Researchers rebuilt the medication histories of 154 patients using an automated prescription-exchange service — a shared, cloud-based record that lets pharmacies and prescribers pull up a patient’s medications without asking the patient to list them. 153 of the 154 (99.4%) had at least one error in their history, a median of six errors per patient. The single biggest cause, responsible for 72.1% of all omissions, was over-the-counter medicine — a category the automated system never saw in the first place.
Here’s what that number means, and why it’s a different failure from a patient simply not mentioning something.
This isn’t the OTC gap we’ve already covered
We’ve written before about what happens when patients are asked directly about over-the-counter (OTC) and herbal drug use. In a survey of 577 primary care patients, 94% said it mattered that their provider knew about their OTC medications. But only 65% said they always mentioned them. That’s a gap in willingness to speak up — a patient choosing, for whatever reason, not to volunteer something. The finding below is a different kind of problem, not just a bigger version of the same one. It’s what happens inside a system built so the patient never has to speak up at all, because the system is supposed to pull the record automatically. It still missed almost everything. The reason is structural, not behavioral: an over-the-counter purchase usually doesn’t create the prescription or pharmacy transaction these exchange systems are built to track. So there’s nothing in the system for it to find.
99.4% of automated medication histories had at least one error
- 153 of 154 patients (99.4%) had at least one error in a medication history pulled from an automated prescription-exchange system connecting prescribers and pharmacies.
- Patients had a median of 6.0 errors each.
- Omission was the most common error type, affecting 33.3% of current medications.
- 72.1% of those omissions were over-the-counter medicines — the single largest category of error in the entire dataset.
(Elliott RA, Taylor SE, Koo SMK, Nguyen AD, Liu E, Loh G, Internal Medicine Journal 2022)
This isn’t a one-off number
A review of 22 studies covering 3,755 patients found that 60–67% of patients had at least one omission or commission error — a medicine left off the list, or a wrong one added — in their medication history at hospital admission. More than a quarter of hospital prescribing errors traced back to an incomplete admission history (Tam VC, Knowles SR, Cornish PL, Fine N, Marchesano R, Etchells EE, CMAJ 2005;173(5):510–515). That review looked at hospital admissions in general, not pharmacy-to-pharmacy record transfers specifically — so it backs up the scale of the problem rather than repeating the exact finding above. But the pattern holds either way: rebuilding a medication history from records, automated or not, leaves a large and persistent gap.
An Australian finding, not a guarantee about every system
This data comes from an Australian prescription-exchange service, not from U.S. pharmacies — so the specific 99.4% figure shouldn’t be read as a prediction for any particular American system. What does carry over is the reason behind it: an over-the-counter purchase doesn’t produce a prescription or a billed pharmacy transaction, so any system that rebuilds a history from those transactions — in any country, from any vendor — has nothing to find for that category in the first place. The number is Australian. The blind spot is built into the design.
A record-exchange system built so nobody has to ask the patient still missed almost everything — because over-the-counter medicines never create the transaction these systems are built to track.
What this does and doesn’t claim
This isn’t a claim that automated medication-history exchange is worthless. It solves a real problem — getting a prescriber’s and a pharmacy’s records to match — that this product doesn’t try to solve at all. ConfirmMyMeds’s own pharmacy import screen takes whatever file or export a pharmacy’s system already produces. It is not a cross-pharmacy exchange system, it doesn’t use the data-exchange formats those systems rely on, and it doesn’t integrate with or replace any prescription-exchange service. On top of that import sits a separate layer: the patient confirms their own list directly, with no pharmacy-to-pharmacy feed involved. That puts the patient in a position to catch exactly the category — OTCs — that never creates a transaction for an automated exchange to find. That’s a statement about where the two approaches are aimed, not a claim that one has been shown to work better than the other, or that either one closes the gap completely. We have no outcome data on that question.
ConfirmMyMeds is a working application. It asks the patient to confirm their clinic’s medication list before the appointment — outside the portal, with the bottles in hand — and returns what they confirmed, what they flagged, and what they never started, timestamped and attributed. It never writes to your record; the clinician decides what goes in the chart and what to do about a gap. We have no outcome data and make no claims about adherence or affordability outcomes. We’re looking for clinics to help us learn where it fails.
This article discusses medication-history accuracy research in general terms. It isn’t medical advice.