Skip to main content

Table 3 Hospice at home

From: A scoping review of initiatives to reduce inappropriate or non-beneficial hospital admissions and bed days in people nearing the end of their life: much innovation, but limited supporting evidence

Author, date and journal

Study design

Sample and setting

Research focus

Relevant results

Wider implications

Limitations

Wong et al.; 2013; Annals, Academy of Medicine, Singapore [38]

Quantitative, prospective cohort study.

44 Singaporean patients with advanced heart failure enrolled in an Advanced Care Programme between July 2008 and July 2010.

To evaluate the impact of a home-based Advanced Care Programme on healthcare utilisation in the last stages of life for advanced heart-failure patients.

Mean all-cause and heart failure related admissions prior to Advanced Care Programme enrolment were 3.6 and 2.0 episodes. Following admission to the programme, these dropped to 1.0 and 0.6 (p < 0.0001), representing 72 and 70% reductions, respectively. After follow-up duration, adjusted hospital admission episodes were 1.2 and 0.5 (p < 0.0001).

Home-based palliative care has shown reduced hospital utilisation for patients with heart failure.

Small sample size.

No control group.

Did not account for placebo effects of a palliative programme.

Lustbader et al.; 2017; Journal of Palliative Medicine [39]

Quantitative, retrospective, comparative analysis.

82 US Medicare Shared Savings Program accountable care organisation deceased patients compared with 596 patients receiving usual care between October 2014 and March 2016.

To evaluate the impact of a home-based palliative care programme within an accountable care organisation on cost and resource utilisation during the final 3 months of life.

A 34% reduction in hospital admissions during the final month of life was observed in hospital-based palliative care-enrolled patients.

Hospital-based palliative care was also linked with a reduction in emergency department visits per 1000 patients compared with standard care (878 vs. 1097).

Hospital-based palliative care in an accountable care organisation is linked with significant cost savings and fewer hospital admissions.

Lack of minority populations.

No case-matched control group (although disease burden was quantified and equal for both groups, p = 0.4).

Spilsbury et al.; 2017; PLOS ONE [40]

Quantitative, retrospective cohort study.

12,763 Western Australian patients who died from cancer, and 7 from non-cancer conditions.

To investigate how community-based palliative care is associated with a reduction in acute care health service use.

During the intervention period, hospital admissions dropped by 34% and the mean length of stay decreased by 6%. For patients < 70 years, receiving community-based palliative care was linked with a reduced rate of hospital admissions approximately 5 months before death, and in those > 70 years, the reduction was observed up to a year before death.

Hospital admission rates were reduced when patients were receiving community-based palliative care. These effects could be seen within 5 months of death and earlier in older patients.

Lack of detail on intervention intensity.

Palliative care may have been different for those living in rural and remote areas.

Seow et al.; 2014; British Medical Journal [41]

Quantitative, pooled analysis of a retrospective cohort study.

6218 Canadian cancer patients, of whom 3109 received community-based palliative care administered by one of 11 specialist palliative care teams (intervention)a, and 3109 received standard community-based palliative care during the last 2 weeks of life (control).

To understand the pooled effect of exposure to any one of 11 palliative care teams operating at patients’ homes.

970 patients from the intervention group visited a hospital, compared with 1219 from the control group (31.2% vs. 39.3%, p < 0.001). 896 patients from the intervention group were admitted to the emergency department, compared with 1070 in the control group (28.9% vs. 34.5% p < 0.001).

The pooled relative risks of being admitted to hospital and an emergency department in late life for the intervention versus control group are 0.68 and 0.77, respectively.

Fewer intervention patients died in hospital than control patients (503 vs. 887, p < 0.001).

Palliative care teams appear to reduce both hospital and emergency department admission, as well as the amount of hospital deaths.

Propensity scores cannot account for unmeasured variables such as preferred place of death or existing care.

Generally, cancer patients only.

Seow et al.; 2016; Journal of Pain and Symptom Management [42]

Quantitative, retrospective cohort study.

54,576 Canadian decedents who used home care nursing services during the last 6 months of life.

To explore the link between the home care nursing rate and emergency department visit rate in the following week during the last 6 months of life.

Patients who received the intervention during any week had a 31% lower emergency department visit rate during the following week than did patients receiving standard care. During the last month of life, those receiving home care end-of-life nursing and standard care for more than 5 h per week were associated with a decreased emergency department visit rate of 41 and 32%, respectively, compared with patients with 1 h of standard nursing per week.

There is a temporal association between receiving home-based end-of-life nursing care in a specific week during the last 6 months of life and a reduced emergency department visit rate. Higher levels of both home-based end-of-life care and standard care are also linked with a lower emergency department visit rate.

