In R4EPI/sitrep: Report templates and helper functions for applied epidemiology
Introduction to this template
This is a template which can be used to create a report from a vaccination
coverage survey from the long dictionary.
- There are sections for reading and cleaning data, followed by
weighting calculations and then survey analysis.
- For a more detailed explanation of this template, please visit https://r4epis.netlify.com/surveys
-
Feedback and suggestions are welcome at the GitHub issues page
-
Text within <! > will not show in your final document. These comments are used
to explain the template. You can delete them if you want.
Installing and loading required packages
## hide all code chunks in the output, but show errors
knitr::opts_chunk$set(echo = FALSE, error = TRUE, fig.width = 6*1.25, fig.height = 6)
# Ensures the package "pacman" is installed
if (!require("pacman")) {
install.packages("pacman") }
# Install and load required packages for this template
pacman::p_load(
knitr, # create output docs
here, # find your files
rio, # for importing data
janitor, # clean/shape data
dplyr, # clean/shape data
tidyr, # clean/shape data
forcats, # manipulate and rearrange factors
stringr, # manipulate texts
ggplot2, # create plots and charts
apyramid, # plotting age pyramids
sitrep, # MSF field epi functions
survey, # for survey functions
srvyr, # dplyr wrapper for survey package
gtsummary, # produce tables
flextable, # for styling output tables
labelled, # add labels to variables
matchmaker, # recode datasets with dictionaries
lubridate, # working with dates
parsedate # guessing dates
)
## set default text size to 18 for plots
## give classic black/white axes for plots
ggplot2::theme_set(theme_classic(base_size = 18))
## generates MSF standard dictionary for Kobo
study_data_dict <- msf_dict_survey("Vaccination_long")
## generates MSF standard dictionary for Kobo in long format (for recoding)
study_data_dict_long <- msf_dict_survey(disease = "Vaccination_long", compact = FALSE)
## generates a fake dataset for use as an example in this template
## this dataset already has household and individual levels merged
study_data_raw <- gen_data(dictionary = "Vaccination_long",
varnames = "name",
numcases = 1000)
### Read in data ---------------------------------------------------------------
## Excel file ------------------------------------------------------------------
## read in household data sheet
# study_data_hh <- rio::import(here::here("vaccination_survey.xlsx"),
# which = "vaccination_survey_erb", na = "")
## read in individual level data sheet
# study_data_indiv <- rio::import(here::here("vaccination_survey.xlsx",
# which = "r1", na = "")
## Excel file with password ----------------------------------------------------
## Use this section if your Excel has a password.
# install.packages(c("excel.link", "askpass"))
# library(excel.link)
# study_data_hh <- xl.read.file(here::here("vaccination_survey.xlsx"),
# xl.sheet = "vaccination_survey_erb",
# password = askpass::askpass(prompt = "please enter file password"))
# study_data_indiv <- xl.read.file(here::here("vaccination_survey.xlsx"),
# xl.sheet = "r1",
# password = askpass::askpass(prompt = "please enter file password"))
## Excel file ------------------------------------------------------------------
## to read in a specific sheet use "which"
# study_data_hh <- rio::import(here::here("vaccination_survey.xlsx"), which = "Sheet1")
## Excel file -- Specific range ------------------------------------------------
## you can specify a range in an excel sheet.
# study_data_hh <- rio::import(here::here("vaccination_survey.xlsx"), range = "B2:J102")
## Excel file with password ----------------------------------------------------
## use this section if your Excel has a password.
# install.packages(c("excel.link", "askpass"))
# library(excel.link)
# study_data_hh <- xl.read.file(here::here("vaccination_survey.xlsx"),
# xl.sheet = "Sheet1",
# password = askpass::askpass(prompt = "please enter file password"))
## CSV file --------------------------------------------------------------------
# study_data_hh <- rio::import(here::here("vaccination_survey.csv"))
## Stata data file -------------------------------------------------------------
# study_data_hh <- rio::import(here::here("vaccination_survey.dat"))
## join the individual and household data to form a complete data set
#study_data_raw <- left_join(study_data_hh, study_data_indiv, by = c("_index" = "_parent_index"))
