Poster P5174 — at the board

Daily monitoring of BPSD · IAGG 2026 · johannes.malm@ju.se

30-second pitch

This is a mixed methods pilot on whether care staff can rate dementia symptoms every day instead of every six months, which is the current standard in Sweden. BPSD fluctuate day to day, but the standard registry misses that. Staff at four nursing-home units used a simple web app to rate the 12 NPI domains up to three times a day. Over 90 days they completed about 19,000 ratings, with 75% adherence, and found it meaningful and quick. The interesting part: the barriers were organizational and interpretive, not technical.

2-min walkthrough · point as you talk

  1. Problem: up to 90% of people with dementia have BPSD, they swing daily, but assessment happens every 6 months. Important change goes undocumented.
  2. What we did: convergent mixed methods, 90 days, 4 units, 8 residents, 44 staff. Web app "Daily-BPSD", 12 NPI-NH domains, 0–3, up to 3×/day.
  3. Findings (the 3 figures): 75% adherence, uneven (B4 lowest) · time fell 77→46 s, faster with use · prevalence 7.5–38%. Middle heatmap = day-to-day variation is visible only because data were daily.
  4. Qualitative: quick, fit the end-of-shift routine, raised awareness & confidence. Barriers: time pressure, uneven participation, interpretive uncertainty.
  5. Where it leads (bottom-right, your strongest card): now running at scale — 230 residents, 5 organizations, >1M registrations. The pilot held.

Numbers cold

NPI scale (0–3 severity)

0Absent — not seen during the shift 1Mild — noticeable, not disruptive 2Moderate — clear, some effort to handle 3Severe — marked, disruptive / strong impact

If pressed: classic NPI-NH scores frequency (1–4) × severity (1–3). We deliberately simplified to a daily 0–3 so staff can do it 3×/day — a trade-off for feasibility.

Q&A · tap to open

n=8, isn't that too small?
It's a feasibility pilot, so the unit of interest is the registration and the routine, not the resident. We have 19,000 ratings and 1,600 shift registrations. And the scale-up with 230 residents confirms the routine holds.
75% adherence, is that good?
There's no clean benchmark, daily BPSD documentation hasn't been done. For a voluntary routine sustained at 75% over 90 days, we read that as encouraging. The spread matters more than the mean, B4 shows what happens when a unit's routine isn't stable.
Different staff rate the same resident, reliability?
(The sharpest one, own the caveat.)
That's exactly the interpretive challenge we found, overlapping symptoms are hard to separate. We didn't test formal inter-rater reliability in the pilot; the in-app definitions were there to support consistency, and staff said they helped. It's a key question for the scale-up.
Did daily rating improve care / outcomes?
We didn't measure clinical outcomes, this was about feasibility. What staff reported was increased awareness and more confident team conversations. Whether that translates to outcomes is the next study.
Why both Bowen and NASSS?
Anchor: Bowen measures, NASSS explains.
They answer different questions. Bowen structures the feasibility outcomes — whether the routine works in practice. But feasibility doesn't tell you why a unit succeeds or struggles, so we used NASSS as an interpretive lens for the interviews, it's built for technology adoption in care. Bowen tells us that it's feasible; NASSS tells us why. That's how we landed on: barriers are organizational and interpretive, not technical.
Isn't adherence just clicks?
Turn the weakness into the reason for mixed methods.
You're right that adherence is only a proxy — someone could register without really integrating it. That's exactly why we paired it with the interviews, to check whether the routine was real, not just the clicks.
What's the tool, is it commercial?
A purpose-built web app for this study.
Grey cells in the heatmap?
No registration that day. Missing data is itself a signal, it maps onto the low-adherence units.

Nuance moves (deploy on purpose)

Feasible → sustainable arc: We showed it's feasible. The open question is whether it's sustainable — and the scale-up now suggests it is.

Granularity: daily data give you granularity the six-month registry can't capture.

Proxy caveat: adherence is a proxy for integration into routine — not a perfect one, which is why it's mixed methods.

Words to reach for · EN · SV · example

feasibility fii-zi-BIL-i-ti
genomförbarhet
"The study was about feasibility, not effectiveness."
adherence ad-HEER-ens
följsamhet
"Adherence to the registration routine averaged 75%."
to fluctuate / fluctuation
svänga, variera / fluktuation
"BPSD fluctuate from day to day, and daily data capture that."
uptake
upptag, att personalen nappar på det
"We saw good uptake among staff from the first weeks."
sustained / sustainable suh-STEIN
ihållande / hållbar över tid
"Adherence was sustained over 90 days, and the scale-up suggests it's sustainable."
sociotechnical
sociotekniskt
"The barriers were sociotechnical, about people and routines, not the technology."
interpretive
tolkningsmässig
"There was interpretive uncertainty when symptoms overlapped."
overlapping
överlappande
"Overlapping symptoms are hard to separate on a single scale."
burden
börda, arbetsbelastning
"The routine added little burden on an already busy shift."
embedded / to embed
inbäddad, förankrad
"It became embedded in the end-of-shift routine."
prospectively
prospektivt, i förväg
"We defined the feasibility outcomes prospectively, using Bowen."
to capture
fånga, fånga upp
"Daily data capture fluctuation that a six-month registry misses."
granularity gran-ju-LA-ri-ti
upplösning, detaljnivå
"Daily rating gives a granularity you can't get every six months."
real-world
i verklig vård, i praktiken
"This is feasible in real-world dementia care, not just in a trial."
complementary
kompletterande
"Bowen and NASSS are complementary, not competing."
to underpin
underbygga, ligga till grund för
"The same routine underpins the larger ongoing study."
scalable / to scale
skalbar / skala upp
"The routine scales, it now runs across 230 residents."
trade-off
avvägning, kompromiss
"There's a trade-off: we simplified the NPI to a daily 0–3 so staff could do it often."
proxy PROK-si
ställföreträdande mått
"Adherence is a proxy for how well the routine is integrated."
signal
signal (mots. brus/noise)
"Missing data is itself a signal, not just noise."

Close with this every time

Daily BPSD observation is feasible in real-world dementia care. The hard parts are organizational and interpretive, not technical — and it scales.