The impact of auditory hallucinations on “living well” with dementia: Findings from the IDEAL programme

To determine whether auditory hallucinations in community‐dwelling people with dementia (PwD) living in the community impacted on quality of life (QoL), subjective wellbeing and life satisfaction.

� The associations remained significant for QoL, wellbeing and life satisfaction after controlling for cognition, depression, antipsychotic prescription, and caregiver stress � The finding that auditory hallucinations are important for people with dementia to live well provides a focused treatment target when developing therapies for psychotic symptoms in people with dementia

| INTRODUCTION
There are 45 million people living with dementia worldwide. 1 Whilst dementia is commonly thought of as a generalized disorder of cognition and functional impairment, 2,3 neuropsychiatric symptoms are nearly universal, and often include psychosis. 4 Psychotic symptoms are frequently distressing to the individual and those in close contact, including the caregiver, 5 and are associated with poorer disease outcomes, including accelerated cognitive decline, 6 more rapid progression of functional impairment, 7 increased hospital admissions, 8 earlier admission to institutional care 7 and increased mortality. 7,9 In addition, psychotic symptoms are often antecedent to or co-morbid with other neuropsychiatric symptoms such as agitation, aggression and depression, 10 further adding to the impact on the individual and others. It is estimated that up to 50% of people with dementia (PwD) will experience a psychotic symptomdelusions, hallucinations or both-during the course of the disease. 11 Auditory hallucinations have received far less attention than visual hallucinations or delusions, but nevertheless have a prevalence of up to 12% in PwD, and prevalence is higher in people with Lewy body dementia. 12,13 Understanding the potential impact of auditory hallucinations on quality of life (QoL) and elucidating their importance as a specific treatment target is therefore an important gap in our current knowledge. QoL has been defined as describing individuals' perceptions of their position in life in the context of the culture and value systems in which they live and in relation to their goals, expectations, standards and concerns. 14 Dementia, and symptoms resulting from dementia, have a key impact on a person's lived experience and perceived QoL. [15][16][17][18] Relatively few studies have examined the impact of psychotic symptoms such as delusions and hallucinations on QoL. [19][20][21][22] The few studies that have been conducted have concentrated on community-dwelling PwD with smaller number of participants and possibly as a consequence have generally not identified associations between hallucinations and impaired QoL. 21,22 Whilst QoL measures have been most frequently used in evaluating subjective life experience in the presence of dementia, the inclusion of perceived wellbeing and life satisfaction may yield a more complete understanding of a person's ability to "live well." 23,24 Subjective wellbeing represents the appropriate balance of positive and negative emotions ascribed to one's current situation, and life satisfaction reflects a global evaluation of one's life up to the present circumstances. 16,25 Although QoL, subjective wellbeing and life satisfaction are interconnected, they include unique elements that examine different facets of "living well." 16,26 Living well with chronic illness and disability has been defined as the best achievable health state that encompasses all dimensions of physical, mental and social wellbeing. 16,27 The specific effects of auditory hallucinations on different facets of a person's ability to live well have not yet been determined.
The presence of other symptoms and features of dementia may serve as a possible confound to any interpretations associating auditory hallucinations with living well. Symptoms commonly related to dementia such as cognitive impairment and depression, and connected features such as antipsychotic prescriptions and caregiver stress, have been associated with lower QoL in PwD. 15,17,18,26,28,29 It is important to delineate the contributions of these factors when assessing the impact of psychotic symptoms in dementia.
The primary aim of this study was to determine the impact of auditory hallucinations on living well in PwD living in the community.
We also aimed to elucidate the role of various possible confounds, such as depression, cognition, whether antipsychotic drugs were prescribed and caregiver stress, when evaluating the impact of auditory hallucinations on these outcomes.

| Study population
The study used data from the Improving the experience of Dementia and Enhancing Active Life (IDEAL) study, a longitudinal cohort study conducted across England, Scotland and Wales to investigate social, psychological and economic factors that may enhance the capacity to "live well" with dementia. 26,29 The IDEAL cohort at baseline For each person with dementia a family member or friend, subsequently referred to as a caregiver, who provided practical or emotional unpaid support was also invited to take part where available and willing. Informed consent was obtained from all PwD and caregivers. Researchers visited PwD and their caregivers in CHOI ET AL. their homes to conduct structured interviews. The caregivers were asked to complete questionnaires separately, and both acted as informants regarding the person with dementia and provided information about their own experiences. Data from PwD were included in the present study only if there was an accompanying caregiver to form a participant dyad and if the caregiver answered the question on auditory hallucinations from the Neuropsychiatric Inventory Questionnaire (NPI-Q). 31,32 The IDEAL study was approved by the Wales