## Data dictionary -------------------------------------------------------------
## read in a kobo data dictionary
## important to specify template is FALSE (otherwise you get the generic dict)
# study_data_dict <- msf_dict_survey(name = here::here("vaccination_survey_dict.xlsx"),
# template = FALSE)
## look at the dictionary by uncommenting the line below
# View(study_data_dict)
## Clean column names ----------------------------------------------------------
## This step fixes the column names so they are easy to use in R.
## make a copy of your original dataset and name it study_data_cleaned
study_data_cleaned <- study_data_raw
## Sometimes you want to rename specific variables
## For example we want to add details to violance_natur_other
## This is because otherwise it will be duplicated after we clean names below
## The formula for this is rename(data, NEW_NAME = OLD_NAME).
## You can add multiple column name recodes by simply separating them with a comma
# study_data_cleaned <- rename(study_data_cleaned,
# age_in_years = q10_age_yr,
# age_in_months = q55_age_mth)
## define clean variable names using clean_names from the janitor package.
study_data_cleaned <- janitor::clean_names(study_data_cleaned)
## create a unique identifier by combining indeces of the two levels
## x and y are added on to the end of variables that are duplicated
## (in this case, parent and child index)
# study_data_cleaned <- study_data_cleaned %>%
# mutate(uid = str_glue("{index}_{index_y}"))
## MSF Survey Dictionary ----------------------------------------------------------
## generates MSF standard dictionary for Kobo
# study_data_dict <- msf_dict_survey("vaccination_long")
## generates MSF standard dictionary for Kobo in long format (for recoding)
# study_data_dict_long <- msf_dict_survey(disease = "vaccination_long", compact = FALSE)
## look at the standard dictionary by uncommenting the line below
# View(study_data_dict)
## You will need to recode your variables to match the data dictionary. This is
## addressed below.
## Clean column names ----------------------------------------------------------
## This step fixes the column names so they are easy to use in R.
## make a copy of your orginal dataset and name it study_data_cleaned
# study_data_cleaned <- study_data_raw
## define clean variable names using clean_names from the janitor package.
# study_data_cleaned <- janitor::clean_names(study_data_cleaned)
## Match column names ---------------------------------------------------------
## This step helps you match your variables to the standard variables.
## This step will require some patience. Courage!
## Use the function msf_dict_rename_helper() to create a template based on the
## standard dictionary. This will copy a rename command like the one above to your
## clipboard.
# msf_dict_rename_helper("vaccination_long", varnames = "name")
## Paste the result below and your column names to the matching variable.
## Be careful! You still need to be aware of what each variable means and what
## values it takes.
## If there are any variables that are in the MSF dictionary that are not in
## your data set, then you should comment them out, but be aware that some
## analyses may not run because of this.
## PASTE HERE
## Here is an EXAMPLE for changing a few specific names. function. In this
## example, we have the columns "gender" and "age" that we want to rename as
## "sex" and "age_years".
## The formula for this is rename(data, NEW_NAME = OLD_NAME).
# study_data_cleaned <- rename(study_data_cleaned,
# sex = gender, # TEXT
# age_years = age # INTEGER_POSITIVE
# )
## Read data -------------------------------------------------------------------
## This step reads in your population data from Excel.
## You may need to rename your columns.
# population_data <- rio::import(here::here("population.xlsx"), which = "Sheet1")
## repeat preparation steps as appropriate
## Enter counts directly -------------------------------------------------------
## Below is an example of how to enter population counts by groups.
