analysis/vaccination.r

In sbfnk/flusurvey: Packages for analysing flusurvey data

library('flusurvey')
library('tidyverse')

dt <- extract_data("flusurvey_raw_2010_2018.rds", years=2012:2015,
                   surveys=c("background", "symptom"),
                   clean=c("limit.season", "n.reports",
                           "unsuccessful.join"))

dt %<>% ## transform variables
    mutate(date.vaccine=as.character(date.vaccine)) %>%
    mutate(date.vaccine=if_else(date.vaccine == "", NA_character_, date.vaccine)) %>%
    mutate(date.vaccine=as.Date(date.vaccine)) %>%
    mutate(season.id=paste(season, participant_id, sep="."))

## create data set for participant numbers
dbg <- dt %>%
  filter(!duplicated(background.id)) %>%
  group_by(participant_id, season) %>%
  mutate(nsurvey.season=n()) %>%
  ungroup %>%
  group_by(participant_id) %>%
  mutate(nseason=length(unique(season))) %>%
  ungroup

## number of seasons per participant
dbg %>%
  filter(!duplicated(participant_id)) %>%
  group_by(nseason) %>%
  summarise(n())

## number of participants per year
dbg %>%
  filter(!duplicated(season.id)) %>%
  group_by(season) %>%
  summarise(n())

## date of vaccine present
dbg %>%
  filter(!duplicated(participant_id)) %>%
  group_by(region) %>%
  summarise(n()) %>%
  remove_rownames

ili.previous.season <- dt %>%
    mutate(season=season-1) %>%
    group_by(participant_id, season) %>%
    summarise(nb.ili.previous=sum(ili=="t", na.rm=TRUE),
              nb.reports.previous=length(ili))

## create data set for analysis
dta <- dt %>%
    group_by(participant_id, season) %>%
    mutate(first.questionnaire=min(date),
           vaccination.updated=if_else((length(unique(date.vaccine))>1), "t", "f"),
           nb.reports=length(ili)) %>%
    filter(row_number()==n()) %>%
    ungroup %>%
    left_join(ili.previous.season) %>%
    mutate(vaccinated=!is.na(date.vaccine),
           any.medication=if_else(no.medication=="t", "f", "t")) %>%
    arrange(user, participant_id, season) %>%
    select(user, participant_id, season, gender, age, birthmonth, region, main.activity, occupation, highest.education, frequent.contact.children, frequent.contact.elderly, frequent.contact.patients, frequent.contact.none, nb.household, nb.household.children, transport, vaccine.this.year, date.vaccine, vaccination.updated, starts_with("why.vaccine"), starts_with("why.not.vaccine"), vaccine.last.year, any.medication, starts_with("medication"), starts_with("risk"), pregnant, smoke, first.questionnaire, nb.ili.previous, nb.reports.previous)

dta_pre_phe <- dta %>%
    filter(season <= 2015)

## part_sample <- sample(unique(dta_pre_phe$participant_id), 100)

## dta_sample <- dta_pre_phe %>%
##     filter(participant_id %in% part_sample)

write_csv(dta_pre_phe, "flusurvey.csv")