The Outcomes of a Person-centered, Non-Pharmacological Intervention in Reducing Behavioral and Psychiatric Symptoms in Residents With Advanced Dementia in Australian Rural Nursing Homes

Background There is limited best-practice evidence to address behavioral and psychiatric symptoms for those with advanced dementia in Australian rural nursing homes. This study aims to evaluate the outcomes of a person-centered, non-pharmacological dementia care model, ‘Harmony in the Bush’, based on the Progressively Lowered Stress Threshold principles and person-centered music in rural Australia.


Background
In Australia, 60-70% of the people residing in nursing homes have dementia [1,2] and about 70-90% of residents with dementia suffer from psychiatric or behavioral symptoms [1,3].These behaviors include agitation, mood dysregulation, and disturbed thoughts and perceptions, which pose a major challenge for the residents with advanced dementia and nursing home staff [4,5].Behavioral and psychiatric symptoms are associated with diminished quality of life of the residents and increased di culty associated with caregiving for the staff [5][6][7][8][9][10].However, nursing home staff in rural Australia have poor understanding of appropriate methods of management and lack of adequate resources to manage behavioral and psychiatric symptoms in dementia [11].There is an overwhelming need to build best-practice evidence to address the residents' behavioral and psychiatric problems, leading to improve patient outcomes and staff quality of life in rural nursing homes.
Evidence of non-pharmacological interventions is growing and have been recommended to be pursued at rst instance, rather than pharmacological treatments, in the management of behavioral and psychiatric symptoms of residents with dementia.Non-pharmacological interventions are recommended to be person-centered to truly mitigate behavioral and psychiatric symptoms in residents with advanced dementia [12].Person-centered dementia care accentuates well-being and quality of life as regarded by the individual with dementia and their family members and creating an environment that considers the person's past life and current status [13].However, there is shortage of staff education and adequate training in rural nursing homes to develop the skills, knowledge and con dence in the use of evidence-based non-pharmacological interventions in person-centered care [11].Staff turnover, challenges in recruiting and retaining skilled staff, lack of resources compared to metropolitan areas pose major challenges for rural nursing homes [9].
The Progressively Lowered Stress Threshold (PLST) is an established model that provides a framework for person-centered dementia care [6].The model suggests that behaviors are in response to internal and external stimuli that may arise from the complex interactions between the residents and the nursing home environment.For instance, the model assumes that residents with dementia have reduced ability to experience stimuli, such as hunger, pain or transient noise and the lower threshold to these stimuli will eventually manifest as behaviors.This model also assumes all stress-related behaviors have an underlying cause and the caregivers can be trained to address the individuals needs and/or modify the care environment to be more person-centered to eliminate or minimize behavioral and psychiatric symptoms [7].
Person-centered music intervention as a non-pharmacological intervention in nursing homes, has shown positive results [14].Person-centered music that incorporates the cultural differences, personal memories, and individual taste has shown to improve attention, ability to access remote memory and emotions [15].However, a recent study did not show statistically signi cant decline in behavioral and psychiatric symptoms after 6 weeks of person-centered music intervention [16].These interventions did not adopt PLST as a theoretical framework for staff education and employed music intervention as stand-alone intervention without a comprehensive person-centered care plan [17,18].As such, a comprehensive PLST model that incorporates person-centered care plan and music intervention has not been tested previously, especially for residents with advanced dementia living in rural nursing homes in Australia.
Harmony in the Bush is a co-designed dementia care model that integrates PLST based person-centered care plan and music (during rest time using MP3 players), for the development of a low stress, person-centered organizational culture.This paper reports the 4-week outcome of the intervention on residents' behavioral outcomes and staff stress in the nursing homes.It is hypothesized that this person-centered, non-pharmacological intervention will reduce behavioral and psychiatric symptoms of residents with advanced dementia with comparable reduction in caregiver stress in staff.

Ethics
Ethics approval for Harmony in the Bush study was granted by Southern Adelaide Clinical Human Research Ethics Committee, Australia [Project Number: 277.17].

