Analytical approach:
The cost-effectiveness analysis was designed for a randomised controlled trial. The time horizon was the duration of the trial, which was 12 weeks. The authors stated that the study perspective was societal.
Effectiveness data:
The effectiveness data were from the multi-site, randomised, controlled trial; individual patient data were used. There were 314 mothers in the usual-care group, and 296 mothers in the peer-support group. The main clinical effectiveness estimate was whether or not a mother had postpartum depression. This was assessed using the Edinburgh Postnatal Depression Scale (EPDS).
Monetary benefit and utility valuations:
The time off work needed by family or friends and partners was valued using the average hourly wage for a similar demographic group in Ontario, Canada. This wage was from Statistics Canada.
Measure of benefit:
The time off work that was not required, and the postpartum depression that was avoided, were captured in the analysis. The cost-effectiveness ratio was the cost per case of postpartum depression avoided.
Cost data:
The resource use was recorded by research nurses, and collected by a structured telephone interview at 12 weeks postpartum. The case reports included a modified version of the Health Service Utilization and Cost of Care Questionnaire. This captured both health service and family resource use, including out-patient health provider visits, nursing visits, new-mother support groups, mental health services, emergency room visits, in-hospital care, work absence, hired child care, and household help. The cost of training of volunteers was included; maternal work absence was not due to maternity leave. Health service use was valued using data from the Ontario Health Insurance Plan, the McMaster University System-Linked Resource Unit, and the Ontario Association of Social Workers. All costs were reported in 2011 Canadian dollars (CAD). The medical care component of the Canadian Consumer Price Index was used to adjust costs to 2011, where necessary.
Analysis of uncertainty:
Cost-effectiveness acceptability curves were produced by bootstrapping the individual patient data. Several one-way sensitivity analyses were conducted, by varying the resource use estimates.