There were 116 studies in total, although in 18 studies only the direction of the treatment effect was provided. The main analysis is based on 98 studies (5,326 patients) representing 116 experimental treatment groups: education only (20 groups); education with counselling (20 groups); nonbehavioural/noncognitive counselling only (8 groups); behavioural/cognitive counselling only (65 groups); and behavioural/cognitive and nonbehavioural/noncognitive counselling (3 groups).
Psychological well-being: Anxiety (68 studies), 95% of studies showed a positive treatment effect, which was significantly larger than the 50% expected by chance. The average effect size (based on 55 studies) was 0.56 (95% confidence interval, CI: 0.42, 0.70), but the effect sizes were heterogeneous. A subgroup analysis by type of pyschoeducational care (tested in 5 or more studies) found effect-size values homogeneous in 4 of the 6 subgroups. Further analysis of the 2 heterogeneous subgroups suggests that the 6 types of care have similar effects on anxiety.
Depression (48 studies), 92% of studies showed a positive treatment effect that was statistically-significant. The average effect size (based on 40 studies) was 0.54 (95% CI: 0.43, 0.65) and the effect sizes were homogeneous. Mood (30 studies), 87% of studies showed a positive treatment effect that was statistically-significant. The average effect size (based on 25 studies) was 0.45 (95% CI: 0.32, 0.58), but the effect sizes were heterogeneous. The number of studies within individual types of treatment was too small to enable useful subgroup analysis.
Physical well-being: Nausea (27 studies), 93% of studies showed a positive treatment effect that was statistically-significant. The average effect size (based on 21 studies) was 0.69 (95% CI: 0.45, 0.92), but the effect sizes were heterogeneous. When analysis was restricted to studies in which patients had documented nausea prior to treatment, the effect on nausea was larger and homogeneous (1.04, 95% CI: 0.69, 1.39).
Vomiting (16 studies), 81% of studies showed a positive treatment effect that was statistically-significant. The average effect size (based on 12 studies) was 0.34 (95% CI: 0.09, 0.69) and the effect sizes were homogeneous.
Pain (13 studies), 92% of studies showed a positive treatment effect that was statistically-significant. The average effect size (based on 11 studies) was 0.43 (95% CI: 0.16, 0.69), but the effect sizes were heterogeneous. Relaxation-type interventions showed a large and homogeneous effect (0.91, 95% CI: 0.35, 1.47). Education only or multiple behavioural strategies produced small, non significant and heterogeneous average effect sizes.
Knowledge (19 studies), 95% of studies showed a positive treatment effect that was statistically-significant. The average effect size (based on 18 studies) was 0.90 (95% CI: 0.61, 1.20), but the effect sizes were heterogeneous. Treatment effect on knowledge was significantly larger in treatments that included written content.