26. Survival: home palliative care versus usual care.
| Study | Measure | Analysis | Significance and direction | Details |
|
Bakitas 2009 US (high quality) |
Months from enrolment to death or study end | Kaplan‐Meier and log rank test (post‐hoc exploratory analysis, patients alive at end of study censored on study last day) | n.s. P value = 0.14 |
Intervention (n = 161): median 14 months (95% CI 10.6 to 18.4 months) Control (n = 161): median 8.5 months (95% CI 7.0 to 11.1 months) |
|
Brumley 2007 US (high quality) |
Days in study | t‐test, Kaplan‐Meier and log rank test (patients alive at end of study censored on study's last day) | n.s. with log rank test 2.98; P = 0.08 (favours control with t‐test P = 0.03) | Intervention: "average" 196 days (± 164) Controls: "average" 242 days (± 200) |
|
Jordhøy 2000 Norway (high quality) |
'Survival' days | Kaplan‐Meier and log‐rank test adjusted for 3 cancer diagnostic groups (not clear how patients alive at end of study were analysed) | n.s. P value = 0.1 (adjusted for diagnosis) |
Intervention: median 99 days (95% CI 79 to 119 days) Control: median 127 days (95% CI 88 to 166 days) |
|
McKegney 1981 US (high quality) |
'Length of survival' | Not stated | No differences stated but significance not stated | "It should be briefly noted that the intensive and non‐intensive patients did not differ in terms of length of survival" (McKegney 1981) |
|
Grande 1999 UK |
Days from referral to death |
MannWhitney U tests | n.s. z 1.666; P value = 0.096 Favours control vs. those who received intervention service (113/186 of intervention) z 3.005; P value = 0.003 |
Intervention (n = 186): median 11 days (quartiles from 4 to 34 days) Control (n = 43): median 11 days (quartiles from 3 to 26 days) Subanalysis of intervention patients Patients who received service (n = 113): median 16 days (quartiles from 5 to 42.5 days) Patients who did not receive service (n = 73): median 8 days (quartiles from 3 to 18 days) |
|
Zimmer 1985 US |
'Survival profiles' | log‐likelihood ratio Chi2 using Aitken and Clayton method assuming exponential survival and adjusting for terminal/ non‐terminal status (unclear how patients alive at end of study were analysed) | n.s. log‐likelihood ratio Chi2 1.4; P value > 0.10 |
No descriptive data provided |
|
Hughes 1992 US |
'Survival days' | Not stated | n.s. (authors stated "n.s.") |
Intervention: M 76.2 days (SD 67.1) Control: M 83.1 days (SD 68.1) Decedents only Intervention: M 48.0 days (SD 43.3) Control: M 54.5 days (SD 47.7) |
|
Ahlner‐Elmqvist 2008 Sweden |
Days after enrolment | Kaplan‐Meier and log‐rank test (patients alive at end of study were excluded) | Favours control log rank 12.04; P value = 0.0005 |
Intervention: median 94 days Control: median 160 days |
|
Axelsson 1998 Sweden |
Months from diagnosis to death |
Wilcoxon signed rank (comparisons included 3 groups: intervention and control groups and a third historical group) | Favours intervention P value = 0.043 |
Intervention: median 23 months (M 35 ± 42) Control: median 6 months (M 16 ± 19) |
|
Greer 1986 (CBA) |
'Survival' | Not stated | n.s. |
"We tested the differences in patient survival in the study groups, controlling for disease, performance status, selected symptoms, as well as the amount of time patients were able to be followed, and found no statistically significant differences" (Greer 1986); no data provided to support this statement |
CI: confidence interval; M: mean; n.s.: non‐significant; SD: standard deviation.