TABLE 2.

Parameters Used to Calculate District Costs of MNH Care, Life-Years Lost Due to Maternal Death, and Incremental Cost-Effectiveness of Deaths Averted

Number	Parameter	Value	Data Source	Notes
Costs (all)
1	Discount rate	3%	WHO-CHOICE recommendation³⁴	Locally published discount rates used in sensitivity analysis (15% in Uganda and 9.7% in Zambia)³⁵^,³⁶
Costs 2012
2	Costs associated with the first delay	Varies by district (see Table 4)	Interviews with health facility staff, district health office staff, provincial health office staff, and implementing partners in comparison districts	Interviews covered the period 2012 through 2016; start-up activities and capital costs were tracked. Costs for existing maternity waiting homes are included.
3	Costs associated with the second delay	Varies by district (see Table 4)	Interviews with health facility staff, district health office staff, provincial health office staff, implementing partners, and review of ambulance log books in comparison districts	Interviews covered the period 2012 through 2016; start-up activities and capital costs were tracked.
4	Unit cost of ANC	Varies by type of facility (see Table 3)	Data collection at health facilities in comparison districts, interviews with implementing partners	Inclusive of facility overhead costs
5	Number of ANC visits	Ratio of ANC visits to number of facility births	Data from health facility registers/district health offices in comparison districts	Number of facility births based on SMGL districts data from 2012
6	Unit cost of vaginal delivery	Varies by type of facility (see Table 3)	Data collection at health facilities in comparison districts, interviews with implementing partners	Inclusive of facility overhead costs and admissions (for mother and newborn)
7	Number of vaginal deliveries	Varies by district	Data from health facility registers/district health offices in comparison districts, Serbanescu and colleagues³⁰	Number for SMGL districts in 2012
8	Unit cost of cesarean delivery	Varies by type of facility (see Table 3)	Data collection at health facilities in comparison districts, interviews with implementing partners	Inclusive of facility overhead costs and admissions (for mother and newborn)
9	Number of cesarean deliveries	Varies by district	Data from health facility registers/district health offices in comparison districts, Serbanescu and colleagues³⁰	Number for SMGL districts in 2012
10	Above community/facility costs	Varies by district (see Table 4)	Interviews with health facility staff, district health office staff, provincial health office staff, and implementing partners in comparison districts	Interviews covered the period 2012 through 2016; start-up activities and capital costs were tracked.
11	Total costs of MNH care in 2012	Calculation	Based on parameters 2–10
Costs 2016
12	Costs associated with the first delay	Varies by district (see Table 4)	Interviews with health facility staff, district health office staff, provincial health office staff, and implementing partners in SMGL districts	Interviews covered the period 2012 through 2016; start-up activities and capital costs tracked. Costs for maternity waiting homes are included.
13	Costs associated with the second delay	Varies by district (see Table 4)	Interviews with health facility staff, district health office staff, provincial health office staff, implementing partners, and review of ambulance log books in SMGL districts	Interviews covered the period 2012 through 2016; start-up activities and capital costs were tracked.
14	Unit cost of ANC	Varies by type of facility (see Table 3)	Data collection at health facilities in SMGL districts, interviews with implementing partners	Inclusive of facility overhead costs
15	Number of ANC visits	Ratio of ANC visits to number of facility births	Data from health facility registers/district health offices in SMGL districts	Number of facility births based on SMGL districts data from 2016
16	Unit cost of vaginal delivery	Varies by type of facility (see Table 3)	Data collection at health facilities in SMGL districts, interviews with implementing partners.	Inclusive of facility overhead costs and admissions (for mother and newborn)
