Brumley 2007.
| Study characteristics | ||
| Methods |
Design: RCT (patient) Methodological quality: 3.5/6a (high quality, EPOC RCT/CCT checklist) |
|
| Participants |
Country and regions: US, Hawaii and Colorado, metropolitan areas Recruitment (2002‐2004, 18 months): from discharge planners, primary care and specialist physicians in 2 non‐profit HMOs (Kaiser Permanente Group), life expectancy ≤ 12 months with ED or hospital admissions in last year and physical function ≤ 70 in Palliative Performance Scale (from 0 death to 100 normal) Number of patients (randomised): 310 (155 intervention and 155 control) Diseases: cancer (138), CHF (97), COPD (62) Patient characteristics: mean age 73.8 years; 49% female; 26.3% lived alone Deaths at end of study: 225 (73%) Withdrawals/other drop‐outs: 5 withdrawals (2 in the intervention and 3 in the control); 8 patients died after randomisation but prior intervention (all in the intervention group) |
|
| Interventions | Home palliative care vs. usual care "In‐Home Palliative Care – IHPC" Type: intermediate palliative care Service base: 2 non‐profit Kaiser Permanente Group HMOs – 1) Hawaii: 18 medical offices of 317 medical group physicians providing all outpatient care and most inpatient care (with internal home health agency, contracts with external providers for hospice care only); 2) Colorado: 16 ambulatory medical offices of more than 500 physicians representing all medical specialities and subspecialities (contracts with external providers for ED, hospital, home health and hospice care) Team: physician, nurse, social worker with support from others (spiritual counsellor/chaplain, bereavement co‐ordinator, home health aide, pharmacist, dietician, volunteer, physiotherapist, occupational therapist, speech therapist) Responsibility: team's physician (but allowed primary provider to continue) Description: modelled on hospice programmes but without forgoing curative treatment; patients assigned to palliative care physician to co‐ordinate care from a variety of healthcare providers; included medical care, symptom management, education, biopsychosocial support, co‐ordination, continuous assessment and management of care across all settings; initial inter‐disciplinary assessment (physical, medical, psychological, social and spiritual needs); joint care plan to decide frequency of visits on individual basis; advance directives with patient and family; training patients and caregivers on use of medication, self management and crisis intervention at home; patient wrist band for ED staff to know patient was in home palliative care and had advance directives on file (1 site only); physicians home visits and nursing care 24 hours on‐call Duration: from enrolment to death or transfer to hospice care (mean survival 196 days in the intervention group) Control: usual care with various amounts and levels of primary care services, home health services (when Medicare criteria met), acute care services and hospice care |
|
| Outcomes | Death at home Patient satisfaction with care Survival Assessment points: satisfaction with care measured at baseline, 30, 60, 90 and 120 days after enrolment Analysis: forwards from enrolment |
|
| Resource use/costs | Hospitalisations (admission and inpatient days) ED visits Hospice care (referral and days) Time horizon: from enrolment to death, transfer to hospice care or study end Costs: mean adjusted total costs and mean adjusted total costs per day (included those associated with physician visits, ED visits, hospital days, skilled nursing facility days, and home health or palliative days) Currency: 2002 USD |
|
| Notes |
Power considerations: study powered for cost differences (planned 300; randomised 310; analysed 297) Data on satisfaction with care at 120 days were not analysed due to "significant reduction in sample size" (n = 136) |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Concealment of allocation (RCT / CCT) | Low risk | "blocked randomization using a computer generated random number chart, stratified according to study site" |
| Follow‐up (RCT / CCT) | Unclear risk | 70% follow‐up for patient satisfaction with care at 30 days (n = 216); 73% follow up for place of death (n = 225); 145/155 intervention (94%) and 152/155 control for survival (98%) |
| Blinding (RCT / CCT) | Unclear risk | Single‐blind: research assistants blinded but patients not stated |
| Baseline measurement (RCT / CCT) | Low risk | Baseline differences in care satisfaction scores (intervention higher satisfaction; P value = 0.03) but analysis conducted with binary variable (no baseline differences) |
| Reliability of outcome measurement (RCT / CCT) | Unclear risk | Collected from patients or their caregivers by research assistants (care satisfaction), HMO's records, death certificates and family report (death at home, survival) |
| Protection against contamination (RCT / CCT) | High risk | Patients randomised |