Table 1.

Papers reporting family involvement (FI) intervention or impact of FI on resident BPSD with a quantitative or mixed method study design.

Authors	Method, approach and setting	N	Key FI domain, measures and time points	Key results	Quality rating and comments
Arai et al. (2021) (Japan) RQ: 2	Correlational (longitudinal) Investigated how social interaction affected BPSD among residents with dementia 10 long-term care facilities 1 region	Residents 312	Family: frequency of visits/contact Resident: activities of daily living (ADL); cognitive function (MFIS); BPSD (NPI-Q); BPSD severity; social interaction; activity participation, resident friendships, quality of family relationships, external contact Baseline, 12 months	Less communication with family associated with increased resident BPSD and BPSD severity. Severity stayed the same for those in frequent communication. 37% of residents had communication from family more than once per week	MMAT: **** Pos: Confounding factors and interactions accounted for, clear results reporting including effect sizes and potential biases Neg: Attrition rate
Bramble et al. (2011) (Australia) RQ: 1	CRCT (MM) Family involvement in care (FIC) intervention 2 long-term care facilities	Family 57 Staff 59	Family: knowledge (FKOD); stress (FPCR); satisfaction (FPCT) Staff: knowledge (SKOD); stress (SPCR; CSI); Attitudes toward family (AFC) Baseline, 1, 5 and 9 months	Sig increase in both family and staff knowledge of dementia, sig decrease in family satisfaction regarding staff consideration and management effectiveness	MMAT: **** Pos: Randomised sites, blinding, power and attrition aims Neg: Small sample, follow-up attrition, no variance reported
Brazil et al. (2018) (Northern Ireland, UK) RQ: 1	CRCT (MM) Evaluated effectiveness of an advance care planning intervention designed to assist families to participate in decision making 24 care homes	Family 197	Family: uncertainty in decision making (DCS); carer satisfaction (FPCS) including involvement related factors of support and communication Baseline, 6 months	Significant reduction in family carer uncertainty in decision making and improved family carer satisfaction in nursing home care. No impact on resident hospitalisations or number of deaths found	MMAT: **** Pos: Cluster randomisation of care homes, balancing confounding variables, variance reporting Neg: Size of care homes unknown, power unclear, no blinding
Chappell et al. (2014) (Canada) RQ: 2	Correlational (longitudinal) Examined predictors of change in social skills among residents with dementia 18 care homes 3 communities	Family 135 Residents 149	Family: involvement (F-INVOLVE); involvement importance (F-IMPORT) Resident: social skills (MAS-R) Baseline (admission), 6, 12 months	FI was not a sig predictor of changes in resident social skills over time, larger decreases in social skills associated with smaller social networks and sig fewer total visits	MMAT: **** Pos: Power, analysis reporting, longitudinal (12m), CI reporting, measures, response rate Neg: Type I error risk
Dobbs et al. (2005) (USA) RQ: 2	Correlational (cross-sectional) Compared dementia care in residential care (RC) / assisted living (AL) to care homes 35 RC/AL, 10 care homes 4 USA states	Family 400 Residents 400	Family: frequency of visits Resident: activity involvement (PAS-AD); Baseline	Families visited at least once in the last week, family assessing activities and social involvement was related to more resident activity involvement	MMAT: *** Pos: Large sample, adjustments, variance reporting Neg: No description of family participants, non-standardised facility measures, missing data
Grabowski & Mitchell (2009) (USA) RQ: 2	Correlational (longitudinal) Examined caregiver visit duration and resident quality end-of-life care 22 care homes 1 USA city	Family 323 Residents 323	Family: oversight (visit hours per week); satisfaction with care (SWC-EOLD) Resident: health and dementia severity (BANS-S); quality of life (QUALID); quality of care (seven domains) Baseline, quarterly for 18 months/death	Most families spent between one and 7 h visiting each week, family satisfaction with care highest in group that did not visit, quality of care sig worse for residents visited over 7 h per week	MMAT: *** Pos: Longitudinal, large sample, confound control, variance and limitation reporting Neg: One non-representative, geographical site
Jablonski et al. (2005) (USA) RQ: 1 & 2	CRCT Family involvement in care (FIC) intervention 14 care home special care units Midwest	Family 164 Residents 164	Resident: cognition (GDS); Function (FAC) Baseline, 3, 5, 7, 9 months	Resident deterioration reversed initially though not sig different by 9 months, no sig effect on resident self-care ability, inappropriate behaviour or agitation	MMAT: ** Pos: Attrition adjustment, site randomisation, cluster effects considered Neg: Blinding, no family description, power calculation or effect size, attrition
