diff --git a/DESCRIPTION b/DESCRIPTION index bec2f7e..ff9909c 100644 --- a/DESCRIPTION +++ b/DESCRIPTION @@ -27,7 +27,7 @@ Config/testthat/parallel: true Encoding: UTF-8 Language: en-GB Roxygen: list(markdown = TRUE) -RoxygenNote: 7.2.3 +RoxygenNote: 7.3.1 Imports: cinterpolate, dplyr, diff --git a/NEWS.md b/NEWS.md index d342220..03cb4dc 100644 --- a/NEWS.md +++ b/NEWS.md @@ -1,3 +1,8 @@ +# 0.1.1 + +* Project vignette now includes Zika in French Polynesia as a case study. + + # 0.1.0 * This project now includes a diff --git a/R/data.R b/R/data.R index 8cff567..b574dab 100644 --- a/R/data.R +++ b/R/data.R @@ -12,3 +12,21 @@ #' } #' #' @source "fiji_2014" + +#' Tahiti 2014. +#' +#' Dataset includes weekly laboratory confirmed Zika cases from Tahiti in French Polynesia, during an outbreak in 2013 and 2014 +#' as well as monthly average temperature data and rainfall for Tahiti during this period +#' +#'#' A data frame with 25 rows and 6 columns: +#' \describe{ +#' \item{date}{Date of epidemiological week} +#' \item{cases}{Number of laboratory confirmed cases in Tahiti} +#' \item{reporting_sites}{Number of sentinel sites in Tahiti reporting that month} +#' \item{total_sites}{Total sentinel sites Tahiti} +#' \item{av_temp}{Average monthly temperature for Tahiti, in °C} +#' \item{av_temp}{Average monthly rainfall for Tahiti, in mm} +#' ... +#' } +#' #' @source +"tahiti_2013" diff --git a/data/tahiti_2013.rda b/data/tahiti_2013.rda new file mode 100644 index 0000000..d9cbb4e Binary files /dev/null and b/data/tahiti_2013.rda differ diff --git a/inst/WORDLIST b/inst/WORDLIST index 699f66b..77fe5c1 100644 --- a/inst/WORDLIST +++ b/inst/WORDLIST @@ -6,7 +6,9 @@ AetaRVFV AetaSINV AetrEEEV AngaPfal +Aubry CMD +Childs Codecov CxanMVEx CxanRRVx @@ -40,5 +42,6 @@ io lifecycle packagename packagetemplate +polynesiensis svg yaml diff --git a/man/tahiti_2013.Rd b/man/tahiti_2013.Rd new file mode 100644 index 0000000..fd374d1 --- /dev/null +++ b/man/tahiti_2013.Rd @@ -0,0 +1,30 @@ +% Generated by roxygen2: do not edit by hand +% Please edit documentation in R/data.R +\docType{data} +\name{tahiti_2013} +\alias{tahiti_2013} +\title{Tahiti 2014.} +\format{ +An object of class \code{tbl_df} (inherits from \code{tbl}, \code{data.frame}) with 25 rows and 6 columns. +} +\usage{ +tahiti_2013 +} +\description{ +Dataset includes weekly laboratory confirmed Zika cases from Tahiti in French Polynesia, during an outbreak in 2013 and 2014 +as well as monthly average temperature data and rainfall for Tahiti during this period +} +\details{ +#' A data frame with 25 rows and 6 columns: +\describe{ +\item{date}{Date of epidemiological week} +\item{cases}{Number of laboratory confirmed cases in Tahiti} +\item{reporting_sites}{Number of sentinel sites in Tahiti reporting that month} +\item{total_sites}{Total sentinel sites Tahiti} +\item{av_temp}{Average monthly temperature for Tahiti, in °C} +\item{av_temp}{Average monthly rainfall for Tahiti, in mm} +... +} +#' @source \url{https://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0004726} +} +\keyword{datasets} diff --git a/vignettes/climateR0.Rmd b/vignettes/climateR0.Rmd index 2f7412b..e92b576 100644 --- a/vignettes/climateR0.Rmd +++ b/vignettes/climateR0.Rmd @@ -20,7 +20,7 @@ knitr::opts_chunk$set( Temperature is an important driver of vector-borne disease transmission, affecting vector reproduction, development, and survival, as well as the probability of pathogen transmission. Previous work by [Mordecai and colleagues](https://onlinelibrary.wiley.com/doi/10.1111/ele.13335) empirically estimated the effect of temperature on different vector traits, and used these to develop models of temperature dependent R~0~. -The {climateR0} package extracts temperature-dependent R0 for an input time series of mean temperature for 14 vector-pathogen combinations, focusing on mosquito-borne diseases that pose a major public health