In epiforecasts/omicron-sgtf-forecast: Real-time estimation of the time-varying transmission advantage of Omicron in England using S-Gene Target Status as a Proxy

knitr::opts_chunk$set(echo = FALSE, include = TRUE,
                      warning = FALSE, message = FALSE,
                      root.dir = here::here(),
                      fig.width = 12, fig.height = 9)
options(digits = 1)
library(here)
library(forecast.vocs)
library(dplyr)
library(loo)
library(scoringutils)
library(knitr)
library(data.table)

# load packages
library(ggplot2)
library(scales)
library(dplyr)
library(here)
library(readr)
library(data.table)

# load functions
source(here("R", "load-local-data.R"))
source(here("R", "plot-summary.R"))
source(here("R", "plot-daily-cases.R"))
source(here("R", "plot-omicron-95.R"))
source(here("R", "plot-cumulative-percent.R"))

1. Centre for the Mathematical Modelling of Infectious Diseases, London School of Hygiene & Tropical Medicine, London WC1E 7HT, United Kingdom

Introduction

Methods

Data

We use S-gene status by specimen date sourced from UKHSA as a direct proxy for the Omicron variant with a target failure indicating a case has the Omicron variant. Data was available by UKHSA region.

# get latest date
date_latest <- get_latest_date()

# load data
daily <- load_local_data(date_latest)

# start date
start_date <- as.Date("2021-11-23")

daily <- daily %>%
  filter(date >= start_date) %>%
  filter(!is.na(sgtf)) %>%
  filter(!(region %in% "England"))

# munge fraction, logit scaling, difference, and dow
daily <- daily %>%
  mutate(frac_sgtf = sgtf / total_sgt) %>%
  mutate(logit_frac_sgtf = log(frac_sgtf / (1 - frac_sgtf))) %>%
  group_by(region) %>%
  arrange(date) %>%
  mutate(diff_logit_sgtf = logit_frac_sgtf - lag(logit_frac_sgtf)) %>%
  ungroup() %>%
  mutate(weekday = weekdays(date) %>%
   factor(levels = c("Monday", "Tuesday", "Wednesday", "Thursday",
                     "Friday", "Saturday", "Sunday"))) %>%
  mutate(mondays = ifelse(weekday %in% "Monday", date, NA_Date_))

Models

Statistical Inference

Implementation

All models were implemented using the forecast.vocs R package [@R; @forecast.vocs] and fit using stan [@stan] and cmdstanr [@cmdstanr]. Each model was fit using 2 chains with each chain having 1000 warmup steps and 2000 sampling steps. Convergence was assessed using the Rhat diagnostic [@stan]. Models were compared using approximate leave-one-out (LOO) cross-validation [@loo; @loo-paper] where negative values indicate an improved fit for the correlated model.

Limitations

Results

Summary

Data description

Daily

daily %>%
  plot_daily_cases(
    truncate_date = max(daily$date), caption = "",
    start_date = start_date, smooth_total = TRUE
  ) +
    geom_vline(aes(xintercept = mondays), linetype = 2, alpha = 0.5)

Fraction of those tested for the S gene with target failure by UKHSA region

daily %>%
  ggplot() +
  aes(x = date, y = logit_frac_sgtf, col = region) +
  geom_vline(aes(xintercept = mondays), linetype = 2, alpha = 0.5) +
  geom_point(alpha = 0.8) +
  geom_line(alpha = 0.4) +
  scale_color_brewer(palette = "Paired") +
  theme_bw() +
  theme(legend.position = "bottom") +
  labs(x = "Specimen date",
       y = "Fraction of those tested for the S gene with target failure
        (logit scale)",
       col = "UKHSA region")

Growth in those with SGTF by day on the week for each UKHSA region

daily %>%
  ggplot() +
  aes(x = weekday, y = diff_logit_sgtf, col = region) +
  geom_violin(aes(col = NULL), alpha = 0.4, fill = "grey",
              draw_quantiles = c(0.25, 0.5, 0.75)) +
  geom_jitter(alpha = 0.6) +
  scale_color_brewer(palette = "Paired") +
  theme_bw() +
  theme(legend.position = "bottom") +
  guides(fill = NULL) +
  labs(x = "Day of week",
       y = "Difference in SGTF fraction from the previous day (logit scale)",
       col = "UKHSA region")

References

epiforecasts/omicron-sgtf-forecast documentation built on Jan. 21, 2022, 9:19 p.m.