This paper is a commentary on the paper by Smeets etal. (Reference Smeets2020), which reports results of the PROPER (Prescription Optimization of Psychotropic drugs in Elderly nuRsing home patients with dementia) II cluster-randomized controlled trial that aimed to establish the effectiveness of a multidisciplinary biannual medication review on improving the psychotropic medication prescribing practice in Dutch nursing homes (aged care homes hereafter). The paper reports the results for the secondary outcomes of the PROPER II trial: 1) the prevalence of psychotropic medication uses (e.g. antipsychotics, antidepressants, hypnotics, and anxiolytics) for treatment of behaviors and psychological symptoms of dementia (BPSD), and 2) the occurrence of BPSD, assessed at 0, 6, 12, and 18 months. The results suggest the PROPER intervention did not reduce the prevalence of psychotropic medication use, based on the review of the residents’ medical records or their electronic pharmacist information system. Instead, the prevalence increased in the intervention group, while it decreased in the control group over time, with an estimated difference of 3.9% per 6 months. No significant changes were observed for the occurrence of BPSD, using the Neuropsychiatric Inventory – Questionnaire, between the groups over time. Notably, albeit not statistically significantly, agitation and aggression, assessed with the Cohen-Mansfield Agitation Inventory (CMAI), was worse in the intervention than the control group (greater increase of CMAI scores with a small slope difference of 0.6 every 6 months) (Smeets etal., Reference Smeets2020).

The high prevalence of BPSD and their detrimental impact on aged care home residents living with dementia are widely known. BPSD is a term used to describe various clusters of behaviors and symptoms associated with dementia, for example, agitation, aggression, repetitive questioning, depression, apathy, psychosis, inappropriate behaviors, nighttime behaviors, and wandering. The exact mechanisms and causes are not fully understood, but they are often attributed to a number of factors including biomedical, environmental, and interactional origins. Management of BPSD should be preceded by the assessment of etiological factors (Kales, Gitlin & Lyketsos, Reference Kales, Gitlin and Lyketsos2015).

In the past 10 years, we have seen a growing number of studies that investigated and established the effectiveness of programs or tools in reducing the (mis)use of psychotropic medications for the management of BPSD, also known as neuropsychiatric symptoms, in aged care homes, for example, Reducing Use of Sedatives (RedUSe) trials on antipsychotic and benzodiazepine prevalence and dosage (Westbury etal., Reference Westbury, Tichelaar, Peterson, Gee and Jackson2011; Reference Westbury2018) and the Halting Antipsychotic use in Long Term care (HALT) study (Brodaty etal., Reference Brodaty2018) in Australia; the Well-Being and Health for People With Dementia Program in the UK (Ballard etal., Reference Ballard2016); and the CHemical Restraints avOidance MEthodology criteria study in Spain (Muniz etal., Reference Muniz2019). These studies followed numerous earlier reports of unnecessarily high prevalence of psychotropic medication use for the management of BPSD and the risk of serious adverse effects (e.g. death, stroke, gait disturbance, falls and fractures, faster cognitive decline, infections, pulmonary embolism, and deep vein thrombosis), with very little effect on the actual BPSD (Ballard & Corbett Reference Ballard and Corbett2020; Harrison etal., Reference Harrison, Cations, Jessop, Hilmer, Sawan and Brodaty2019). At the same time, the effectiveness of person-centered care and psychosocial therapies for the management of BPSD has become more widely accepted as a result of several large-scale trials and systematic reviews (e.g. Chenoweth etal., Reference Chenoweth2009; Harrison etal., Reference Harrison, Cations, Jessop, Hilmer, Sawan and Brodaty2019; Kales, Gitlin & Lyketsos, Reference Kales, Gitlin and Lyketsos2015).