# population_data_age <- gen_population(
# groups = c("0-2", "3-14", "15-29", "30-44", "45+"),
# counts = c(3600, 18110, 13600, 8080, 6600),
# strata = NULL) %>%
# rename(age_group = groups,
# population = n)
## Create counts from proportions ----------------------------------------------
## This step helps you estimate sub-group size with proportions.
## You need to replace the total_pop and proportions. You can change the groups
## to fit your needs.
## Here we repeat the steps for two regions (district A and B) then bind the two
## together
## generate population data by age groups in years for district A
population_data_age_district_a <- gen_population(total_pop = 10000, # set total population
groups = c("0", "1", "2", "3", "4", "5+"), # set groups
proportions = c(0.0164, 0.0164, 0.015, 0.015,
0.015, 0.1386), # set proportions for each group
strata = c("Male", "Female")) %>% # stratify by gender
rename(age_group = groups, # rename columns (NEW NAME = OLD NAME)
sex = strata,
population = n) %>%
mutate(health_district = "district_a") # add a column to identify region
## generate population data by age groups in years for district B
population_data_age_district_b <- gen_population(total_pop = 10000, # set total population
groups = c("0", "1", "2", "3", "4", "5+"), # set groups
proportions = c(0.0164, 0.0164, 0.015, 0.015,
0.015, 0.1386), # set proportions for each group
strata = c("Male", "Female")) %>% # stratify by gender
rename(age_group = groups, # rename columns (NEW NAME = OLD NAME)
sex = strata,
population = n) %>%
mutate(health_district = "district_b") # add a column to identify region
## bind region population data together to get overall population
population_data_age <- bind_rows(population_data_age_district_a,
population_data_age_district_b)
cluster_counts <- tibble(cluster = c("Village 1", "Village 2", "Village 3", "Village 4",
"Village 5", "Village 6", "Village 7", "Village 8",
"Village 9", "Village 10"),
households = c(700, 400, 600, 500, 300,
800, 700, 400, 500, 500))
## view the first ten rows of data
head(study_data_raw, n = 10)
## view your whole dataset interactively (in an excel style format)
View(study_data_raw)
## overview of variable types and contents
str(study_data_raw)
## get summary:
## mean, median and max values of variables
## counts for categorical variables
## also gives number of NAs
summary(study_data_raw)
## view unique values contained in variables
## you can run this for any column -- just replace the column name
unique(study_data_raw$sex)
## check for logical date inconsistencies
## for example check vaccine age greater than current age and return corresponding IDs
# study_data_raw %>%
# filter(age_routine_vacc > age_months) %>%
# select("uid")
## use the dfSummary function in combination with view
## note that view is not capitalised with this package
# summarytools::dfSummary(linelist_cleaned) %>%
# summarytools::view()
## Kobo standard data --------------------------------------------------------
## If you got your data from Kobo, use this portion of the code.
## If not, comment this section out and use the below.
## make sure all date variables are formatted as dates
## get the names of variables which are dates
DATEVARS <- study_data_dict %>%
filter(type == "date") %>%
filter(name %in% names(study_data_cleaned)) %>%
## filter to match the column names of your data
pull(name) # select date vars
## find if there are date variables which are completely empty
EMPTY_DATEVARS <- purrr::map(DATEVARS, ~all(
is.na(study_data_cleaned[[.x]])
)) %>%
unlist() %>%
which()
## remove exclude the names of variables which are completely emptys
DATEVARS <- DATEVARS[-EMPTY_DATEVARS]
## change to dates
## use the parse_date() function to make a first pass at date variables.
## parse_date() produces a complicated format - so we use as.Date() to simplify
study_data_cleaned <- study_data_cleaned %>%
mutate(
across(.cols = all_of(DATEVARS),
.fns = ~parsedate::parse_date(.x) %>% as.Date()))
