BPSD — assessment, non-drug care and antipsychotic risk (MEQ)

(i) Assessment (DICE/ABC). Describe: exact behaviours (resisting care, one strike, nocturnal calling), timing, who is present, frequency/severity, and antecedents/consequences (ABC chart for showering and night shifts). Investigate: delirium screen (attention, fluctuation, acuity over 10 days); infection; constipation (already risk); urinary retention; pain from osteoarthritis; medication changes; sensory impairment; environment (noise, lighting, staffing); carer approach and distress; DLB features (visual hallucinations, fluctuations, parkinsonism, RBD). Create: person-centred plan before drugs. Evaluate: define measurable outcomes (episodes of aggression/week, sleep, carer injury). Capacity/substitute decision-maker and least-restrictive care framework (local statute). [1]

(ii) Differentials. Delirium (acute 10-day change — must exclude); pain-driven agitation; constipation/retention; depression with irritability; psychosis of AD; primary anxiety; medication toxicity/withdrawal; environmental mismatch; elder abuse (always consider). Discriminators: attention/fluctuation, exam for abdomen/bladder/joints, vitals and basic labs, collateral on hallucinations/parkinsonism, medication chart review. [1]

(iii) Non-drug and pain-first plan. Warm room, two-staff personal care, explain each step, preferred music, consistent carers, daytime activity and light, night-time noise reduction, scheduled toileting, treat constipation aggressively. Stepwise analgesia for osteoarthritis pain (e.g. regular paracetamol if appropriate) — Husebo cluster RCT supports treating pain to reduce behavioural disturbance. Staff training and carer support. Document target behaviours and review in days, not months. [1][5]

(iv) If antipsychotic needed. Only if severe ongoing risk of harm after medical and non-drug optimisation. Example: risperidone 0.25–0.5 mg orally daily (or divided), lowest effective dose, short intended course. Discuss increased mortality (Schneider meta-analysis of RCTs) and stroke/cerebrovascular risk, sedation, falls, EPS, metabolic effects; local product-information context for limited Alzheimer behavioural use does not remove black-box harms. Monitor response, BP, falls, EPS, sedation; ECG if cardiac risk. Review within 1–2 weeks; plan taper when stable, with monitoring for relapse (Devanand). Not "long-term by default." [2][6]

(v) CATIE-AD and DART-AD counselling. CATIE-AD: atypical antipsychotics have limited effectiveness for psychosis/aggression/agitation in AD with high discontinuation for intolerability/inefficacy — set realistic expectations. DART-AD long-term follow-up: continued neuroleptic treatment associated with higher mortality signal — reinforces time-limited prescribing and structured deprescribing rather than indefinite "settling" medication. [3][4]