The latest guidance for the management of BPSD (National Institute for Health and Care Excellence, NICE, guideline [NG97] 2018) recommends psychosocial and environmental interventions as first line of treatment, as well as for ongoing management of BPSD, and limits the use of psychotropics to exceptional circumstances only (e.g. person being at risk of harming themselves or others, BPSD causing the person severe distress). The NICE guideline (2018) also emphasizes the importance of involving the person with dementia and their family in discussing the benefits and harms of psychotropic medications and decision making when considering such pharmacological treatment for BPSD. The underlying principle for all those trials mentioned above, including the PROPER II trial, is that a high prevalence of psychotropic medication use in residents with BPSD is an indicator of poor clinical practice, so the goal of the intervention is to reduce the prevalence and encourage deprescribing of psychotropics when deemed unnecessary or inappropriate. In fact, most trials use the prevalence of psychotropic medication prescriptions as a primary outcome so that reduction of the prevalence indicates the positive/successful outcome, and the effectiveness of the program can be established.

The results of the PROPER II trial reported in the Smeets etal.’s (Reference Smeets2020) paper are therefore somewhat disappointing. The authors provide a number of theories as to what might have contributed to the undesired effects including the intervention’s narrow focus on the knowledge aspect of psychotropic medication reviews, without linking it to complex interplays between various types of BPSD and medication, and its lack of consideration of the attitudes and behaviors of clinicians and families as well as collaboration of all stakeholders involved.

Most critically, the PROPER II intervention did not address the contextual and systemic factors that influence the prescribing practice of clinicians (e.g. their behaviors, attitudes, and feelings) in aged care homes. Several barriers are associated with the implementation of interventions to reduce psychotropic prescribing in aged care homes, with a unique influence of organizational culture. Organizational culture is a complex phenomenon, underpinned by assumed beliefs, which profoundly influence health professionals’ decisions to use psychotropic medications (Sawan etal., Reference Sawan, Jeon, Fois and Chen2016; Sawan etal., Reference Sawan, Jeon, Fois and Chen2017; Sawan etal., Reference Sawan, Jeon and Chen2018). Aspects of organizational culture that impede the appropriate use of psychotropic medications include limited managerial support for and facilitation of measures to reduce psychotropic prescribing, staff feeling helpless to do the right thing due to perceptions of limited resources, and physicians feeling pressured to prescribe psychotropic medications in response to inadequate staff levels and training (Sawan etal., Reference Sawan, Jeon, Fois and Chen2016; Sawan etal., Reference Sawan, Jeon, Fois and Chen2017; Sawan etal., Reference Sawan, Jeon and Chen2018). There is a need for interventions seeking to change the behavior of aged care home staff and health professionals to evaluate and address the organizational culture of aged care homes to achieve reduced psychotropic prescribing.

This points to another drawback in PROPER II intervention design: the education program was only targeted at nurses, physicians, and pharmacists and focused on prescribing of psychotropic medications. The Australian HALT study (Brodaty etal., Reference Brodaty2018) reported the effect of a multicomponent intervention on the reduction of antipsychotic prescribing. The intervention comprised a comprehensive education program of nursing staff, physicians, and pharmacists about person-centered and non-pharmacological management of BPSD. The impact of the intervention was significant as the number of older adults on regular antipsychotics reduced by 82% over 12 months (Brodaty etal., Reference Brodaty2018). There is a need to involve all care staff in the monitoring of psychotropic medication and implementation of person-centered approaches to the management of BPSD (Harrison etal., Reference Harrison2020; Sawan etal., Reference Sawan, Jeon, Fois and Chen2017).

The PROPER II study reported that a component of the intervention included evaluation meetings of the medication review process with all stakeholders at 6 and 12 months. However, it is unclear as to what function of the medication review was evaluated and what key features were identified to improve follow-up medication reviews. Medication reviews alone have been shown to be effective in the short term but not sustainable (Thompson Coon etal., Reference Thompson Coon2014). The RedUSe trials (Westbury etal., Reference Westbury, Tichelaar, Peterson, Gee and Jackson2011; Reference Westbury2018), involving on-site and visiting staff education and pharmacists-led medication reviews in collaboration with a nurse, demonstrated success in reducing antipsychotic prescribing, but the sustained impact of related studies over longer periods (up to 12 months) was mixed (Westbury etal., Reference Westbury, Tichelaar, Peterson, Gee and Jackson2011). What is needed is a multi-pronged intervention that addresses specific organizational cultural aspects of aged care homes in order to achieve sustainable behavior and practice change.

