Table 1 Characteristics of included studies

Structures of care

 Dorr et al. [20], USA

Registered Nurse (RN) direct care time per resident per day:

30–40 min

Less than 10 min

1376

Retrospective cohort study

Cost-benefit analysis; societal; institutional

1 year

2001;

Secondary sources including national databases, with true costs obtained where possible;

USD

No

NH

Annual net societal benefit of $3191 per resident per year in nursing home units with 30–40 min of RN direct care time per resident per day compared to less than 10 min

 Grabowski and O’Malley [ 34], USA

Off-hours physician coverage via telemedicine vs. on-call physician

N/A

Cluster randomised controlled trial

Cost-benefit analysis; insurance provider (medicare)

2 years

Oct 2009–Sep 2011;

Estimated cost of hospitalizations to Medicare from recent literature;

USD

No

NH

15.1 hospitalisations avoided. Net savings of $120,000 per facility per year

 Jenkens et al. [36], USA

Green House model

Usual care

N/A

Cross-sectional

Cost analysis; institutional

N/A

2009;

Wages derived from salary.com and payscale.com with 5% increase applied to Green House CNA wage;

USD

No

SNF

GH facilities use 1.97–2.49% more staff than traditional nursing homes

 Maas et al. [23]; Swanson et al. [38]; Swanson et al. [44], USA

Special care unit

Traditional unit

44

Prospective cohort study

Cost analysis; health care

1 year

Date not disclosed;

Resource use measured and unit costs assigned—source of unit cost data not disclosed;

USD

Yes

NH

Costs of care for residents with dementia in special care units were 29% higher than cost of care on traditional units

 Mehr and Fries [24], USA

Special care unit

Traditional unit

6663

Cross-sectional

Cost analysis; institutional

N/A

Date not disclosed;

Resource use data from the resident status measure database, a preliminary version of the national nursing home minimum Data set;

USD

Yes

NH

Unadjusted resource use was 18% lower on SCUs than other units in the facility; when adjusted for case mix no significant difference in resource use was found

 Przybylski et al. [28], CAN

Physical Therapy (PT) & Occupational Therapy (OT) staffing levels:

1.0 FTE PT and 1.0 FTE OT per 50 beds

1.0 FTE PT and 1.0 FTE OT per 200 beds

115

Randomised controlled trial

Cost analysis; institutional

2 years

1993/1994;

Direct care nursing costs calculated based on the Alberta resident classification system (case mix measure) which estimates average amount of nursing care required per category. Source of wage data not disclosed;

Currency not disclosed.

No

NH

PT/OT delivered at a 1:50 ratio was more effective at promoting, maintaining, or limiting decline in functional status. The resulting reduction in required care delivery resources was estimated to provide an annual cost saving of $283 per bed (a 1% cost reduction)

 Schneider et al. [35], GBR

1.0 FTE occupational therapist

Usual care

190

Non-randomised experimental trial

Cost analysis; health and social services

1 year

2002–2003;

Published unit costs, inflated to 2005;

GBP

No

CH

Intervention group showed a significant increase in the likelihood of using social services. At 2005 levels, net cost of providing occupational therapy was £16 per resident per week

 Sharkey et al. [37], USA

Green House model

Usual care

240

Cross-sectional

Cost analysis; institutional

N/A

2008–2009;

Observational, interview, and survey methods at participating facilities;

N/A

No

SNF

Total staffing time (excluding administration) in Green House facilities was 18 min less per resident per day that traditional facilities. CNAs in Green House facilities spent 24 min per resident per day more time in direct care activities than CNAs in traditional facilities

 Teresi et al. [29], USA

Implementation of an evidence-based education and best practice program:

Training staff vs. training staff and nursing home inspectors vs. usual training

N/A

Quasi-experimental

Cost-benefit analysis; Societal

2.5 years

2008;

Aggregate cost data based on local estimates and published literature;

USD

No

NH

Training staff was associated with a 15% reduction in annual falls, while training staff and inspectors was associated with a 10% reduction in falls.

