Zimmer 1985.

Study characteristics
Methods	Design: RCT (patient) Methodological quality: 3/6a (EPOC RCT/CCT checklist)
Participants	Country and regions: US, Rochester, New York, metropolitan area Recruitment (1979‐1982; 27 months): from "variety of sources" following publicity campaign informing health professionals and the community; homebound patients living in study county with chronic illness or terminal illness (the latter were largely cancer patients receiving palliative care only and with life expectancy ≤ 3 months) who had a caregiver not necessarily living in same household and who wished to remain at home, with no physician to make home visits. Exclusion criteria: primary psychiatric illness, unable to fill in questionnaires Number of patients (randomised): 167 (85 intervention and 82 control); (baseline): 158 (82 intervention and 76 control) ; 22% intervention and 18% control classified as terminal (subanalysis performed)  Diseases (overall baseline sample): cancer (21% intervention, 17% control), stroke (12% intervention, 17% control), arthritis/rheumatism (9% intervention, 12% control), others, all below 10% (59% intervention, 54% control); those classified as terminal were "largely cancer" patients Patient characteristics: mean age 76 years, median age 77 years; 68% female; 23.2% intervention patients and 18.4% controls lived alone Deaths at end of study: 48 (29 intervention and 19 control, including 3 patients who died before the initial interview) Withdrawals/other drop‐outs: 21: 5 drop‐outs/changed their minds and 1 patient moved away before initial interview; 15 drop‐outs after initial interview (17 deliberate dropped out: 4 intervention and 13 control)
Interventions	Home palliative care vs. usual care "Home Health Care Team" Type: intermediate palliative care Service base: ambulatory care unit at University of Rochester Medical Center Team: physician‐led multiprofessional team with geriatric nurse practitioner (Masters' medical nurse practitioner) and social worker; weekly team conferences to assure co‐ordination of patient care Responsibility: 1 team member designated as primary provider in care plan following initial interdisciplinary assessment Description: existing service (since 1977) aimed to provide care for home‐bound, chronically or terminally ill, or disabled patients (largely elderly) for whom transportation to clinic had become too costly or cumbersome; initial home visit by each team member for initial assessment; interdisciplinary care plan then established which designated 1 team member as primary provider with consultative visits by others as needed; nurse conducted physical assessments in community (routine and emergency), was responsible for the nursing plan and supported medical care in close cooperation with physician; physician conducted home visits and evaluation at intake and follow‐ups as needed, and acted as patient's attending physician during any hospitalisations; social worker helped to deal with social, financial and emotional problems, freeing physician and nurse to devote mostly to medical and nursing care; encouraged informal care by family and friends by providing them with needed physical and psychological support and education to take on or continue HC; 24‐hour telephone service backed up by physician (all patients advised to restrict calls and service requests to office hours if possible, but were given nurse and physician's home numbers for emergencies and the physician's pager as ultimate resource); low number of out‐of‐hours calls (> 70% handled on telephone) Duration: not stated Control: usual care, including HC services available in community (area described as with well‐developed long‐term care services in general)
Outcomes	Death at home  Survival Assessment points: not stated
Resource use/costs	Out‐of‐home care (hospital admission, hospital inpatient days, ED visits, clinic visits, MD office visits, ambulance or chairmobile rides, nursing home days) In‐home care (MD home visits, nurse home visits, social worker home visits, RN/LPN hours, aide/homemaker visits, laboratory technician home visits, meals‐on‐wheels visits) Time horizon: terminal 2 weeks before death (subanalysis relevant to this review) Costs: out‐of‐home, in‐home, total costs (mean costs per patient of terminal 2 weeks) Currency: USD (date not stated)
Notes	Power considerations: none stated: "A sample size of 200 to 250 was desired over the period of study based on ability of the team to care for half that number (estimated from previous team experience of turnover, visit frequency, and case load)" The analyses relevant for our review were those conducted with the subsample of terminal patients/deaths, and the outcomes were: death at home for patients who died during 6 months' follow‐up (45; 28 intervention and 17 control) survival, with analysis adjusted for terminal/non‐terminal status healthcare use subanalysis of patients who died during 6 months follow‐up and who had at least 2 weeks in study before death with care utilisation data (33; 21 intervention and 12 control). Other outcomes included physical and psychosocial health, morale, and patient and caregiver care satisfaction but no subanalysis was conducted
Risk of bias
Bias	Authors' judgement	Support for judgement
Concealment of allocation (RCT / CCT)	Unclear risk	Sample stratified into terminal and non‐terminal; "randomized into experimental (team) and control groups"; process not stated
Follow‐up (RCT / CCT)	Unclear risk	45 (27%) for place of death; 80/85 intervention (94%) and 66/82 control for survival (81%)
Blinding (RCT / CCT)	Unclear risk	Double‐blind at baseline only: "The patients were informed of the result of randomization immediately following the initial interview, in order to keep the base‐line information as free as possible of bias from knowledge of the result either by the patient or interviewer"
Baseline measurement (RCT / CCT)	Low risk	"The only demographic variable which showed a statistically significant difference was sex, with 61 per cent females in the team group compared with 76 per cent in the controls; this is taken into account in the analysis by adjustment for sex. Previous health care utilization and diagnoses in the two groups were quite similar"
Reliability of outcome measurement (RCT / CCT)	Unclear risk	Interviewers (10/14 had previous survey experience) trained by study co‐ordinator with practice of interviews, role play, biweekly feedback meetings
Protection against contamination (RCT / CCT)	High risk	Patients/caregivers randomised

^aIn all Risk of bias tables, "high risk" is equivalent to the EPOC criteria “not done” (scored 0), unclear risk is equivalent to “not clear or varied across outcomes” (scored 0.5) and "low risk" is equivalent to “done” (scored 1)

AIDS: acquired immune deficiency syndrome; ARIMA: autoregressive integrated moving average; CBA: controlled before and after study; CC: conventional care; CCT: controlled clinical trial; CES‐D: Center for Epidemiological Studies Depression; CHF: congestive heart failure; COPD: chronic obstructive pulmonary disease; ECOG: Eastern Cooperative Oncology Group; ED: emergency department; EORTC QLQ‐C30: European Organisation for Research and Treatment of Cancer Quality of Life questionnaire; EN: enrolled nurse; EPOC: Effective Practice and Organisation of Care; ESAS: Edmonton Symptom Assessment Scale; FEV₁: forced expiratory volume in 1 second; GHQ‐28: 28‐item General Health Questionnaire; GP: general practitioner; HB: hospital based; HC: home care; HIV: human immunodeficiency virus; HMO: health management organisations; ICD: International Classification of Diseases; ITS: interrupted time series; KPS: Karnofsky Performance Status; LPN: licensed practical nurse; MD: medical doctor; MND: motor neurone disease; MS: multiple sclerosis; NYHA: New York Health Association; POS: Palliative Care Outcome Scale; QLI: quality of life index; QWB: quality of well‐being; RCT: randomised controlled trial; RGN: registered general nurse; RN: registered nurse; SD: standard deviation; SF‐36: Short Form 36; VA: Veterans Administration; WHO: World Health Organization