The denominator of a minute
A minute saved by a clinical scribe cannot be compared across hospitals without knowing its denominator. Time per note, daily time per clinician and after-hours work answer separate questions. The most useful contribution of this review, in my view, is to put that distinction at the center of evaluation. Amama Mahmood and colleagues’ November 10 preprint finds 96 operational definitions for 40 metrics across 54 studies in real patient care. That diversity can help explain why findings about one technology resist accumulation. Individual positive results can remain useful while we ask which common unit makes the overall evidence comparable.[1]
When visit volume rises, time per note can fall while total daily documentation time increases. Both results can be correct because the hospital has asked different questions. The review uses this example to show that even an automatically extracted EHR metric is not inherently comparable. My interpretation is that the denominator and time window belong inside the result. Institutions can also differ in workflow and patient mix, however. Treating every disagreement as a calculation artifact would go beyond what the evidence supports.[1]
The hidden effect of use
The problem grows when we ask who continues using the tool. Only 15 studies report both documentation efficiency and adoption or utilization. According to the review, 21 report time changes without stating what share of clinicians or notes used the scribe. People who find it unhelpful may stop, leaving a more favorable average among continuing users. A hospital-wide result per clinician and a result restricted to retained users therefore describe different effects. The review provides a concrete reason to report exposure alongside outcomes; it does not calculate that every existing benefit comes from this selection process.[1]
Note length is another easily misread metric. A longer note can reflect restored detail or unnecessary repetition. Automatically counting its volume does not identify which interpretation applies. This makes the authors’ proposal for direct audits of clinically significant errors and omissions valuable. Measuring time and accuracy on the same notes can show what correction work remains alongside the minutes saved. The accuracy audit supplies an additional outcome needed to interpret an efficiency result for a particular product.[1]
A separate question for safety
The distribution of measurement makes the distinction pressing: 44 of the 54 studies assess efficiency and time, while only 2 cover safety and clinical outcomes. The three randomized trials last 3–4 months. Random assignment can reduce some comparison problems, but it does not remove expectation effects from self-report when clinicians know which tool they use. The outcome receiving that stronger design matters. This review provides no basis for reading an improved workload survey as a patient-safety result. Measuring efficiency and safety separately makes it possible to assess them together.[1]
The review itself deserves the same restraint. Its fixed search cutoff excludes newer work. The review maps evaluation practice, with counts unweighted by study size or quality, rather than estimating one pooled treatment effect in a meta-analysis. My priority from that map is to pair a time measure with a defined denominator, a utilization rate and a direct audit of note accuracy. That reporting would support questions about which work a faster note simplifies and which work it leaves to downstream readers. Understanding the benefit requires placing those questions alongside the minutes saved. If time savings persist among all eligible clinicians while audited errors and omissions remain stable, that combination provides a more interpretable local result.[1]