Study does not imply causality.

The database does not hold other important factors which may influence emergency department visit rates.

Gagnon et al.; 2015; Journal of Pain and Symptom Management [43]

Population study.

52,316 Canadian cancer patients between 2003 and 2006. 27,255 had received home palliative care services

To test the association between the quality of home palliative care services and quality of end-of-life palliative care indicators.

Increased quality of home palliative care services was associated with a lower proportion of men having 1 or more visit to an emergency department during the last month of life (risk ratio 0.924; 95% CI 0.867–0.985).

In women, higher-quality home palliative care services were also associated with a higher proportion of patients dying at home (risk ratio 2.255; 95% CI 1.703–2.984) and spending less time in hospital (risk ratio 0.765; 95% CI 0.692–0.845).

Effective home palliative care services were associated with improved end-of-life palliative care indicators.

Lack of a needs assessment.

Iupati and Ensor; 2015; Internal Medicine Journal [44]

Retrospective, quantitative study of consecutive patient records over 6 years.

73 New Zealand-based patients with advanced chronic obstructive pulmonary disease who were referred to 2 community hospice programmes.

To examine the impact of community-based hospice programmes on hospital admission in patients with advanced chronic obstructive pulmonary disease.

Following entry into the hospice programme, a mean decrease of 2.375 hospital admissions over 1 year was observed.

Community-based hospice programmes may be linked with a reduction in hospital admissions in patients with advanced chronic obstructive pulmonary disease.

Highly selected and small sample size.

Only 2 regional hospices.

Rosenwax et al.; 2015; Palliative Medicine [45]

Quantitative, retrospective, population-based cohort study.

All patients in Western Australia who died with dementia between January 2009 and December 2010 (n = 5261) and a comparator of patients who died from other conditions who received palliative care (n = 2685)

To report on the use of hospital emergency departments during the last year of life by patients who died with dementia and the effect that community-based palliative care had upon that use.

The rate of emergency department visits significantly increased if community-based palliative care was not being received, varying over the last year of life. During the initial 130 days, patients who received standard care visited the emergency department 1.4 more times often than those receiving community-based palliative care. In the last month of life, those receiving standard care in a private residence visited the emergency department 6.7 times more frequently and those receiving standard care in a care facility visited emergency department 3.1 times more frequently than those who received community-based palliative care.

Community-based palliative care of people who die with dementia is linked with a significant reduction in emergency department visits during the last year of life.

A retrospective study using data not collected for the purpose of the study.

Alonso-Babarro et al.; 2013; Palliative Medicine [46]

Quantitative population-based comparison.

549 cancer patients in 2 areas in the Madrid region. Only 1 area had a palliative home care team.

To explore the impact of palliative home care team in the last 2 months of life on place of death, emergency room visits, admissions and use of hospital resources.

Frequency of patients dying in hospital was significantly lower in the palliative home care team area (61% vs. 77%, p < 0.001), as were the number of patients using emergency services (68% vs. 79%, p = 0.004) and number of patients using in-patient services (66 vs. 76%, p = 0.012). After adjusting for other factors, patients in the palliative home care team area had a lower risk of hospital death than those in the non-palliative home care team area (adjusted odds ratio 0.4, 0.2–0.6).

Suggests that palliative home care team is associated with reduced in-patient deaths and use of hospital services in the last 2 months of life.

Focused on 2 specific urban areas, meaning unidentified factors may influence in-patient deaths.

McNamara et al.; 2013; Journal of Palliative Medicine [47]

Quantitative, retrospective cross-sectional study.

746 Australian deceased cancer patients between August 2005 and June 2006.

To investigate whether early entry to community-based palliative care reduces emergency department admissions during the last 90 days of life.

During the final 90 days of life, fewer patients who had early access to palliative care visited an emergency department than did patients without early access (31.3% vs. 52.0%).

Early access to community-based palliative care can reduce the frequency of patients’ emergency department visits.

The data were not originally collected to answer the research question.

Ranganathan et al.; 2013; Journal of Palliative Medicine [48]

Quantitative, retrospective cohort study.

391 American palliative home care patients compared with 890 palliative care patients receiving standard care at home, post-acute care programme.

To study the impact of palliative home care on 30-day hospital readmission rates.

Those receiving palliative care at home had a 30-day readmission probability of 9.1% vs. 17.2% in those who received standard care at home.

Compared to regular care at home, palliative home care may reduce 30-day hospital readmission.

The study could not adjust for treatment intensity, which may have affected readmission.

Cassel et al.; 2016; Journal of the American Geriatric Society [49]

Quantitative, retrospective study using propensity-based matching.