Livingston et al. (2017) (UK) RQ: 2	Correlational, cross-sectional Reported prevalence and determinants of agitation in residents with dementia 86 care homes	Family 1281 Resident 1483 Staff 1701	Family: visits Resident: agitation (CMAI); quality of life (DEMQOL); dementia severity (CDR); Neuropsychiatric symptoms (NPI) Baseline	Clinically significant agitation shown by 40% of residents with dementia. Agitation was not associated with number of visits by the main family carer	MMAT **** Pos: Large sample, sensitivity analyses, confound control, variance reporting, generalisability of results Neg: Possible underestimation of agitation level
Mariani et al. (2018) (Italy and Netherlands) RQ: 1 and 2	Quasi-experimental (MM) Analysed shared decision making (SDM) on agreement of residents’ ‘life-and-care plans’ 2 care homes (IT 2; NL 1)	Family 49 Residents 49 Staff 34	Family: quality of life (EuroQoL); sense of competence (SSCQ) Resident: care plans (case report Form); dementia stage (GDS); Katz Index of ADL Staff: sense of competence (SSCQ) Baseline and 6 months post intervention	Overall, care plans showed higher level of agreement with policy recommendations post SDM. Improvements in resident and family involvement in care planning found in Italy	MMAT: **** Pos: Use of control groups in each location, clearly operationalised care plan measures, inter-rater agreement applied, group difference considered Neg: Small sample size per outcome, site control group variance, no estimate of variance reported
Mbakile-Mahlanza et al. (2020) (Australia) RQ: 1 and 2	CRCT crossover Evaluated impact of Montessori activities implemented by family on visitation experiences of people living with dementia 9 care homes 1 Australian state	Family 20 Residents 20 Staff 9	Family: quality of visits and satisfaction (5-point Likert scale); personal mastery (truncated PMS); quality of relationship with relative (5-point Likert scale) and across 4 dimensions (MSFCI); carer mood (CESDS); quality of life (Carer-QOL), frequency and length of visits Resident: affect and engagement (PGCARS and MPES) Baseline, during and post intervention	Visits ranged from 2 to 32 per month, average 2 h per visit. Resident displays of pleasure and constructive engagement were significantly higher; anger, anxiety and passive engagement were significantly lower in Montessori versus control. Families experienced higher total visit satisfaction, and higher care-resident quality of relationship than in the control group	MMAT: **** Pos: Use of control groups and randomisation, group interaction and crossover effects analysed, effect size reported, low missing data rates Neg: Small sample size for multiple testing, no adjustment for group differences, control group components
Minematsu (2006) (Japan) RQ: 2	Correlational (longitudinal) Investigated family visits and behavioural and psychological symptoms of dementia (BPSD) 1 care home	Residents 67	Family: hours per week visiting/talking Resident: cognition (HDS-R); BPSD suppression (DBD) Baseline, 12 months	Majority of residents visited between none and ten times per month on average, frequency of visits associated with positive change in HDS-R and DBD in residents with initial moderate HDS-R, change was lower where visit frequency was above average	MMAT: * Pos: Longitudinal (12 m), measures, description of analysis, multiple appraisers Neg: Small single site sample, minimal description of participants and data collection, missing measure reference and limitations
Reinhardt et al. (2014) (USA) RQ: 1 and 2	RCT Palliative care conversation with follow-up calls intervention 1 care home Northeast	Family 90	Family: satisfaction with care (SWC-EOLD) Resident: symptom control (SM-EOLD); single item rating across seven end-of-life domains Baseline, 3, 6 months	Families had sig increased care satisfaction and had documented sig more end-of-life care decisions in care records, no sig difference in symptom management	MMAT: *** Pos: Randomisation, blinding, control group Neg: Sample size, no power calculation, description of randomisation
Robison et al. (2007) (USA) RQ: 1 and 2	CRCT (MM) Partner in caregiving intervention adapted for special care units (PIC-SCU) 20 care homes 1 USA state	Family 388 Staff 384	Family: conflict (ICS); staff provision (SPRS); staff behaviour (SBS); staff empathy (SES); hassle (NHHS); family involvement (FIS) Resident: agitation (CMAI) Staff: Conflict (ICS); family behaviour (FBS); family empathy (FES) Baseline, 2 and 6 months	Improvements in ease of talking with staff, and resident behaviours. Spouse/same-generation visits increased, number of programmes offered to families increased	MMAT: *** Pos: Sample size, 6m follow-up, confounding accounted for, response rates Neg: No variance reported, measure reliability
