Extract basic rates from malariasimulation model output

Summarises clinical incidence `clinical`, severe incidence `severe` and mortality `mortality`. Output also includes `severe` and `mortality` outputs broken down by in hospital (`_hospital`) or in the community (`_community`) rates as well as per-capita years lived with disability `yld`, years of life lost `yll` and daly `dalys`. All rates are expressed per person per day.

Usage

get_rates(
  x,
  baseline_year = 2000,
  ages_as_years = TRUE,
  treatment_scaler = 0.42,
  hosp_sev_cfr = 0.065,
  community_sev_cfr = 0.6,
  mild_disability_weight = 0.006,
  moderate_disability_weight = 0.051,
  severe_disability_weight = 0.133,
  clinical_episode_length = 0.01375,
  severe_episode_length = 0.04795,
  life_expectancy = life_expectancy_africa,
  infer_ft = TRUE,
  infer_ft_sev = TRUE
)

Arguments

x: Input data.frame
baseline_year: Baseline year (assumes simulation starts on the first day of the year)
ages_as_years: Convert ages to be in units of years
treatment_scaler: The impact of first-line treatment coverage on progression to severe disease and death. The probability of being hospitalised for untreated cases compared to treated cases is difficult to study and is not known, but can be estimated from available data with several assumptions. Our estimate is derived from data used in a meta-analysis on the impact of delayed treatment of uncomplicated malaria on progression to severe malaria (Mousa et al, 2020, Plos Medicine), which estimated the relative risk ratio for treatment delays of 1-2, 2-3 and >3 days compared to treatment within 1 day. Here, we assume that the risk of hospitalisation with severe malaria among untreated cases would be double the risk in this last “delay category” of >3 days (2 * RR 1.186), giving a relative risk of any severe malaria of 0.578 for treatment within 1 day compared to no treatment (i.e. a treatment scaler of 0.42 for the proportion of severe malaria cases that are prevented by treatment of uncomplicated malaria). However, in sensitivity analysis a range of 0.281-0.843 should be considered for the treatment scaler, which was derived by assuming the risk among untreated cases would be the same as the risk with a treatment delay of >3 days (lower estimate) or 3 times the risk as those with a treatment delay of >3 days (upper estimate).
hosp_sev_cfr: Severe case fatality ratio in hospital The original estimate fitted to Reyburn et al data in Griffin et al (2016) 0.065
community_sev_cfr: Severe case fatality ratio in the commnity (non-hospitalised) The original estimate fitted to Lubell et al data in Griffin et al (2016) 0.6
mild_disability_weight: disability weight for mild malaria. Assigned to clinical cases in those over 5 years old
moderate_disability_weight: disability weight for moderate malaria. Assigned to clinical cases in those under 5 years old
severe_disability_weight: disability weight for severe malaria. Assigned to all severe cases
clinical_episode_length: average length of an episode of clinical malaria
severe_episode_length: average length of an episode of severe malaria
life_expectancy: data.frame of expected years left to live. See example in data for format
infer_ft: If ft not found in model output (usually if ft = 0), assume ft = 0. If FALSE an error will be thrown
infer_ft_sev: If ft_sev not found in model output assume ft_sev = 0.8. If FALSE an error will be thrown

Details

A note on severe disease and deaths: We follow an approach that broadly aligns with the methods presented by Griffen et al (2016), See Griffin et al (2016) SI for more details. However, we allow more flexibility. The user may specify a column `ft_sev` to signify the assumed proportion of severe cases that receive hospital treatment. This is then used to estimate community severe cases from model output hospitalised severe incidence. Hospitalised and community severe incidence are combined with estimates of the severe case fatality ratio for hospitalised severe cases (from data originally from Reyburn et al) and community severe cases (from data originally from Lubell et al). In the absence of a user specified ft_sev, a default of 0.8 is used. This follows the original fit from Giffen et al (2016) and reproduces the original estimate of the ratio of total deaths to hospitalised cases of 0.215.

For DALYs: disability weights are from the Global Burden of Disease study. To estimate YLL we assume the average life expectancy for a person aged x years taken from the UN WPP sub-Saharan-Africa profile. To estimate the expected age in a given range we assume exponentially distributed ages with the range. Disability weights sourced: here. This is an approximation of YLD estimation from the GBD study; disability due to comorbid conditions such as motor impairment and aneamia are excluded.

Note: Default parameter values are for _Plasmodium falciparum_

Note: Some model outputs (for example _Plasmodium vivax_) do not include severe incidence (`n_inc_severe_...`) columns. When these are absent they are created to match the clinical incidence age groups and filled with `NA`, with a warning. The severe, mortality and DALY (`yld`, `yll`, `dalys`) columns in the returned output will then be `NA`.