# Tailored interventions for inappropriate psychotropic drug use in nursing home residents with dementia: participatory action research in a special case of a stepped-wedge cluster randomized controlled trial

## TL;DR

This study tested a tailored intervention to reduce inappropriate use of psychotropic drugs in dementia patients in nursing homes, finding some reduction in drug use but not in inappropriate prescribing.

## Contribution

The study introduces a novel implementation strategy combining participatory action research and tailored coaching in a stepped-wedge trial design for psychotropic drug use in dementia.

## Key findings

- No significant reduction in inappropriate psychotropic drug use was observed at 8 or 16 months.
- Overall psychotropic drug use decreased significantly at 16 months, especially for anxiolytics and antidepressants.
- Professionals prioritized implementing alternatives over improving prescribing quality.

## Abstract

Psychotropic drugs are modestly effective and may cause adverse effects. Efforts to reduce inappropriateness and increase usage of psychosocial interventions often suffer from suboptimal implementation. The purpose of this study was to evaluate effectiveness of an innovative study using implementation promoting elements in nursing home residents with dementia and neuropsychiatric symptoms.

A multicenter cluster randomized controlled trial with a special case of a stepped-wedge design with two arms and one stap was designed. The intervention comprised participatory action research, tailored information provision and external coaching, leading to the implementation of tailored action and implementation plans. The primary outcome was inappropriateness of psychotropic drug use (Appropriate Psychotropic Drug Use in Dementia [APID] index) and the secondary outcome was percentage of psychotropic drug use at baseline, 8 months, and 16 months. Homes were allocated to start with usual care or the intervention. After 8 months, the control group crossed over to receive the intervention. The other homes continued the intervention to 16 months. Patients were eligible if they were diagnosed with dementia, had a life expectancy of at least 3 months, and resided in psychogeriatric units.

An adjusted multilevel model revealed no effect on the APID index sum score at 8 months (0.564; 95% confidence interval [CI], -2.449–3.577; p = 0.71) or 16 months (2.165; 95% CI, -1.113–5.443; p = 0.20). An adjusted generalized estimation equation (GEE) model showed an effect at 16 months for percentage of use (OR 0.654; 95% CI, 0.481–0.889; p = 0.007). Adjusted GEE models showed an effect especially at 16 months for anxiolytics (OR 0.573; 95% CI, 0.382–0.859; p = 0.007) and antidepressants (OR 0.678; 95% CI, 0.475–0.968; p = 0.033).

No reduction of inappropriateness was found although overall usage was reduced. Professionals focused on implementing alternatives to compensate for usage, rather than prescribing quality. Future studies may focus on changing physicians’ prescribing behaviors in combination with multicomponent and multidisciplinary psychosocial alternatives.

Netherlands Trial Registry (NTR5872) on 27/05/2016, https://onderzoekmetmensen.nl/nl/node/26060/pdf.

The online version contains supplementary material available at 10.1186/s12877-025-06206-y.