Toles et al. (2018) (USA) RQ: 2	Correlational (cross-sectional) Compared family perceptions of quality of communication with staff and clinicians, and links with resident and family characteristics 22 care homes 1 USA state	Family 302 Residents 302	Family: quality of communication including involvement and interactions, demographics Resident: Demographics Baseline	Family decision makers rated quality of communication with NH staff higher than that with clinicians and reported poor quality end-of-life communication for both staff and clinicians. 26% of staff and 50% of clinicians did not involve family decision makers in decisions about treatment residents would want.	MMAT **** Pos sample size, diverse homes, data collection, measure, adjusted for clustering effects, non-significant result included in reporting Neg: 1 geographical state, uncontrolled potential confounds identified, no effect size reporting
Van der Steen et al. (2012) Canada, Netherlands, Italy) RQ: 1	Quantitative retrospective study Evaluated families’ perspectives on acceptability, usefulness, preferred timing and way of obtaining a booklet on comfort care in dementia 38 care homes, (NL 28; IT 4; Canada 6)	Family 138	Family: Author developed 8-item scales Residents: Demographics & health problems assessed Baseline	Most families perceived the booklet as useful. Approximately half of the families endorsed availability not through practitioners. Italian families’ ratings differed from other countries across several domains including way of obtaining, profession preferred and timing	MMAT: **** Pos: Factor adjustments made, confounding and clustering factors considered, missing data management detailed Neg: Non-standard scales, discrepancy in care home representation across locations. Retrospective design may have introduced bias (acknowledged)
Verreault et al. (2018) (Canada) RQ: 1 & 2	Quasi-experimental study Assessed intervention to increase quality of care and quality of dying in people with advanced dementia including education provision, early and systematic communication with families 2 long-term care facilities	Family 124 Residents 193	Family: quality of care (FPCS); Resident: symptom management (SM-EOLD); quality of dying (CAD-EOLD); Pain (PACSLAC) 48 h before death and 4 weeks post or within 6 months of relative death	Sig increase in family satisfaction with care. Frequency of discussion with families and provision of information booklet higher than in care as usual. Sig increase in families’ perception of comfort assessment and symptom management	MMAT: **** Pos: Control group, validated measures, factor adjustments. confounding and clustering factors considered, full data management detailed Neg: Response rate disparity between study groups, no estimate of variance reported
Zimmerman et al. (2005) (USA) RQ: 2	Correlational (longitudinal) Compared dementia care in residential care (RC) / assisted living (AL) to care homes 35 RC/AL, 10 care homes 4 USA states	Family 170 Residents 170	Family: frequency of visits Resident: activity involvement (PAS-AD); quality of life (QOL in AD-activity); behaviour (DCM) Baseline, 6 months	Families spent almost 7 h per week on average visiting or talking with the resident, FI was associated to higher resident quality of life	MMAT: **** Pos: Longitudinal, randomisation within site, confound adjustments, limitation reporting Neg: Missing data, no power analysis or effect size

Note: ADL = Activities of daily living; AFC = Attitudes towards family checklist; BANS-S = Bedford Alzheimer’s Nursing Severity subscale; CAD-EOLD = Comfort Assessment in Dying; Carer-QOL = Carer’s quality of life; CDR = Clinical Dementia Rating; CESDS = Center for Epidemiological Studies Depression Scale; CMAI = Cohen-Mansfield Agitation Inventory; CRCT = Clustered randomised controlled trial; CSI = Caregiver stress inventory; DBD = Dementia behaviour disturbance scale; DCM = Dementia Care Mapping; DCS = Decisional Conflict Scale; DEMQOL = Dementia Quality of Life Measure; EuroQOL = EQ-5D Standardised Health Outcome Measure; FAC = Functional Abilities Checklist; FBS = Family Behaviors Scale; FES = Family Empathy Scale; FIS = Family Involvement Scale; FKOD = Family Knowledge of dementia test; FPCR = Family perceptions of caregiving role; FPCS = Family perceptions of care scale; FPCT = Family perceptions of care tool; GDS = Global Deterioration Scale; HDS-R = Hasegawa Dementia Scale-Revised; ICS = Interpersonal Conflict Scale; MAS-R = Multi-Focus Assessment Scale Revised; MFIS = Mental Function Impairment Scale; MPES = Menorah Park Engagement Scale; MSFCI = Mutuality Scale of the Family Caregiving Inventory; NHHS = Nursing Home Hassles Scale; NPI = Neuropsychiatric inventory; PACSLAC = Pain Assessment Checklist for Seniors with Limited Ability to Communicate; PAS-AD = Patient Activity Scale–Alzheimer’s Disease; PGCARS = Philadelphia Geriatric Center Affect Rating Scale; PMS = Pearlin mastery Scale; QUALID = Quality of Life in Late-Stage Dementia; RCT = Randomised controlled trial; RQ = Research Question; SBS = Staff Behaviors Scale; SES = Staff Empathy Scale; sig = significant; SKOD = Staff knowledge of dementia test; SM-EOLD = Symptom Management at the End-of-Life in Dementia Scale; SPCR = Staff perceptions of caregiving role; SPRS = Staff Provision to Residents Scale; SWC-EOLD = Satisfaction with Care at the End-of-Life in Dementia Scale; SSCQ = Short Sense of Competence Questionnaire.