Table 1. Distributions for model parameters.
| Parameter | Distribution 1 | Description | Reference(s) |
|---|---|---|---|
| Duration of incubation period (days) | Survivors: Gamma(2.8640, 3.5058) Non-survivors: Gamma(2.5988, 3.3515) | Incubation period was assumed to be distributed according to a gamma distribution, the parameters of which were fitted to a subset of the Case Classification Data (CCD) for cases who identified the funeral of a case as the source of transmission. We restricted the distribution to between 1 and 21 days and also stratified for survivors and non-survivors. | Fitted to CCD data |
| Overall duration of symptoms (days) | Triangular(5,8,14) | A triangular distribution with mode of eight days was derived from published estimates. | [14–16] |
| Duration of late symptoms | Uniform(1,5) | The duration of the late symptoms phase duration was drawn from a uniform distribution ranging from one to five days, as has been clinically characterized by more severe symptoms including vomiting, diarrhea, hemorrhaging, and organ failure. | [14, 17] |
| Daily viral load 2 per case | Gamma (Varies per case) | Survivorship-specific daily viral load was sampled per person from a gamma distribution, fitted to published viral load data from the 2000–2001 Uganda outbreak. | [18] |
| Rate ratio of transmission risk | Gamma(4.5824, 0.5874) (S2 Fig) | Relative risk distributions for contact and survivorship-specific viral load were sampled using a Monte Carlo scheme to generate a distribution for the rate ratio. A gamma distribution was fit to the empirical distribution for viral load and truncated between 1 and 100. | [11, 18, 19] |
| Number of contacts per epidemic month | Raw data (S3 Fig) | A distribution was derived using available CCD and Contact Tracing Data (CTD) for each month. The range was truncated between 1 and 40 contacts. | Calculated from data |
| Days from onset to hospitalization per epidemic month | Raw data (S4 Fig) | A distribution was derived using available CCD for each epidemic month. The range was truncated at 20 days and sampled from only for cases reporting care-seeking but no date. | Calculated from data |
| Days from onset to death per epidemic month | Raw data (S5 Fig) | A distribution was derived using available CCD for each epidemic month. The range was truncated at 100 days. | Calculated from data |
| Days from death to burial per epidemic month 3 | [3, 3, 1, 1, 0, 0] | A distribution was derived using available CCD for average number of days between death and date of funeral practices. | Calculated from data |
| Probability of community burial after hospitalization per epidemic month 3 | [0.2727, 0.2697, 0.3306, 0.3883, 0.4293, 0.4150] | A distribution was derived using available CCD for non-survivors with reported care-seeking and for whom the burial location (i.e. hospital or community) was documented. Distributions were calculated for each epidemic month. | Calculated from data |
| Probability of sanitary funeral per epidemic month 3 | [0, 0, 0, 0, 0.8, 0.8] | The probability of a sanitary burial given a community-based funeral. | [20] |
| Zone contact matrix per epidemic month | Raw data (Heat maps for contact matrices presented in S1 Fig) | Frequencies of inter- and intra-zone interactions were determined using the zones of residence of cases and their contacts reported in the CTD or CCD | Calculated from data |
1 Distributions are presented as name of distribution followed by relevant parameters: Gamma(shape, scale), Triangular(lower bound, mode, upper bound), and Uniform(lower bound, upper bound).
2 Viral load was measured based on the mean and standard deviation counts of daily RNA copy levels following symptoms onset and are stratified by case fatality.
3 Data presented for each epidemic month and reported as [June, July, August, September, October, November].