McMillan 2007 found that caregivers overestimated symptom intensity for pain (the difference in mean scores was 1.33 in a 0 to 10 scale), breathlessness (0.78 difference in mean scores in a 0 to 10 scale) and constipation (1.08 difference in mean scores in 0 to 16 scale); all three differences were highly significant (paired t‐tests; P value < 0.001).
Although there was a marginally significant difference in the mean percentage of time spent at home in the last month of life (48% and 41%, respectively; P value = 0.06), this was lost (P value = 0.15) when adjusted for other predictive factors (gender and living with spouse).
Pooled data from five trials (four RCTs, three of high quality, and one CCT with 899 patients) showed that although home palliative care patients had lower odds of dying in a nursing home than those under usual care, the difference did not reach statistical significance (OR 0.64, 95% CI 0.40 to 1.03; Z = 1.83, P value = 0.07; Chi 2 = 2.77, df = 4, P value = 0.60; I 2 = 0%; Analysis 1.5 ).
The higher frequency of deaths at home in the intervention group in Zimmer 1985 failed to reach statistical significance, although a clinically significant difference cannot be ruled out (OR 2.86, 95% CI 0.78 to 10.53 in a sample of 43 patients; Analysis 1.1 ).