Research design and setting
Harmony in the Bush is a quasi-experimental study conducted in ve rural nursing homes of Australia [South Australia (n = 2) and Queensland (n = 3)].The nursing homes were purposely selected, and their managers were approached by the investigators and the former expressed interest to participate.To re ect on the diversity of the participating nursing homes, two were privately owned; one was a public funded (state-run center); and one was a not-for-pro t facility.The fth facility was a not-for-pro t, partly funded by the government, Aboriginal Residential and Community Aged Care home.
The total number of residents living with dementia across the ve participating aged care facilities in this this study is approximately 170-180.Allowing for 10% attrition, the investigators have estimated that 70-80 residents from this cohort can ensure a statistical power of 0.80.This estimation is based on a recent meta-analysis on effect of music interventions on agitation, which reports an overall effect size (Cohen's d) of 0.61 calculated from 12 studies [19], and by adopting an alpha level of 0.05 and employing the algorithm [20].This study adheres to the CONSORT guidelines.

Participants
Participants included eligible residents with advanced dementia and staff from the participating nursing homes.Resident participants were recruited following receipt of informed consents to participate in the study by their legal guardians/power of attorney appointees.In each of the four nursing homes, 15-20 residents were selected as potential participants; but in one aged care home, which has a total of 15 residents, all residents with dementia were approached to participate.As a result, a total of 77 residents with advanced dementia were recruited as participants.Inclusion criteria for the residents were: (a) diagnosis of advanced dementia within the Diagnostic and Statistical Manual of Mental Health Disorders 5 (DSM-5) [21]; (b) a Standardized Mini Mental Status Examination (SMMSE) score of less than 24 [10]; (c) being ambulant; and (d) displaying dysfunctional behaviors.The residents who could not meet the inclusion criteria or had a medical condition that affects normative behavioral patterns were excluded.Nursing home staff, including senior management staff, registered nurses, enrolled nurses, and care staff, were approached through the facility managers to participate in this research, resulting in 87 staff participants after voluntary consent.

Data collection procedure and intervention
The study intervention included: One each of the cognitive, affective and conative losses associated with Alzheimer's disease and related disorders; Five non-cognitive symptoms associated with Alzheimer's disease and related disorders; and Four stressors that may potentiate non-cognitive All staff on duty were responsible for observing and recording the residents' behavior on each shift.At least two staff in each facility were appointed as change champions by senior managers and trained in the process to ensure sustainability of the research outcomes after the project completion.

Intervention and post evaluation phase (Four weeks)
A person-centered care plan and personalized preferred music playlists were provided to participants based on the PLST principles [22] (Table 1; Figure 1).

Person-centered music:
Individualized music forms the basis for the music intervention.A key aspect to the success of individualized music was to identify speci c music preferences including exact song/composition titles and performers.We used The Assessment of Personal Music Preference Protocol in the Evidence-Based Guideline of Individualized Music for Persons with Dementia to assist in the process of music selection for each of the participating residents.The individualized music songs were determined after discussion with the resident, the resident's family member or next of kin/legal guardian, and the staff, and produced onto a multimedia MP3 player for listening, for not more than 30 minutes during rest periods and at an appropriate time requested by the residents.Applicable music activities were personalized and dependent on the participants' preferences, and their appropriateness to the participants' capability, participation, and response.Staff continuously gauged activity and stimulation levels and adjusted music intervention.

Outcome measures Baseline measures
Standardized Mini Mental State Examination (SMMSE) [10], was used to assess orientation, memory, attention and calculation, language and visual construction in residents.Kimberley Indigenous Cognitive Assessment (KICA-Cog) developed by LoGiudice et al. [23], was administered for the Indigenous The 10 item version of The Barthel Index of Activities of Daily Living (ADL) [24], was administered to assess ADL of the Pain Assessment in Advanced Dementia (PAINAD) scale, a frequently used simple ve-item observational tool with a range of 0 to 10 [25], was administered to assess pain.The ve indicators were breathing, vocalization, facial expression, body language and Cornell Scale for Depression in Dementia (CSDD) [26], a tool used to speci cally assess signs and symptoms of a major depression in people with a dementia, which includes items concerning physical well-being, sleep, appetite and other vegetative symptoms.The Kimberley Indigenous Cognitive Assessment -Depression (KICA-Dep) [27], a culturally acceptable screening tool for depression among older Indigenous Australians living in remote areas, was alternatively administered for the Indigenous Baseline and post-intervention measures Cohen-Mans eld Agitation Inventory (CMAI) [28], is a validated tool that is being used in this study to assess level of agitation in residents with dementia.CMAI is found to be a valid and reliable tool for assessing agitation in individuals living in residential aged care facilities.CMAI was considered a useful tool that was easy to complete by the nursing staff and direct care workers [29].Cohen-Mans eld, Marx, and Rosenthal (1989) classi ed the manifestations of agitation into the following three syndromes: (a) aggressive behavior (e.g., hitting, kicking, cursing); (b) physical nonaggressive behavior (i.e., restlessness, pacing); and (c) verbally agitated behaviors (e.g., complaining, negativism, repetitious sentences).CMAI was administered consecutively for ve days pre and post intervention.A maximum of three assessments per day consistent with staff shifts (7.00-15.00,15. 15-23.15, 23.30 -7.15).An aggregate score that re ects the number of incidents per shift was calculated.
Caregiver Stress Inventory (CSI) [30], which is a 43-item, self-reporting tool developed to determine individual staff caregiver stress to incidents (behaviors) that occur in residents with dementia, is used for assessing staff stress levels.
Staff stress is measured as the response of individual staff members' experience to incidents that occur in day-to-day care of persons with dementia.Reponses are self-rated on a seven-point Likert-type scale (1=not stressful, 7=extremely stressful) based on current perceptions of stress.The tool consists of four subscales representing staff stress from aggressive behavior, inappropriate behavior, resident safety, and resource de ciency.The tool is valid for use with paid carers and informal carers and has been recently used in Australian residential care facilities [31].
The paper reports the outcomes related to resident behavior and caregiver stress, while other elements of the Harmony in the Bush study such as sleep and psychotropic medication use will be reported separately.We recently reported on the determinants of person-centered care in rural nursing homes using follow-up qualitative data from the Harmony in the Bush study [32].