| Measures
Auditory hallucinations: Caregivers completed the NPI-Q and were asked whether the person with dementia was experiencing symptoms in 12 discrete behavioral domains. 31 For the purpose of the present study, we only used the domain for auditory hallucinations.
Specifically, presence of auditory hallucinations was defined as a caregiver's positive response to the question "Does your relative/ friend act as if s/he hears voices? Does s/he talk to people who are not there?" 32 Living well: Three self-report measures were completed by PwD to assess living well with dementia. These were the QoL in Alzheimer's disease scale (QoL-AD) 33 total score, the World Health Organization-Five Well-Being Index (WHO-5) 34,35 percentage score, and Satisfaction with Life Scale (SwLS) 25 total score. For each measure, a higher score indicates greater capability to "live well"; QoL-AD range 13-52, WHO-5 range 0-100, and SwLS range 5-35. For convenience the three measures together will be referred to as "living well" measures.
Cognition, mood and caregiver stress: The following additional measures from IDEAL were included in this analysis. The MMSE was used to measure cognition in PwD, with higher scores indicative of better cognition. 30 The Geriatric Depression Scale-10 (GDS-10) was used to measure depression in PwD, with higher scores indicating more self-rated depressive symptoms. 36 The Relative Stress Scale (RSS) measured the level of self-reported caregiver stress; possible scores range from 0 to 60 with higher scores indicating greater caregiver stress. 37 Demographic data: Information on age and sex were obtained from PwD and caregivers. Additionally, information on education was obtained from PwD. Dementia diagnosis was provided by researchers from medical notes. Information about whether antipsychotic drugs were prescribed was obtained from a joint interview with the dyad using a pre-defined list of medications. Educational level was classified into four groups: no qualifications; school leaving certificate at age 16; school leaving certificate at age 18; university. Relationship with the participant was classified into two groups: spouse/partner and other. Whether antipsychotic drugs were prescribed was coded as a binary yes/no variable.

| Planned analyses
All statistical analyses were carried out using SPSS Statistics v26. hence the total number of dyads included in the analysis was 1251.

| Impact of auditory hallucination on living well
The first research question focused on whether there was an association between the presence of informant-rated auditory hallucinations and living well in PwD. Auditory hallucinations were significantly associated with lower scores for QoL, WHO-5, and SwLS, although effect sizes were small; see Table 2.
The second research question asked whether the difference between informant-rated auditory hallucinations and living well could be explained by cognition, depression, whether antipsychotic drugs were prescribed, or caregiver stress. ANCOVA suggested that for QoL-AD scores the main effect of auditory hallucinations remained statistically significant after controlling for a priori confounding variables, but effect sizes were small; see Table 3.
ANCOVA also indicated that for WHO-5 scores the association between auditory hallucination and wellbeing remained statistically significant. The ANCOVA did not find a statistically significant main effect of auditory hallucinations on life satisfaction scores.
The results show that the association between auditory hallucinations and reduced QoL and wellbeing remained significant after controlling for a priori covariates but not life satisfaction.
Depression was a significant confounding variable with large effect sizes.

| DISCUSSION
The present study examined the associations of auditory hallucinations with self-reported QoL, wellbeing and life satisfaction in community-dwelling PwD. To our knowledge, this is the first study to use data from a large cohort to examine the relationship between auditory hallucinations and living well in community-dwelling PwD. Auditory hallucinations were associated with significantly lower scores on measures of living well, and these findings remained significant for QoL and wellbeing measures after adjusting for cognitive impairment, depression, antipsychotic usage and caregiver stress, confounding factors known to affect QoL in PwD. 15,17,18,28,29 The prevalence of auditory hallucinations in this cohort (7.7%) was similar to that found in previous literature 11,13 which suggests that prevalence of auditory hallucinations in community-dwelling people with mild-to-moderate dementia is likely to be around 10%. We found prevalence to be higher in PwD with Lewy bodies and Parkinson's disease dementia, which is consistent with previous research. 12,13 No significant associations were detected for prescribed antipsychotics or cognitive impairment. Importantly, relatively few PwD in this cohort were prescribed antipsychotic medication (2.1%) as compared to previous studies that had detected an association between poor QoL and antipsychotic prescription, 15,19 and therefore the statistical power to detect associations with antipsychotic usage was low. Caregiver stress had a small effect on QoL for the person with dementia, but not wellbeing and life satisfaction, which is consistent with previous literature where caregiver stress has a stronger association with informant-rated living well rather than selfrated living well in PwD. 17 As highlighted in previous literature, depression is a frequent comorbidity in PwD with psychosis. 10 In the current study it was a key covariate alongside auditory hallucinations and associated with lower QoL, wellbeing and life satisfaction in PwD. Furthermore, depression had the largest effect size on all three living well measures of any of the co-variates. This result aligns with previous studies, in which depression has been a common predictor of poor QoL. 15,17,18,28,29 Importantly, the large effect size of depression detected in our analysis along with evidence from previous literature may indicate a potential association chain with depression having substantial  "Improving the experience of Dementia and Enhancing Active Life: living well with dementia. The IDEAL study" was funded jointly by the Economic and Social Research Council (ESRC) and the National