Table 5

Results of primary and sensitivity economic analyses comparing CSNAT intervention with usual care

Net costs (95% CI)Net QALYs (95% CI)ICER (£/QALY)
Primary analysis
 Multiple imputed datasets (n=410)*£39.05 (−69.61 to 147.71)−0.004 (−0.020 to 0.012)Intervention is dominated
Sensitivity analyses
 Complete cases (n=131)£41.24 (−29.01 to 111.49)−0.0001 (−0.026 to 0.026)Intervention is dominated
 Per-protocol analysis† (n=374)£42.55 (−71.77 to 156.88)−0.0002 (−0.016 to 0.016)Intervention is dominated
 Exclude training and intervention costs (n=410)*£23.33 (−98.21 to 144.87)−0.004 (−0.020 to 0.012)Intervention is dominated
Alternative outcome measure
Net costs (95% CI)Net change (95% CI)ICER: (£/1 point improvement)
 FACQ strain, complete cases (n=139)£57.32 (−15.77 to 130.41)−0.02 (ie, lower score in intervention group) (−0.30 to 0.26)Intervention is dominated
  • All analyses adjusted for covariates: carer’s age, time since stroke, stroke severity, whether or not carer has long-term health conditions, length of experience of cluster staff, size of cluster and cluster ID.

  • CIs for all analyses calculated following bootstrapping: 2000 times for imputed datasets, 10 000 times for complete case datasets.

  • *Four participants with no baseline EQ-5D data were excluded from the imputation, leaving 410 participants.

  • †Thirty-six participants in the imputed dataset excluded who violated protocol conditions (multiple carers per stroke survivor or questionnaires returned late).

  • CSNAT, Carer Support Needs Assessment Tool; FACQ, Family Appraisal of Caregiving Questionnaire; ICER, incremental cost-effectiveness ratio; QALYs, quality-adjusted life years.