This trial tested a sequence, not a single medication rule
Alzheimer's & Dementia published the peer-reviewed StaN trial online July 27, 2026. Researchers enrolled adults aged 50 or older with possible or probable Alzheimer's disease dementia and clinically significant agitation. The 185 participants came from five psychiatry inpatient units and seven long-term-care homes in three Canadian provinces. Of the 92 LTC residents, 46 were assigned to the integrated care pathway and 46 to treatment as usual.
The pathway began with a cleanup of medicines considered unnecessary or ineffective for agitation. At the same time, the team completed a structured behavioral assessment and implemented a personalized behavioral plan. If needed, a sequential, measurement-guided psychotropic phase typically began after three weeks while behavioral work continued. The research team discussed participants weekly and made recommendations; clinical teams remained free to follow them or not. This was a supported care process, not a stand-alone checklist or software alert.
The week-12 primary results matter more than the early snapshot
The two primary outcomes were agitation and psychotropic polypharmacy at 12 weeks. Among LTC residents, change in Cohen-Mansfield Agitation Inventory scores did not differ significantly between the pathway and usual-care groups over time (p=.48). That is not proof that the approaches were equivalent; it means this trial did not detect a group difference in agitation under its analysis.
The medication trajectory was more complicated. The study defined psychotropic polypharmacy narrowly as at least two scheduled psychotropics used to treat agitation. In LTC, the pathway group moved from 28% at baseline to 20% at week 1 and 25% at week 12. Usual care moved from 26% to 30% and then 24%. Adjusted post hoc comparisons separated at weeks 1 and 3, but the paper does not report a multiplicity adjustment for those time-point comparisons, and the groups did not differ at week 12. The early finding was not a sustained endpoint advantage.
Keep the medication count beside the resident outcome
The authors also reported no group differences in overall neuropsychiatric symptoms, falls, caregiver burden, health-related quality of life, time to agitation response, or adverse events. Falls were uncommon and the trial was not powered to establish a falls difference. Investigators judged no adverse event causally related to the pathway, although mild events were numerically higher in that group, driven by more urinary tract infections.
That combination blocks two easy overstatements. The study does not show that reducing a medication count improved agitation or prevented harm. It also does not show that cleanup was pointless. Instead, it demonstrates why a medication-use measure and a resident-centered outcome should stay visible on the same timeline. One can change while the other does not separate from usual care.
Build a trajectory record, not a before-and-after trophy
For an independent consultant pharmacist, the transferable lesson is about measurement. Preserve the baseline target behavior and exact scheduled medicines attributed to agitation. Then date each cleanup decision, behavioral assessment, non-drug intervention, authorized medication change, response check, and later restart or escalation. Keep PRN use and psychotropics for other indications outside the study's narrow count unless they are deliberately analyzed as separate measures.
A dashboard should be able to show baseline, week 1, week 3, and the planned endpoint without silently promoting the most favorable snapshot. It should also distinguish a research recommendation from an order and an order from actual administration. If a count rises again, the record needs the resident evidence, decision owner, rationale, and next review—not an assumption that the pathway failed or that returning to baseline was appropriate.
Carry the trial's limits into any local pathway discussion
The investigators planned for 220 participants and randomized 185. Randomizing people within the same units and homes created a risk that teams applied pathway ideas to usual-care participants. The settings were affiliated with academic centers, follow-up lasted 12 weeks, and the intervention included weekly research-team review. The study recruited 96 participants during the COVID-19 period, when several assessments were moved online or suspended. Those conditions limit transfer to a small U.S. consultant pharmacy practice.
This paper should not be converted into a resident-specific treatment protocol, a claim of safer care, or evidence of compliance with U.S. nursing-home psychotropic and gradual-dose-reduction requirements. Its useful challenge is narrower: decide the endpoint before reviewing the graph, define exactly which medicines count, and keep the clinical response beside the medication trajectory. If the early gain disappears, that is part of the result—not a footnote to hide.