Impact of the Covid‐19 pandemic on neuropsychiatric symptoms and antipsychotic prescribing for people with dementia in nursing home settings

Abstract Objectives This study aimed to determine the impact of the Covid‐19 pandemic on neuropsychiatric symptoms and antipsychotic use in people with dementia living in nursing homes. Methods This was a comparative analysis of baseline data from two large nursing home studies, one conducted during (COVID‐iWHELD study) and one prior (WHELD study) to the pandemic. It involves data from 69 and 149 nursing homes, and 1006 and 666 participants respectively. Participants were people with established dementia (score >1 on Clinical Dementia Rating Scale). Resident data included demographics, antipsychotic prescriptions and neuropsychiatric symptoms using the Neuropsychiatric Inventory Nursing Home version. Nursing home data collected were nursing home size and staffing information. Results Overall prevalence of neuropsychiatric symptoms was unchanged from pre‐pandemic prevalence. Mean antipsychotic use across the sample was 32.0%, increased from 18% pre‐pandemic (Fisher's exact test p < 0.0001). At a nursing home level, the medians for the low, medium and high tertiles for antipsychotic use were 7%, 20% and 59% respectively, showing a disproportionate rise in tertile three. Residents in these homes also showed a small but significant increase in agitation. Conclusion There has been a significant increase in antipsychotic prescribing in nursing homes since the COVID‐19 pandemic, with a disproportionate rise in one third of homes, where median prescription rates for antipsychotics were almost 60%. Strategies are urgently needed to identify these nursing homes and introduce pro‐active support to bring antipsychotic prescription rates back to pre‐pandemic levels.


| INTRODUCTION
The unnecessary use of antipsychotic medications in people with dementia has been a major issue in research, policy and practice over recent decades. Historically used widely to manage the neuropsychiatric symptoms of dementia such as agitation, aggression and psychosis, atypical antipsychotics confer only modest clinical benefit when used in the short-term. Meta-analyses of clinical trials indicate borderline statistical significance for clinical benefit to aggression following treatment with risperidone, 1 but not for agitation or psychosis. The benefit is even less clear with other antipsychotics such as quetiapine and importantly, many trials reported high placebo response which further limits the benefits reported. 2 This marginal benefit is even less favourable when considered in the context of the well-established harms associated with these medications. A review of 15 trials in people with AD showed a significant mortality risk following treatment with these medications, and studies focusing on the withdrawal of atypical antipsychotics have reported a reduction in this risk following halting of prescriptions. 3 Additional important adverse effects include a clinically significant increase in cognitive decline, a three-fold increase in cerebrovascular events such as stroke, as well as extra-pyramidal symptoms, peripheral oedema, sedation, prolonged Q-Tc interval, infections and abnormal gait. 4 These effects worsen with longer-term use, and trials of antipsychotics lasting 6 months or more show mortality rates of up to 59%. 2,5,6 These harms are particularly prevalent in nursing home settings, where many people have late-stage dementia with complex co-morbidities, including a high prevalence of neuropsychiatric symptoms. Prescription rates of antipsychotics for people with dementia in nursing homes exceeded 45% in the early 2000s. 7 This led to high profile initiatives across the US, Europe and elsewhere which have successfully reduced antipsychotic prescribing to less than 20% of nursing home residents, 8,9 as part of programmes internationally leading to a reduction of the prescription of antipsychotics for people with dementia. For example, in the UK antipsychotic prescription for people with dementia was reduced by 50% nationally between 2009 and 2015. 10,11 These reductions were achieved by clinicians adopting updated prescribing and treatment guidelines, 12,13 combined with an increased focus on person-centred care and use of non-drug psychosocial interventions as a first-line measure. 7,9 The latter not only reduce antipsychotic use but are also associated with benefit to quality of life, neuropsychiatric symptoms, apathy and pain. 7,9 The Covid-19 pandemic represents a considerable potential risk to the positive changes that were achieved prior to 2020. During the pandemic, social care systems worldwide experienced an unprecedented level of burden amongst frail, vulnerable residents. There were severe shortages of staff due to ill health and the social restrictions and society-wide lockdowns prevented families from visiting and had marked effects on residents. 14,15 Preliminary studies have reported accelerated cognitive decline and an increase in depression, anxiety and neuropsychiatric symptoms in people with dementia in the community during the pandemic. [16][17][18] In addition, one study reported a 10% rise in antipsychotic use from March to July 2020 in people with dementia and this observation has been supported by further reports. 19 However, there has been very little focus on people with dementia living in nursing homes. 20,21 This leaves key questions unanswered, such as the impact of the pandemic on neuropsychiatric symptoms and antipsychotic prescribing amongst these individuals. 22 The potential adverse effects associated with atypical antipsychotics may be of particular concern in the context of the COVID pandemic, where mortality and morbidity rates are already increased. Here we report an analysis of neuropsychiatric symptoms and antipsychotic prescribing rates across 149 UK nursing homes during the Covid-19 pandemic and compare the findings with a baseline dataset collected for a clinical trial conducted prior to the pandemic. The two datasets were collected using closely similar methodologies and settings, thus allowing for a unique insight into the impact of the pandemic on these key indicators of health and wellbeing in nursing homes.