368 American patients who received a home-based palliative intervention, Transitions, between 2007 and 2014 were propensity matched with 1075 control individuals. The 4 disease types were cancer, chronic obstructive pulmonary disease, heart failure and dementia.

To assess non-clinical outcomes of a home-based palliative care programme.

For all 4 diseases, the number of participants admitted to hospital and the number of hospital days were lower for the Transitions group than the control (p < 0.001). The transitions group also had a lower rate of admission during the final 30 days of life (p < 0.001) and hospital deaths (p < 0.001).

The mean 30-day readmission rate was lower in the Transitions group for COPD (p = 0.005), heart failure (p < 0.001) and dementia (p = 0.01), but not for cancer (p = 0.08).

The home-based Transitions programme could reduce hospital utilisation for the 4 included disease types.

Retrospective studies have the potential for selection bias.

Tan el al; 2016; Palliative medicine [50]

Retrospective cohort study.

321 Singaporean cancer-diagnosed patients who had an expected prognosis of ≤1 year and were enrolled in an integrated hospice home care programme were compared with 593 patients who were referred to other home hospices.

To evaluate the impact of an integrated hospice home care programme on acute care service usage and on rate of home deaths.

Hospital deaths were significantly lower in programme participants (12.1% vs. 42.7%).

The programme group had significantly lower emergency department visits and hospitalisations at 30 days, 60 days and 90 days prior to death.

Integrating acute care and home hospice care could reduce acute care service usage.

Only cancer patients were included.

Riolfi et al.; 2014; Palliative Medicine [51]

Retrospective cohort study.

402 patients who died in 2011 of cancer as a primary cause in Veneto region, Italy.

To assess the effectiveness of appropriate palliative home care services in reducing hospital admissions and identify predictive factors in hospitalisations of patients treated at home.

The 39.9% of patients who were enrolled in the palliative care programme were more likely to die at home (53.8% vs. 7.9%), had fewer hospital admissions (0.4 vs. 1.3 admissions) and shorter stays (4.4 vs. 19.6 days) in their last 2 months of life.

The findings indicate that palliative home care teams enable consenting patients to stay at home with reduced hospital usage compared with those not under the care of palliative home care teams.

Variables between the intervention and control groups, such as symptoms and socio-economic status, were not considered.

Lukas et al.; 2013; Journal of Palliative Medicine [52]

Longitudinal case series.

369 patients with advanced complex illness referred for home-based palliative consultation in the mid-Atlantic United States.

To evaluate home-based palliative care on hospital outcomes within a fee-for-service environment.

There was a significant reduction in hospital utilisation (number of hospitalisations, total hospital bed days, total costs, variable costs and probability of 30-day readmission).

The only exception was no change in the probability of an emergency department visit.

These results suggest a home-based palliative medicine practice may reduce hospital utilisation for patients with advanced, complex illness.

Non-randomised.

Only accessed hospital utilisation data from 1 health care network.

Goldenheim et al.; 2014; Journal of Palliative Medicine [53]

Retrospective cohort study.

59 patients (19 readmitted within 30 days and 40 controls not readmitted) aged ≥65 years who were newly discharged to home hospice care between February 2005 and January 2010.

To identify factors associated with 30-day readmission among older adults newly discharged to hospice.

25% of admitted patients received a palliative care consultation, compared with 47% of patients who were not readmitted.

Patients discharged to hospice and readmitted within 30 days were less likely to have palliative care consultations.

Small sample and selection bias.

Kerr et al.; 2014; Journal of Palliative Medicine [54]

Prospective, observational database study.

149 decedents who received the home care programme. The control population (n = 537) was derived using propensity score matching.

To evaluate the impact on healthcare costs and utilisation of an innovative home-based palliative care programme implemented through a hospice-private payer collaboration.

Home care patients had much lower hospital admission rates, particularly in the last 6 months of life.

There was no significant difference in emergency department use between home care patients and those in the control group.

Palliative care programs partnered with community hospice providers may achieve cost savings while providing care across the continuum.

Selection bias.

Statistically significant differences in diagnoses and disease types between the control and intervention groups.

Pouliot et al.; 2017; American Journal of Hospice and Palliative Care [55]

Prospective cohort study.

123 participants with serious illness enrolled in a programme in New York.

To evaluate the effectiveness of Care Choices, a new in-home palliative care service.

The group had fewer emergency department visits (p < 0.001) and inpatient stays (p < 0.001)) after enrolment in the intervention.

An in-home palliative care programme may be a successful model that satisfies patients’ desire to remain at home and avoid hospital admissions.

Observational results and no randomisation.

  1. aThe specialist palliative care teams included a core group of physicians, nurses and family physicians who provided integrated palliative care in patients’ homes