One population, six different measures
Published September 15 in Drugs & Aging, the retrospective cross-sectional study compared six measures using 2017–2018 administrative data for 330,684 Western Australian adults aged 65 or older. The proportion classified as exposed ranged from 13.1% to 53.6%; high-exposure classifications ranged from 0.7% to 6.9%. Each instrument used its own high-exposure threshold, so those categories did not represent equivalent absolute burden.
The study compared classification, not software performance or clinical outcomes. Dispensing records did not establish consumption and incompletely captured over-the-counter, private-prescription, and hospital-supplied medicines. Annual measures could obscure shorter exposure peaks; excluding people who died during the year could omit frailer adults. Older data and medication lists also limit present-day interpretation. These are not U.S. nursing-facility prevalence estimates.
Start with the number on the screen
Our editorial implication is a purchasing question: can the person receiving a burden score tell how it was produced? For a small consultant pharmacy practice, an unexplained number leaves another task for the pharmacist who must interpret and communicate it. Put that task into the demonstration rather than accepting a feature label as sufficient evidence.
Ask the supplier to open a de-identified demonstration record and identify the instrument by its full name. Request the version or revision date, supporting documentation, medication list, scoring formula, and threshold definitions. If the implementation modifies a published method, ask which parts changed and where those modifications are documented. Keep the answer with the evaluation notes.
Then ask what the displayed score represents: the current list, a particular review date, a rolling interval, or an accumulated history. Have the demonstrator point to the actual dates used. A report should give its recipient enough context to understand that choice without needing the original sales presentation.
Follow the inputs into the report
Choose one entry in the demonstration and follow its contribution to the total. Ask which fields the calculation uses, how a medicine is matched to the instrument, and how an unmatched entry is presented. If a required dose or date is missing, inspect whether the software marks the result incomplete, uses an assumption, or omits the entry. Record the behavior actually shown.
Next, ask the demonstrator to change one input and explain the resulting calculation. The purpose is to establish whether the output can be checked and understood. Do not turn this exercise into an invented clinical case or a test of which medicine someone should stop. Keep it focused on the software’s handling of information.
Export the ordinary report. Can its reader identify the measure and time window, see the relevant contributors, and distinguish unavailable information from a calculated zero? If those details live only behind a tooltip or in a separate module, ask how the receiving professional obtains them. Save the sample output alongside the supplier’s written explanation.
Ask what survives an update
Use a second demonstration to examine history. Ask what happens when a medication list, potency assignment, threshold, or formula changes in a software update. Does an earlier report retain its original result and method? Can a later recalculation be distinguished from the record issued at the time? Ask for the release note or documentation that explains the behavior.
For a practice comparing facilities or reporting changes over time, request a view that makes method changes visible alongside the dates. An apparent difference deserves a reproducible explanation before it becomes a performance claim. If the product cannot provide that context, record the limitation and the manual work needed to supply it.
Keep the pharmacist’s decision visible
Finish by following the flag into the review workflow. Where does the pharmacist record an assessment, a recommendation, or a reason for taking no action? Where does a response appear? Ask the demonstrator to show those separate records rather than treating the presence of a score as completion of the review.
These questions do not select a preferred instrument or establish that any reviewed vendor supplies these capabilities. They give a practice concrete evidence to request and compare. A useful acceptance record names the method, preserves the sample output, states what was demonstrated, and identifies what still needs verification.