| RESULTS
Demographic information about both study samples is shown in Table 1. Antipsychotic data were available for 3794 residents with an ICD 10 diagnosis of dementia based on nursing home records. 1060 participants were approached, for whom legal representatives gave consent for 707 individuals. 40 of these individuals withdrew before baseline assessment, and data was incomplete, giving a cohort of 666 individuals up to December 2021, who were the analysis cohort for the current study. More detailed data were collected for the 666 residents who were registered as iWHELD study participants. This sample of 666 residents was 69.5% female with a mean age of 85.5 (SD 8.0). In this group the overall prevalence of delusions (20.3%), hallucinations (15.3%) and agitation (46.6%) was consistent with prepandemic levels ( Table 2). Mean antipsychotic frequency in the whole nursing home population was 33% (n = 1252) amongst all residents with dementia in the participating nursing homes, and 32.0% in the iWHELD study sample (n = 213). This represents a significant increase from 18% (n = 181) in the WHELD sample pre-pandemic (Fisher's Exact Test p < 0.0001, Table 3). Antipsychotics were pre-  highlighting a disproportionate increase in at least one-third of nursing homes. Comparatively, in the pre-pandemic study all homes had prescribing rates of six to 33%, with only two outliers with rates of 40% and 52% respectively, indicating a far lower overall prescribing rate and less variability pre-pandemic.
To further investigate this substantial rise in antipsychotic prescriptions in tertile 2 and 3, we examined the potential associations with nursing home factors and neuropsychiatric symptoms. There were no significant differences in nursing home size, the number of residents with dementia, sick days or use of temporary staff across the three groups ( Table 3). The ANOVA showed a significant but modest elevation in levels of agitation and delusions in nursing homes with higher antipsychotic use (F = 3.23; p = 0.04 and F = 4.09; p = 0.017 respectively) ( Table 2).

| DISCUSSION
This study provides the first direct insight into the impact of the  psychosocial interventions in nursing homes report improvements to staff burden, perceived skills and confidence following structured training, 27 and it is likely that these benefits contributed to an increased resilience across the majority of the workforce during the challenges imposed by the pandemic.
However, in the remaining 32% of nursing homes where antipsychotic prescribing rose to 59% there is a considerable challenge moving forward. The reasons for the increase in prescribing within these settings is not clear. There was no significant association with key nursing home characteristics such as home size, resident numbers, staffing structure or staff absences. The analysis did however show a modest increase in neuropsychiatric symptoms in residents in these homes, particularly in agitation and delusions. It is likely that this uptick in symptoms resulted, in part, in the increased prescribing rate, although the reason for symptom emergence and worsening is itself unclear. Whilst it is reassuring that the main increase in prescribing occurred in people with clinically significant neuropsychiatric symptoms, the limitations exerted by the pandemic on provision of alternative interventions, combined with the challenges of monitoring symptoms during this time may have led to more rapid prescribing in these groups and the training and support processes available in participating care homes were not captured and so no conclusions can be drawn on the influence of existing staff training and knowledge in these homes. However, it will be essential to proactively identify these nursing homes that have experienced an increase in prescribing and ensure they are fully supported in their institutional recovery from the pandemic, including ensuring evidence-based training and support is fully available in these settings. Programmes such as the evidence-based WHELD training package will be valuable for embedding this practice across settings.
These findings also further emphasise the need for a proactive strategy to revisit antipsychotic prescribing guidance in order to prevent resurgence of use back to pre-pandemic levels. These efforts will be essential to ensure health and wellbeing outcomes for people with dementia stabilise and continue to improve, and to avoid any polarisation in care quality across the social care sector.
This study provides a unique insight into the impact of the Overall the outputs of this study have considerable import for shaping the future of policy in this field, as well as highlighting the strengths of resilience and care practice that are already in place within the care sector. It will be critical to build on these aspects to ensure any discrepancies in resource and training availability do not drive differential trends within the nursing home sector, and to build policy to ensure all homes are supported in providing high quality care to achieve and maintain low levels of antipsychotic prescribing. Jeffrey Cummings has provided consultation to Acadia, Alkahest,