Data analyses
All analysis of pre-and post-test data was analyzed using SPSS IBM Version 23.Statistical tests were considered signi cant at p<0.05.Descriptive data was checked for outliers and assumptions for parametric analyses.Paired-t test was used to analyze the difference between baseline and 4 weeks follow-up of CMAI and CSI scores.Cross tabulation and chisquare tests were used to analyze the factors associated with caregiver stress.Repeated measures ANOVA were used to understand difference in baseline and follow-up assessments to understand the changes in behavioral and psychological symptoms in residents after adjustments for age and gender.Similarly, repeated measures ANOVA were used to analyze changes in staff stress in caregiving measured using CSI adjusted for age and gender.Thirteen Focus group discussions and 38 staff interviews were thematically analyzed and coded for resident behavior and caregiver stress to allow triangulation of ndings.

Effectiveness of the intervention in reducing behavioral and psychiatric symptoms
Data was available for 74 residents with advanced dementia who took part in the intervention.Their demographic characteristics and baseline measures for the participants are described in Table 2.The mean age of the participants was 82.4 (7.7) years, and 68.9% were females.The mean age of admission of the residents was 80.1(8.4)years.The mean SMMSE score was 9.21 (6.7); 51.4% has SMMSE score less than 10, indicating advanced dementia.The PAINAD scale indicated that 23.0% had mild pain and 9.5% with moderate to severe pain.On the Cornell Scale for Depression in Dementia (CSDD), a signi cant 68.9% reported anxiety symptoms, 40.5% had mild-severe sadness, 41.9% reported mildsevere irritability and 46.0%exhibited mild-severe agitation.Barthel Index score of 58.4 (24.0) for Activities of Daily Living (ADL) was measured in residents with advanced dementia at the baseline.Table 3 illustrates the change in behaviors in residents with advanced dementia, pre-and -post-intervention.The common symptoms were pacing, aimless wandering (86.5%); general restlessness (74.3%); trying to get to a different place (63.5%); repetitive sentences (60.8%); cursing and verbal aggression (55.4%).Paired-analyses between baseline and postintervention observed signi cant decline in behavioral and psychological symptoms of dementia across domains, in resident participants (Table 4).The mean scores total CMAI declined from 3.05/shift to 1.35/shift after 4 weeks (paired t=8.5, <0.001).Analyses using repeated measure ANOVA indicated the decline in scores was signi cant after controlling for age and gender (F=6.3.p=0.015).Supplementary Table 1 provides extracts from qualitative analyses on the staff feedback on changes in behavioral outcomes for the residents in the study.

