Table 2. Input values for multivariate sensitivity analysis, by scenario.
| Parameter | Type of probability distribution1 | Minimum Value2 | Likeliest value | Maximum Value2 | Sources |
| Uncertain variables | |||||
| Mortality effect | |||||
| Base case | Triangular | 4.1 | 17.9 | 31.8 | LUNESP trial |
| Optimistic scenario | Triangular | 4.1 | 17.9 | 31.8 | LUNESP trial |
| Conservative scenario | Triangular | 3.1 | 13.4 | 23.9 | Estimate3 |
| Births attended per BA/month | |||||
| Base case | Triangular | 0.70 | 1.29 | 1.87 | LUNESP trial4 |
| Optimistic scenario | Triangular | 1.82 | 3.34 | 4.86 | Estimate5 |
| Conservative scenario | Triangular | 0.49 | 1.21 | 1.93 | LUNESP trial6 |
| Cost of food/day/participant during training workshops7 | |||||
| Base case | Triangular | 10.6 | 14.1 | 17.7 | LUNESP trial |
| Optimistic scenario | Triangular | 10.6 | 14.1 | 17.7 | LUNESP trial |
| Conservative scenario | Triangular | 10.6 | 14.1 | 17.7 | LUNESP trial |
| Cost of fuel, car, staff for program monitoring/month7 | |||||
| Base case | Triangular | 466 | 622 | 777 | LUNESP trial |
| Optimistic scenario | Triangular | 466 | 622 | 777 | LUNESP trial |
| Conservative scenario | Triangular | 466 | 622 | 777 | LUNESP trial |
| Policy-determined variables | |||||
| Number of trainings/year | |||||
| Base case | 3 | LUNESP trial | |||
| Optimistic scenario | 2 | Estimate8 | |||
| Conservative scenario | 4 | Estimate8 | |||
| Number of days/training | |||||
| Base case | 2 | LUNESP trial | |||
| Optimistic scenario | 1 | Estimate8 | |||
| Conservative scenario | 2 | LUNESP trial | |||
| Months/year of monitoring | |||||
| Base case | 12 | LUNESP trial | |||
| Optimistic scenario | 6 | Estimate8 | |||
| Conservative scenario | 12 | LUNESP trial | |||
| Number of TBAs | |||||
| Base case | 60 | LUNESP trial | |||
| Optimistic scenario | 80 | Estimate8 | |||
| Conservative scenario | 60 | LUNESP trial |
Triangular distribution was chosen to be consistent with the approach used in other cost-effectiveness studies [24]–[26].
For parameters except births/TBA, these are minimum and maximum values. For births/TBA, these are 5% and 95% values determined by the 95% CI estimated from the standard deviation of monthly mean births/TBA from the LUNESP trial.
25% less than the LUNESP trial's mortality effect. Using a +/−25% range is typical in similar studies [17].
Based on the average number of births attended/TBA/month during the final year of the LUNESP trial.
Estimated value for typical area in Zambia, calculated by multiplying the LUNESP trial's mean value by the factor: (Zambia national population density/average population density in Lufwanyama).
Based on the average number of births attended/TBA/month during the 27 months of LUNESP trial implementation.
Costs in estimated 2011 US$.
Authors' estimates based on experience during the LUNESP trial.