Range of estimates for the cost-benefit analysis is between a net loss of $26,000 and a net savings of $52,000

Processes of care

 Chenoweth et al. [39], AUS

Person-centred care (PCC)

Person-centred environment (PCE)

Both PCC + PCE

Usual care

601

Cluster randomised controlled trial

Cost analysis; institutional

8 months

2009–2011;

Resource use measured and unit costs assigned using market rates;

AUD

Yes

RACF

PCC: 7169 per home; PCE: 9198 per home; PCC + PCE: 22,857 per home. Reduced agitation and improvements in resident quality of life for care homes which instituted PCC and PCE. The PCC + PCE intervention produced significant improvements in quality of care interactions and care responses, but no improvements in agitation or quality of life

 Chenoweth et al. [31]; Norman et al. [32], AUS

Person-centred care (PCC)

Dementia-care mapping (DCM)

Usual care

289

Cluster randomised controlled trial

Cost-effectiveness analysis; institutional

8 months

2008;

Pharmaceutical costs: Australian pharmaceutical benefit schedule

Training costs: Bradford University, UK

Staff costs: Commonwealth Government Aged Care Nurses’ Award;

AUD

Yes

RACF

Dementia care mapping was found to be a more expensive and less effective intervention than person-centred care. The cost per negative behaviour averted in the person-centred care group was $8.01 post-intervention and $6.43 at follow-up relative to usual care

 MacNeil Vroomen et al. [22], NED

Multidisciplinary Integrated Care (MIC)

Usual care

301

Cluster randomised controlled trial

Cost-effectiveness analysis; societal

6 months

2007;

Health care utilisation collected via patient/proxy interview and medical records. Source of cost data not disclosed. CPI figures sourced from the Dutch bureau of statistics;

EUR

No

RH

For functional health and QALYs, multidisciplinary integrated care was not found to be cost-effective compared to usual care. For patient-related quality of care, the probability that the intervention was cost-effective compared to usual care was 0.95 or more for ceiling ratios greater than €129

 Molloy et al. [25], CAN

Advance Directive program

Usual care

1292

Cluster randomised controlled trial

Cost analysis; health care

1.5 years

Date not disclosed;

Unit prices sourced from local and provincial fee schedules;

CAD

No

NH

Intervention nursing homes reported 44% fewer hospitalisations per resident (0.27 versus 0.48), and 33% less resource use ($3490 versus $5239) than the control facilities.

 Müller et al. [33], DEU

Multifactorial fracture prevention program

Usual care

N/A

N/A

Markov-based simulation model

Cost-utility analysis; insurance provider

20 years

2012;

Retrospective dataset of costs for NH residents from an insurance fund (n = 60,091), a public German dataset for fracture treatment costs, and catalogue of non-physician care for physical therapy costs;

EUR

No

NH

Base-case analysis of multifactorial fall prevention resulted in a cost-effectiveness ratio of €21,353 per QALY

 Ouslander et al. [43], USA

INTERACT II tools (Interventions to Reduce Acute Care Transfers)

N/A

Controlled before-and-after

Cost analysis; institutional

6 months

2010;

Wages based on national data;

USD

No

NH

Intervention group reported 17% reduction in hospitalisation rates. The average cost of the 6-month intervention was $7700 per facility

 Paulus et al. [21], NED

Integrated care

Traditional care

342

Quasi-experimental

Cost analysis; societal

1.2 years

Date not disclosed;

Activity based costing, data obtained from participating nursing homes and a published guide for cost research;

EUR

No

NH

Integrated care had 31% lower informal direct care costs per resident. Total average costs per resident were on average 4% higher in integrated care than traditional care

 Rantz et al. [26], USA

Multilevel intervention with expert nurses vs. monthly info packs on ageing and physical assessment

N/A

Cluster randomised controlled trial

Cost analysis; institutional

2 years

Date not disclosed;

Medicaid cost reports;

USD

No

SNF

Total costs per resident per day increased 6% in the intervention group, and decreased 3% in the control. The intervention demonstrated improvements in quality of care, pressure ulcers and weight loss

 Rovner et al. [27], USA

A.G.E. dementia care program (activities, medication guidelines, educational rounds) vs. usual care

81

Randomised controlled trial

Cost analysis; institutional

6 months

Date not disclosed;

Monthly billing records;

USD

Yes

ICF

At 6 months, intervention residents were more than 10 times more likely to participate in activities than controls. Additional cost of the intervention was $8.94 per resident per day

 van de Ven et al. [30], NED

Dementia-care mapping (DCM)

Usual care

318

Cluster randomised controlled trial

Cost-minimisation analysis; health care

1.5 years

2010–2012;

Data collected over a period of 18 months. Sources included the Dutch manual of health care cost, and cost prices delivered by a pharmacy and a nursing home;

USD (EUR 1.00 = USD 1.318)

Yes

NH

No significant effect on total costs for the intervention