Effectiveness of the intervention in reducing staff stress in caregiving
Of the nursing home staff involved in the Harmony in the Bush intervention, 87 completed the Caregiver Stress Inventory questionnaires at baseline (n = 87) and 4 weeks post-intervention (n = 58) (Table 6).About 39 completed both baseline and follow-up measures.The changing work shifts, and staff turnover reduced the response rate for follow-up assessment.The mean age of staff was 44.9 (13.5) years, 84.1% were female.Among the staff who completed the CSI, 69.3% were personal carers, 18.2% were registered nurses, and 11.4% were enrolled nurses.More than one-third of the staff had country of origin other than Australia and the majority (59.8%) had work experience of more than 6 years in nursing homes.
The Caregiver Stress Inventory scores are presented in four domains i.e., aggressive behavior, inappropriate behavior, resident safety, and resource de ciency (Table 6).The mean score of Total CSI was 3.73 (0.92) at baseline.The subscale scores were aggressive behavior 4.35 (1.2); inappropriate behavior 3.3 (0.95); resident safety 4.2 (1.1) and resource de ciency 3.6 (1.2) at baseline.Table 5 explains the factors associated with Caregiver Stress Inventory scores.Female staff reported higher mean scores than males regarding stress related to inappropriate behavior (t=-2.3,p=0.02).The other signi cant difference was that staff whose country of birth was not Australia reported high stress scores in the domains of aggressive behaviors (t=-1.8,p=0.07) and resource de ciency (t=-2.3,p=0.02).The total CSI reduced from 3.6 (0.9) to 3.1 (1.1) between the two time points.Paired analyses of 39 staff showed consistent decline in staff stress in all subscales (Table 5).Repeated measures ANOVA indicated that the decline in caregiver stress was signi cant after controlling for age and gender (F=4.3,p=0.05).Supplementary Table 2 provides extracts from qualitative analyses on the staff feedback on changes in staff stress in caregiving in the study.

Discussion
The present study provides evidence that 'Harmony in the Bush' model results in the reduction of behavioral and psychiatric symptoms and reduces staff stress in caregiving.We found statistically signi cant decline in aggressive behaviors, physically non-aggressive behaviors, verbally agitated behavior and hiding and hoarding behaviors in the study sample after 4 weeks of intervention in persons with advanced dementia.There was comparable reduction in staff stress in the domains of aggressive behaviors, inappropriate behaviors, resident safety, and resource de ciency measured using Caregiver Stress Inventory tool.Thus, a person-centered, non-pharmacological intervention using PLST framework was effective in rural nursing home settings.
Best practices to improve dementia care in nursing homes are di cult to establish and comply with factors intrinsic to the resident and complexity of care relating to advanced dementia.Rural nursing home staff are highly stressed due to factors including long hours and heavy workload in managing the complex needs of the residents [9].Further, higher prevalence of behavioral and psychiatric symptoms in residents with dementia correlates with higher stress in the caregivers [33,34].Poor knowledge of behavioral and psychiatric symptoms increase staff stress in caregiving and affects the quality of care [11] .Inadequate training and knowledge can lead to dementia residents being subjected to stigmatization and unnecessary use of restraint [35].Our study included training for staff on behavior and psychiatric symptoms in dementia and the PLST framework for a person-centered care plan and music intervention.Studies have found that positive attitudes towards dementia and person-centered care correlated with job satisfaction [35,36].We anticipate staff training in personcentered care will increase quality of care to the residents, job satisfaction and retention in rural nursing homes.
Our study results were consistent with a previous study, which found staff education and care based on the PLST model was clinically effective in reduction of behavior problems of care recipients, decreased caregiver burden and depression of family members, and in increasing their quality of life [33].Similarly, Huang et al. [37] showed reduction in physically non-aggressive behavior, verbally aggressive and nonaggressive behavior subscales as well as the overall CMAI decreased signi cantly after three weeks of PLST intervention and improved self-e cacy of family caregivers.Our study adds to the knowledge base of PLST based interventions and illustrates the effectiveness of the PLST framework for a person-centered care intervention in reducing behavioral and psychiatric symptoms in residents with advanced dementia in rural Australian nursing home.
Our study also illustrates that implementation of individualised music as part of routine care by nursing and direct care workers provide a great avenue for person-centered care in rural nursing homes.While there is growing evidence of the positive impact of music in dementia care [19]; most are therapist-led music interventions, which are di cult to sustain in low-resourced rural nursing homes.Evidence that music therapy served advantages over non-therapist led listening is minimal [38,39].Several studies have found potential bene t of person-centered music intervention in dementia care.For example, listening to favorite music was effective in reducing agitation [40].Other studies have revealed the use of prerecorded personalized music can be effective in reducing, anxiety, depression, pain and emotional well-being including quality of life [17,38,41].Many of these studies have been criticized on methodological limitation and were suggested as low-quality evidence [42].Our intervention includes a person-centered care plan, reducing noise levels and creating a lowstress, person-centered environment for listening to music.The bene t of utilizing a pre-recorded music playlist intervention as an alternative to formal music therapy is that the pre-recorded individualized music playlist is easy to access and can be conducted by nursing and direct care workers as part of their daily care regime.
The present study explains a best practice model for dementia care in Australian rural nursing homes with a good sample of 77 participants which yielded sample power of 80%.The quasi-experimental pre-post design was appropriate and feasible in rural nursing homes and the measurement tools were reliable and validated.However, there are some inherent limitations in the study.The lack of a control group prevented comparison with treatment as normal practice.Additionally, we were not able to disassociate the effect related to environmental modi cation, attitude change or music intervention on patient and carer outcomes.Nevertheless, Harmony in the Bush model facilitates and provides a structure for staff education in personcentered practice to reduce the impact of behavioral and psychiatric symptoms of dementia among residents and improve staff wellbeing in rural Australia.Tables Table 1: Harmony in Bush Study: PLST principles and strategies.Supplementary Files

