Nine studies (18,304 patient, range 82 to 14,827) were included in the review, comprising two randomised controlled trials, two non-randomised studies, and five retrospective record reviews. Two studies were rated as high quality, three as moderate quality and four as low quality.
For incidence of home death, there was a significant effect that favoured palliative home nursing (OR 4.45, 95% CI 3.24 to 6.11; Ι²=93%; nine studies). There was little evidence of publication bias. The effect was not statistically significant in high-quality studies (OR 1.40, 95% CI 0.97 to 2.02; two studies).
None of the three studies that reported symptoms/quality of life found a significant effect favouring the comparator; a meta-analysis of the two high quality studies showed no significant difference between groups, but there was substantial heterogeneity (Ι²=96%).