There are also methodological issues that might have contributed to the results of the PROPER II trial. The authors admitted that, due to a high attrition rate in the intervention group, almost 50% (80/172) of newly admitted residents compared to 38% (57/151) of controls who had been added to the trial after the baseline assessment, might have not received any benefit from the medication review. The authors acknowledged that the changes in the case mix of the study participants due to those newly added residents might have contributed to the findings. However, this is difficult to judge as it seems that the newly admitted residents’ baseline data were not reported. Furthermore, the intervention group had more severe dementia and more Alzheimer’s disease than the control. None of these differences or case mix changes was taken into account in their analysis. Given that the rationale for the psychotropic medication prescriptions was treatment of BPSD, it would have been helpful to investigate the role of individual BPSD in the prevalence of psychotropic medication prescriptions. This could have been particularly useful because agitation and aggression (assessed with CMAI) increased more in the intervention group than the control group (Smeets etal., Reference Smeets2020).

One notable aspect of this paper (Smeets etal., Reference Smeets2020) is the conflict between its results and those of the authors’ earlier report on the primary outcome of the PROPER II trial (van der Spek etal., Reference van der Spek2018). The primary outcome was the appropriateness of psychotropic medication prescriptions for BPSD using the Appropriate Psychotropic drug use In Dementia (APID) index sum score, which showed a significant improvement for all types of psychotropic medication prescriptions in the intervention group compared with control (−5.28, P = 0.005). The authors concluded the PROPER program was successful in improving the appropriateness of the psychotropic medication prescribing practice in Dutch aged care homes. In their second paper published in this issue of International Psychogeriatrics, the authors suggest that these opposing results between the primary and secondary outcomes “illustrate that conscious decisions per individual drug do not necessarily lead to a reduction in the prescription of psychotropic drugs” (Smeets etal., Reference Smeets2020, p.xx).

However, this discrepancy requires a further examination. What is more important, the appropriateness or the rate of psychotropic prescription? One may argue that these contradictory results could be due to the way in which the outcome measures were administered. For example, it appears that measurement in the trial was not blinded. While the audit of prevalence and dose/duration of the psychotropic medication prescribing practice is less likely to cause bias, the assessment of the APID index sum score would have a degree of bias risk if the person carrying out the assessment was aware of the allocation of the intervention. In addition, we note that analysis of the APID scores was performed “assuming no systematic difference between groups at baseline due to the randomisation” (van der Spek etal., Reference van der Spek2018, p.434). This is an unusual way of examining the effectiveness of an intervention in a cluster trial with only 29 clusters, especially when the baseline APID scores for the two groups seem to have been quite different (van der Spek etal., Reference van der Spek2018) and there were reported systematic differences between the two groups (Smeets etal., Reference Smeets2020). Furthermore, the assessment did not include pro re nata prescriptions, which is a critical omission if these were used as replacement. Another question is does the APID truly assess the appropriateness of the psychotropic medications for the management of BPSD? The APID is partly based on Dutch drug formularies which do not allow for some psychotropic medications to be scored as inappropriate for therapy duration (van der Spek etal., Reference van der Spek2015). More importantly, the appropriateness of psychotropic medications for treatment of BPSD should also consider the availability of psychosocial and/or environmental programs. If these programs are not offered as first line of treatment, simply deprescribing the medications will not necessarily bring about desired outcomes for residents experiencing BPSD.

The PROPER II trial results highlight the complexities in reducing psychotropic medication prescriptions in aged care homes and measuring outcomes in the most valid and appropriate ways. There has been a continued emphasis from the scientific community on eliminating the overuse of psychotropic medications in residents with dementia in aged care homes, and frequent public outcries when severe adverse events or deaths are reported. It is critical we continue our scientific effort to build a rigorous implementation pathway to improve evidence-based practice in the management of BPSD in aged care homes.