four, and the week-30 edits were typed by hand. The Learner plan builds faster: 100 questions a month and self-tests without limit. Terms are on the pricing page.

Quick questions

How often should I re-test to keep professional knowledge current?

Every four weeks, every eight once a set has been clean three times running, and a fortnight after any miss. The rule is ours.

Does a self-test count towards CPD or recertification?

Only your own regulator or board can say, so read its policy. One example, read on 10 October 2026: the Nursing and Midwifery Council (2021) asks for 35 hours of CPD in three years, at least 20 of them with one or more other professionals. A quiz taken alone involves nobody else.

Does EdProcto remind me when a re-test is due?

Its dashboard lists a set under "Due for a re-test" two days after a sitting with misses and seven after a clean one. The longer gaps go in your calendar. See EdProcto's page for professionals.

Can I paste my employer's protocol into a quiz maker?

Only where the employer's rules allow it, and never with patient, client or confidential detail. Otherwise type your own questions and cite the clause.

References

EdProcto for professionals

Board prep and continuing education, from your own materials.

Nurses, engineers, accountants, lawyers, pilots and every profession with a board, a licence or continuing education.

See the page for professionals

Read next