Medication Safety Review: Catching What Is Wrong on the ChartMedication Safety Review. Every course in this field tests whether somebody remembers the rules.
Medication safety review is the skill of telling a genuinely unsafe medication record from one that only looks unusual. It is scored by signal detection, which means marking everything is penalised exactly as much as marking nothing, and the report separates how well you tell the two apart from where you set your line.
Ten records to mark, sixty entries, and no reward for marking them all
Ten real-looking medication records are handed to you and you mark every entry that must be queried before anything is given. Sixty entries in all. Twenty-seven carry a genuine defect - a unit shortened so it reads two ways, a weight-based entry with no weight recorded, a short course with no stop date, a spoken order never countersigned, a blank box with no code against it. Thirty-three are correct practice that only looks irregular: a generic name with the brand beside it, a dose written in words as well as figures, an entry stopped with one ruled line and initials and a date, a hold-if instruction that names the reading to check.
That structure is the whole instrument. A reviewer who marks everything and a reviewer who marks nothing can score the same percentage on a test that counts defects found, and they are opposite problems. So this one reports two figures that a percentage destroys: how far apart the two kinds of entry are for you, and where you have set your threshold. Two people who separate them equally well can sit at opposite ends of that second scale - one stops a round over an entry that was correct, the other lets the ambiguous one through - and a ward needs to know which it has.
Nothing in it requires clinical or pharmacological knowledge. No medicine is named, no dose figure appears anywhere, and no item can be answered by knowing what a drug does. Every defect is a defect of the record or the process, which is what makes it readable in any country and what makes it a test of reviewing rather than of remembering. Twenty-two further situations cover the interruption part-way through a round, the discrepancy found at handover, the prescriber who is hard to reach and hard to challenge, and what you do about the near miss that harmed nobody.
What you walk away with
How far apart a real defect and a correct entry are for you, drawn as a ribbon rather than a point, with the two rates it is built from printed underneath it.
Whether you mark readily or sparingly. It is a preference and not an error, it is reported as its own finding, and it is never folded into the number above.
Every defect class you met, sorted by what you caught least often, so the gap is a specific kind of absence rather than a general score.
Caught, missed and marked in error on every record, in a table that survives a black-and-white print.
The interruption, the handover discrepancy, the hard-to-challenge prescriber, the near miss - graded and corrected for chance before any band is assigned.
Inside your report
Illustrative sample - your report is generated from your own responses.
One number, its band, and the parts it is made of. Marking everything scores no better than marking nothing.
Four rungs of rising pressure per situation. The cell where the shade changes is the finding.
Built for
- Hospitals, care homes, domiciliary services and pharmacies hiring or reviewing people who handle medication records
- Nurses, care staff, medication assistants and pharmacy technicians who want to know how their reviewing actually reads
- Educators and training providers running medication-safety programmes who need a measure that is not a rules quiz
Find out whether you are separating the defects from the entries that only look odd
32 scored exercises - about 40 minutes - a full bespoke report with your discrimination figure and its band, your threshold reported apart, and every defect class you met.
₹999 (incl. GST) · assessment and full report, nothing further to pay
Frequently asked questions
Whether a person can tell a genuinely unsafe medication record from one that only looks unusual, and what they do about it. Four capabilities are reported: reading the record closely, the checks before it is given, raising it and recording it, and what you do after something goes wrong. It does not measure clinical or pharmacological knowledge, dose calculation, any country's regulations, or diligence as a character trait.
No, and that is deliberate. No medicine is named, no dose figure appears anywhere, and no item can be answered by knowing what a drug does. Every defect is a defect of the record or the process - a missing route, an ambiguous shortening, an unsigned spoken order, a short course with no stop date. That keeps the construct clean, keeps it fair across countries, and means it does not go out of date when a formulary changes.
By signal detection. Two figures come out: a discrimination figure for how far apart the real defects and the correct entries are for you, and a threshold figure for where you set your line. A single percentage cannot separate them, so a reviewer who marks everything and a reviewer who marks nothing can score the same on an ordinary test and are opposite problems. The situational block is graded separately and corrected against the declared choice rates before it is banded.
No, and the scoring says so. Marking an entry that was correct stops a round, costs a colleague their morning and delays the person waiting, and delay is a harm. Three of the situations key the less cautious action for the same reason. An instrument where more caution always wins can be passed by ticking the careful box, and it teaches a team to query everything, which is how genuine queries stop being heard.
Rs 999 in India including GST, or US$9.99 elsewhere, one time, for one full sitting and report. Medication administration courses run from about $20 for a refresher to well over $150 for a certificate, and they teach and test remembered rules; competency sign-off in a workplace is an observation by a colleague with no comparable data behind it. Nothing on the market scores this as reviewing judgment. Organisations can use AssessAll credits at 26 credits per person.
Each one takes a single capability, puts you inside the situations where it is actually tested, and scores your choices against published evidence — with a report designed for that capability alone, not a template. They span hiring, compliance, education, operations and personal skill.
Browse the catalogue →Methodology: Measures whether a person can find the defects in a medication record and act on them, through original review exercises, situational items and self-report statements. Construct statement: it measures whether a person can tell a genuinely unsafe medication record from one that merely looks unusual, and what they do about it - and it does not measure clinical or pharmacological knowledge, dose calculation, any country's regulations, or a person's diligence as a character trait. No item names a medicine, states a dose figure, or asks whether a dose is right; every keyed defect is a defect of the record or of the process, findable by a trained reviewer who knows nothing about the medicine, which also keeps the instrument country-neutral. Declared response instructions, stated to the respondent and split by block: the fourteen situational items use behavioural-tendency framing and ask what the respondent is most likely to do, never what should be done; the ten review exercises use a marking instruction and ask the respondent to mark every entry that must be queried before anything is given. The two blocks are reported separately and are never fused into one number. Item format: ten six-entry review exercises whose number of real defects varies from chart to chart and is never none and never all; fourteen four-option situational items scored with graded partial credit; eight balanced-keyed self-report statements, half of them reverse-scored. Scoring design: the review block is scored by signal detection rather than by counting correct marks. Each of the sixty entries is one trial; a mark on a real defect is a hit and a mark on correct practice is a false alarm. Discrimination is reported as d-prime, the distance between the two distributions, and threshold as the criterion c, which says how much evidence the person wants before marking an entry. Both are computed with the log-linear correction, half an observation added to each cell, so that extreme rates stay finite. They are reported apart from one another because a cautious flagger and a lax flagger at the same discrimination are usefully different people, and because marking everything is penalised exactly as much as marking nothing. Construct areas drawn on across the instrument: signal-detection theory and the separation of sensitivity from response bias; medication-error taxonomy research distinguishing prescribing, transcription, dispensing and administration errors; work on error-prone abbreviations and notation as a documented cause of misreading; independent double-check research and its limited and contested effect; interruption research during medication rounds; the rights-of-administration checking tradition and its critiques; look-alike and sound-alike confusion research; speaking-up and hierarchy research in clinical teams, including the observed reluctance to challenge a senior prescriber; just-culture and blame-free reporting research on why errors go unreported; second-victim research on what happens to the person who made the error; and reconciliation research at transfer points, where the majority of discrepancies arise. All items are original works, no medicine, dose, regulator, formulary or trademarked instrument is named or reproduced, no clinical advice is given or implied, and no affiliation with any source is claimed. AssessAll original design.