ABCDEFGHIJKLMNOPQRSTUVWXYZ
1
Note: We produced the model before we decided that layworkers would be a more promising approach. Hence the model uses community health workers. We don't expect the cost-effectiveness to be drastically different because of this.
2
Color code
3
Input
4
Calculation
5
Subjective judgement
6
7
Guide to tabs
8
We modeled for three countries: Sierra Leone, Nigeria and Guinea. We modeled for the benefits to include <5 child mortality benefits (which are quite uncertain) and without them for each country. The training timeline tab is to map out how many facilitators we plan to train, and the remaining tabs are standard useful data used for each CEA.
9
10
Navigating the CEA
11
The main estimate we are considering for the input at hand ⬇️ We use this header to tell our Monte Carlo software what type of distribution to use. ⬇️The 5% interval ⬇️The 95% interval ⬇️⬇️Other parameters used in some distributions ⬇️The unit of measurement for the estimate ⬇️The mean of the distribution, which in theory should be a number close to the "expected" number ⬇️The median of the distribution ⬇️A description of our thinking in relation to the input.⬇️We use these columns to discuss reviews among researchers, sometimes some comments are left in our CEAs after publishing ⬇️
12
ExpectedUseCarlop05p95Param1Param2Param3NO EDITUnitsDistr. meanDistr. medianDescriptionSourceCommentsReplies
13
Description of intervention
DO NOT
EDIT COLUMN		FORMULA DO NOT EDIT THIS COLUMN
14
We usually include a short description of our envisioned intervention as modelled.
15
Modelled for [COUNTRY/ ETC.]
16
17
18
Tractability & counterfactuals
⬅️To aid understanding, we split our CEAs in general "areas". The groupings are purely for communication purposes and do not affect the actual model output.
19
Probability of success etc
⬅️To aid understanding, we split the general areas into smaller sub-groups. The groupings are purely for communication purposes and do not affect the actual model output.
20
Probability per year that change would happen anyway
2%Beta1%3%%#N/A#N/A
21
⬆️The input name - we try to make these as descriptive as possible. If we think the name is sufficiently descriptive, we may skip adding a description in the "Description column" ⬆️ The point estimate. These point estimates are ignored in our Monte Carlo simulations, but are what feed the inter-dependent cells in our spreadsheet CEAs and the final outputs. ⬆️The distribution name ⬆️The disribution parameters. These are used for our Monte Carlo simulations. #N/A#N/A
22
23
24
The Results section provides a series of different results
25
26
Results
27
These cells ↙️ contain the name (and unit) of the result we are referring to
28
DALYs averted per person reached per year
#REF!⬅️ These cells contain an outcome calculated from the point estimates in the CEA. It usually helps to work backwards looking at the formula in the cell, to understand how a result was reached. #N/A#N/A
29
30
Charity costs only
⬅️ Sometimes we want to understand how results vary depending on different assumptions (e.g., who pays for what). These grey subsections should clarify what exact assumptions are being made for the below results.
31
Number of people reached total, NPV
32
↗️ NPV is the Net Present Value of our model, it essentially tells us what the value of the intervention is at this point in time when considering how things are going to change in the future.
33
Charity costs and government costs
34
Cost effectiveness ($/DALY)
35
Cost effectiveness (DALY/$1000)
Output #N/A#N/A
36
⬆️ Most often, the results we care about are the top-line cost-effectiveness of the intervention. We usually depict this in two ways, as a dollar cost per DALY/income doubling/unit of value, and as a DALY/income doubling/unit of value "gained" per USD 1,000. These are two different ways of saying the same thing. ⬆️ We use "Output" to tell our Monte Carlo simulation software that this is the outcome we want to see depicted. #N/A#N/A
37
38
⬇️ Other tabs are usually very context dependent, and are used differently by different researchers. Hopefully, they should be relatively intuitive, and have some additional explanation where required. Often, a researcher may add a second country or different way of modelling things to double-check results. In some cases, these are done with significant time constraints and may drop a bit in terms of description or accuracy. ⬇️
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100