17	Number of vaginal deliveries	Varies by district	Serbanescu and colleagues³⁰	Number for SMGL districts in 2016
18	Unit cost of cesarean delivery	Varies by type of facility (see Table 3)	Data collection at health facilities in SMGL districts, interviews with implementing partners	Inclusive of facility overhead costs and admissions (for mother and newborn)
19	Number of cesarean deliveries	Varies by district	Data from health facility registers/district health offices in SMGL districts, Serbanescu and colleagues³⁰	Number for SMGL districts in 2016
20	Above community/ facility costs	Varies by district (see Table 4)	Interviews with health facility staff, district health office staff, provincial health office staff, and implementing partners in comparison districts	Interviews covered the period 2012 through 2016; start-up activities and capital costs were tracked.
21	Total costs of MNH care in 2016	Calculation	Based on parameters 12–20	In Uganda, included cost of patients referred to Fort Portal referral hospital
Deaths in 2012
22	Number of facility-based deliveries	Varies by district	POMS and unpublished district data,³¹ district offices in SMGL districts	Number of deliveries for SMGL districts in 2016 multiplied by the institutional delivery rate in 2012
23	Maternal death ratio	534 deaths (Uganda) and 370 deaths (Zambia) per 100,000 live births	Serbanescu and colleagues³⁰
24	Perinatal death rate	39.3 (Uganda) and 37.9 deaths (Zambia) per 1,000 births	Serbanescu and colleagues³⁰
25	Number of maternal deaths	Calculation	Parameter 22 × proportion of deliveries with live births/100,000 × Parameter 23
26	Number of perinatal deaths	Calculation	Parameter 22/1,000 × Parameter 24
27	Total number of deaths	Calculation	Parameter 25 + Parameter 26
28	Life-years lost due to death	Years of life left estimated as 62.5 and 45.6 for perinatal and maternal death in Uganda and 62.3 and 45.7 for perinatal and maternal death in Zambia	WHO life tables⁴⁰^,⁴¹	Assume average age at death for maternal death is 27.5, for perinatal in first 2 days of life
Deaths in 2016
29	Number of facility-based deliveries	Varies by district	POMS and unpublished district data,³¹ district offices in SMGL districts	Number for SMGL districts in 2016; varied in sensitivity analysis based on results for all SMGL districts²⁴
30	Maternal death ratio	300 deaths (Uganda) and 231 deaths (Zambia) per 100,000 live births	Serbanescu and colleagues³⁰	Decreased the percentage reduction in deaths results by 10 percentage points in sensitivity analysis
31	Perinatal death rate	34.4 (Uganda) and 28.2 deaths (Zambia) per 1,000 births	Serbanescu and colleagues³⁰
32	Number of maternal deaths	Calculation	Parameter 29 × proportion of deliveries with live births/100,000 × Parameter 30
33	Number of perinatal deaths	Calculation	Parameter 29/1,000 × Parameter 31
34	Total number of deaths	Calculation	Parameter 32 + Parameter 33
35	Life-years lost due to death	Years of life left estimated as 62.5 and 45.6 for perinatal and maternal death in Uganda and 62.3 and 45.7 for perinatal and maternal death in Zambia	WHO life tables⁴⁰^,⁴¹	Assume average age at death for maternal death is 27.5, for perinatal in first 2 days of life. Years of life left estimated as 62.5 and 45.6 for perinatal and maternal death in Uganda and 62.3 and 45.7 for perinatal and maternal death in Zambia.
Incremental cost-effectiveness
36	Incremental costs	Calculation	Parameter 21 − Parameter 11	In sensitivity analysis, reassess with all donor costs treated as incremental costs.
37	Incremental deaths averted	Calculation	Parameter 34 − Parameter 27
38	Incremental life-years gained	Calculation	Parameter 35 − Parameter 28
39	Incremental cost per death averted	Calculation	Parameter 36/Parameter 37
40	Incremental cost per life-year gained	Calculation	Parameter 36/Parameter 38

Abbreviations: ANC, antenatal care; MNH, maternal and newborn health; POMS, Pregnancy Outcome Monitoring Survey; SMGL, Saving Mothers, Giving Life; WHO CHOICE, World Health Organization's Choosing Interventions that are Cost-Effective.