threat. Temperature dependent R~0~ is a relative measure bounded between 0 and 1, where 1 indicates maximum temperature suitability for transmission. This is a useful indicator for the epidemic potential of a vector-borne disease which can be used for situational awareness or be incorporated into forecasting models to predict future cases. Note that we use a relative measure of R~0~ as other factors affect the absolute magnitude of R~0~ such as immunity, control measures and population behaviour, which are not considered here. +The {climateR0} package extracts temperature-dependent R~0~ for an input time series of mean temperature for 14 vector-pathogen combinations, focusing on mosquito-borne diseases that pose a major public health threat. Temperature dependent R~0~ is a relative measure bounded between 0 and 1, where 1 indicates maximum temperature suitability for transmission. This is a useful indicator for the epidemic potential of a vector-borne disease which can be used for situational awareness or be incorporated into forecasting models to predict future cases. Note that we use a relative measure of R~0~ as other factors affect the absolute magnitude of R~0~ such as immunity, control measures and population behaviour, which are not considered here. ```{r setup, include = FALSE} library(climateR0) @@ -30,29 +30,29 @@ library(cowplot) ### Case study - 2013/14 DENV3 outbreak in Fiji -As a case study, we'll use data from a 2013/4 DENV-3 outbreak in Fiji. Here we show laboratory confirmed cases over time in Central Division. +As a first case study, we'll use data from a 2013/4 DENV-3 outbreak in Fiji. Here we show laboratory confirmed cases over time in Central Division. ```{r} fiji_cases <- fiji_2014 |> ggplot2::ggplot() + ggplot2::geom_line(ggplot2::aes(x = date, y = cases), col = "#016c59") + ggplot2::scale_x_date(breaks = "month", date_labels = "%b-%y") + - ggplot2::labs(x = "Date", y = "Cases") + + ggplot2::labs(x = "Date", y = "Reported dengue cases") + ggplot2::theme_classic() fiji_cases ``` -In the same dataset, we have a time-series of monthly mean temperatures (in °C) for Fiji. To extract corresponding temperature-dependent R~0~ values from the temperature-relative R0 curves estimated by Mordecai et al, we use the `temperature_R0()` function. Within the function call, we specify a vector-pathogen code of `AeaeDENV` for the vector *Aedes aegypti* and pathogen dengue virus. +In the same dataset, we have a time-series of monthly mean temperatures (in °C) for Central Division, Fiji. To extract corresponding temperature-dependent R~0~ values from the temperature-relative R0 curves estimated by Mordecai et al, we use the `temperature_R0()` function. Within the function call, we specify a vector-pathogen code of `AeaeDENV` for the vector *Aedes aegypti* and pathogen dengue virus. ```{r} fiji_2014$rR0 <- temperature_r0(fiji_2014$av_temp, "AeaeDENV") ``` -Now we can plot relative temperature-dependent R0 values alongside case data. +Now we can plot relative temperature-dependent R~0~ values alongside case data. -```{r, include = FALSE} +```{r} fiji_rR0 <- fiji_2014 |> ggplot2::ggplot() + ggplot2::geom_line(ggplot2::aes(x = date, y = rR0), col = "#54278f") + @@ -67,4 +67,57 @@ fiji_rR0 cowplot::plot_grid(fiji_cases, fiji_rR0, nrow = 2) ``` -As discussed in [Kucharski et al 2018](https://elifesciences.org/articles/34848#s2), comparing the case time series with temperature-dependent R~0~ indicates that a fall in transmission due to seasonal temperature variation cannot fully explain the fall in cases from March 2014. In this paper, Kucharski and colleagues found that a model including the build-up of herd immunity and a decrease in transmission resulting from a vector control campaign in March 2024 better captured the observed pattern of cases. \ No newline at end of file +As discussed in [Kucharski et al (2018)](https://elifesciences.org/articles/34848#s2), comparing the case time series with temperature-dependent R~0~ indicates that a fall in transmission due to seasonal temperature variation cannot fully explain the fall in cases from March 2014. In this paper, Kucharski and colleagues found that a model including the build-up of herd immunity and a decrease in transmission resulting from a vector control campaign in March 2024 better captured the observed pattern of cases. + +### Case study - 2013/14 Zika outbreak in French Polynesia + +As a second study, we'll use data from a 2013/4 Zika outbreak in Tahiti, French Polynesia. Here we show laboratory confirmed cases over time from sentinel reporting sites in Tahiti. + +```{r} +tahiti_cases <- tahiti_2013 |> + ggplot2::ggplot() + + ggplot2::geom_line(ggplot2::aes(x = date, y = cases), col = "#016c59") + + ggplot2::scale_x_date(breaks = "month", date_labels = "%b-%y") + + ggplot2::labs(x = "Date", y = "Reported Zika cases") + + ggplot2::theme_classic() + +tahiti_cases + +``` + +In the same dataset, we have a time-series of monthly mean temperatures (in °C) for Tahiti. Again, we use the `temperature_R0()` function. Within the function call, we specify a vector-pathogen code of `AeaeZIKV` for the vector *Aedes aegypti* and pathogen Zika virus. (Note that the main vectors in French Polynesia include both *Aedes aegypti* and *Aedes polynesiensis*.) + +```{r} +tahiti_2013$rR0 <- temperature_r0(tahiti_2013$av_temp, "AeaeZIKV") +``` + +Now we can plot relative temperature-dependent R~0~ values alongside case data. + +```{r} +tahiti_rR0 <- tahiti_2013 |> + ggplot2::ggplot() + + ggplot2::geom_line(ggplot2::aes(x = date, y = rR0), col = "#54278f") + + ggplot2::scale_x_date(breaks = "month", date_labels = "%b-%y") + + ggplot2::labs(x = "Date", y = "Relative R0") + + ggplot2::theme_classic() + +tahiti_rR0 +``` + +```{r} +cowplot::plot_grid(tahiti_cases, tahiti_rR0, nrow = 2) +``` + +As discussed in [Kucharski et al (201)](https://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0004726), the outbreak ended before temperature or rainfall levels changed in a way that would imply a transmission reduction, indicating that immunity played a major role in driving the epidemic decline. This is consistent with later antibody data reported by [Aubry et al ( 2017)](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5367400/) + +### Caveats to the model approach + +The models used here do not account for other aspects of weather and climate variability that are likely to influence vector-borne disease dynamics, including variation in precipitation, relative humidity, and temperature over time. This approach also does not account for time lags between temperature and R0; instead, it calculates a static snapshot of the relative temperature suitability for transmission at a given point in time, assuming the temperature had been constant up to that point. In reality, multiple time lags may be relevant to R0. For example, temperature 4-8 weeks ago may affect the mosquito population size and development rate, whereas temperature 2-4 weeks ago may affect the vector competence and extrinsic incubation period. + +Despite these limitations, temperature-dependent relative R0 has been used to predict geographical and temporal variation in different settings for malaria, dengue, Zika, Ross River virus and West Nile virus (further details in the references below). + +* [Mordecai et al 2017](https://journals.plos.org/plosntds/article?id=10.1371/journal.pntd.0005568) +* [Tesla et al 2018](https://royalsocietypublishing.org/doi/full/10.1098/rspb.2018.0795?rfr_dat=cr_pub++0pubmed&url_ver=Z39.88-2003&rfr_id=ori%3Arid%3Acrossref.org) +* [Shocket et al 2018](https://elifesciences.org/articles/37762) +* [Shocket et al 2020](https://elifesciences.org/articles/58511) +* [Childs et al 2024](https://www.medrxiv.org/content/10.1101/2024.01.08.24301015v1)