Principles
Baseline assessments (One week) Demographic information and baseline clinical data were collected for each of the residents.The demographic pro le of the residents included age, gender, education level and occupation.During this week, each participant was assessed using the SMMSE or Kimberley Indigenous Cognitive Assessment (KICA-Cog); Pain Assessment in Advanced Dementia (PAINAD) scale; Cornell Scale for Depression in Dementia (CSDD) or Kimberley Indigenous Cognitive Assessment-Depression (KICA-Dep); Barthel Index of Activities of Daily Living (ADL); Apathy Evaluation Scale (AES); and Cohen-Mans eld Agitation Inventory (CMAI).Education and development of person-centered care plan (Three weeks) An experienced nursing educator/investigator(s) conducted the staff training workshops focusing on the theory and delivery of the PLST, person-centered dementia care and the use of the measurement instruments listed in stage 1 -Baseline assessment.The Harmony in the Bush intervention entailed the introduction of person-centered care plan based on the following principles of the PLST theory: Introduce consistent individualized routines to compensate for conative losses; Organize small group activities to eliminate overwhelming stimuli; Allow residents to set their own sleep/wake cycle to prevent fatigue; Plan activities based on past experiences and practices considering present cognitive and functional abilities; and Eliminate misleading stimuli that trigger The expected learning outcome after completion of the training was that staff who attended training should be able to identify: of the intervention program.Acknowledgement also goes to research assistants at Flinders University Rural Health SA, for her valuable support in data collection.41. Park H: Effect of music on pain for home-dwelling persons with dementia.Pain management nursing : o cial journal of the American Society of Pain Management Nurses 2010, 11(3):141-147.42.van der Steen JT, Smaling HJ, van der Wouden JC, Bruinsma MS, Scholten RJ, Vink AC: Music-based therapeutic interventions for people with dementia.The Cochrane database of systematic reviews 2018, 7(7):Cd003477.
Strategies Maximize safe function by supporting losses in a person-centered manner Activity based on their ability and preference (previous interests) To set their own sleep and wake-up cycle as much as possible Limit choices and serve food one after another and preferred menu Food and drink stall Familiar routine; Scheduled afternoon nap time Assess internal stressors -pain, hunger Provide unconditional positive regard Polite language, address them with their names and treat them as mature adults Gentle touch Provide distraction instead of correction or argument Use anxiety and avoidance to gauge activity and stimulation levels Look for early signs of anxiety (toe tapping, dgety) and provide personalized music Provision of individualized music -preferred playlist Teach caregivers to observe and listen to patients Observe and listen to person's needs Attend to their needs or stressors (internal or external) Modify environment to support losses and enhance safety Reduction of noise level in the environment (transient noise due to TV) Reduce noise level in the kitchen and avoid dragging of devices on the oor Reduce outside noise (example mowing during rest time) Remove any hazards in the room and outside Reduce unnecessary lighting during sleep Provide ongoing education, support and problem solving On-going education; mentoring Building relationships among and between staff and residents

Table 2 :
Demographic characteristics of the

Table 3 :
Prevalence of behavioral and psychological symptoms in residents with dementia.

Table 4 :
Effectiveness of intervention on changed behaviors of residents with dementia.

Table 5 :
Factors associated with caregiver stress at baseline.

Table 6 :
Effectiveness of intervention on stress associated caregiving.
FiguresFigure1Harmony in the Bush personalized dementia care intervention.