## Non-Kobo data -------------------------------------------------------------
## Use this section if you did not have Kobo data.
## use the parse_date() function to make a first pass at date variables.
## parse_date() produces a complicated format - so we use as.Date() to simplify
# study_data_cleaned <- study_data_cleaned %>%
# mutate(
# across(.cols = matches("date|Date"),
# .fns = ~parsedate::parse_date(.x) %>% as.Date()))
## Fix wrong dates -------------------------------------------------------------
## Some dates will be unrealistic or wrong.
## Here is an example of how to manually fix dates.
## Look at your data and edit as needed.
## set specific unrealistic dates to NA
# study_data_cleaned <- mutate(study_data_cleaned,
# date < as.Date("2017-11-01") ~ as.Date(NA),
# date == as.Date("2081-01-01") ~ as.Date("2018-01-01"))
## Age group variables ----------------------------------------------------------
## This step shows you how to create categorical variables from numeric variables.
## We have some intermediate steps on the way.
## make sure age is an integer
study_data_cleaned <- study_data_cleaned %>%
mutate(age_years = as.integer(age_years),
age_months = as.integer(age_months))
## create an age group variable (combining those over 5 years)
study_data_cleaned <- study_data_cleaned %>%
mutate(age_group = factor(
if_else(age_years >= 5,
"5+",
as.character(age_years)
)
))
## create an age group variable by specifying categorical breaks (of years)
# study_data_cleaned <- study_data_cleaned %>%
# mutate(age_group = age_categories(age_years,
# breakers = c(0, 3, 15, 30, 45)
# ))
## create age group variable for under 1 years based on months
study_data_cleaned <- study_data_cleaned %>%
mutate(age_group_mon = age_categories(study_data_cleaned$age_months,
breakers = c(0, 6, 9, 12),
ceiling = TRUE))
## alternatively, create an age group variable specify a sequence
# study_data_cleaned$age_group <- age_categories(study_data_cleaned$age,
# lower = 0,
# upper = 100,
# by = 10)
## If you already have an age group variable defined, you should manually
## arrange the categories
# study_data_cleaned$age_group <- factor(study_data_cleaned$age_group,
# c("0-4y", "5-14y", "15-29y", "30-44y", "45+y"))
## to combine different age categories use the following function
## this prioritises the smaller unit, meaning:
## it will order your age_category from smallest to largest unit
## i.e. days first then months then years (depending on which are given)
study_data_cleaned <- group_age_categories(study_data_cleaned,
years = age_group,
months = age_group_mon)
## recode values with matchmaker package (value labels cant be used for analysis)
study_data_cleaned <- match_df(study_data_cleaned,
study_data_dict_long,
from = "option_name",
to = "option_label_english",
by = "name",
order = "option_order_in_set")
## Change a yes/no variable in to TRUE/FALSE
## create a new variable called consent
## where the old one is yes place TRUE in the new one
## (this could also be done with the if_else or case_when function -
## but this option is shorter)
study_data_cleaned <- study_data_cleaned %>%
mutate(consent = consent == "Yes")
## Recode character variables
## This step shows how to fix misspellings in the geographic region variable.
## Ideally, you want these values to match and population data!
# study_data_cleaned <- study_data_cleaned %>%
# mutate(village_name = case_when(
# village_name == "Valliages 1" ~ "village 1",
# village_name == "Village1" ~ "village 1",
# village_name == "Town 3" ~ "village 3"
# village_name == "Town3" ~ "village 3",
# TRUE ~ as.character(village_name))
# ))
## create a character variable based off groups of a different variable
study_data_cleaned <- study_data_cleaned %>%
mutate(health_district = case_when(
cluster_number %in% c(1:5) ~ "district_a",
TRUE ~ "district_b"
))
## create a new variable by reordering the levels of a factor
study_data_cleaned <- study_data_cleaned %>%
mutate(no_consent = fct_relevel(no_consent_reason,
"There is no benefit to me and my family",
"I do not have time",
"I have had a negative experience with a previous survey.",
"Other")
)
## create simplified groupings for each of the vaccine status variables
## simplified, TRUE (regardless of date) otherwise FALSE (or NA if missing)
## select the variables from dictionary for vaccines by card
CARDVARS <- study_data_dict %>%
filter(type == "vaccine_card") %>%
filter(name %in% names(study_data_cleaned)) %>%
## filter to match the column names of your data
pull(name) # select vars
## create simplified variables
study_data_cleaned <- study_data_cleaned %>%
mutate(
## for each of the vaccine variables
dplyr::across(all_of(CARDVARS),
## if a row contains "Yes" return TRUE
~str_detect(.x, "Yes"),
## rename with original name with "_simple" on the end
.names = "{.col}_simple")
)
## create a list of vaccines with simplified variable names
CARDVARSSIMP <- str_glue("{CARDVARS}_simple")
## create a list of vaccines with date on the front of variable names
CARDVARSDATE <- str_glue("date_{CARDVARS}")
## years since vaccination variables -------------------------------------------
## adapt date of vaccination and then create a new time_since variable
study_data_cleaned <- study_data_cleaned %>%
mutate(
across(
all_of(CARDVARSDATE),
~as.numeric(date - .x),
## rename with original name with "_simple" on the end
.names = "{.col}_time"
))
## create a grouping variable for years since vaccination
study_data_cleaned <- study_data_cleaned %>%
mutate(
across(
all_of(str_glue("{CARDVARSDATE}_time")),
~age_categories(.x,
lower = 0, upper = 5, by = 1) %>%
## remove what comes after a dash (i.e. repeated year)
str_remove("-.*") %>%
factor(levels = c("0", "1", "2", "3", "4", "5+")),
## rename with original name with "_grp" on the end
.names = "{.col}_grp"
)
)
## create a list of vaccines with time group on the of variable names
CARDVARSTIMEGRP <- str_glue("{CARDVARSDATE}_time_grp")
## create variable names with labelled
## define a list of names and labels based on the dictionary
var_labels <- setNames(
## add variable labels as a list
as.list(study_data_dict$short_name),
## name the list with the current variable names
study_data_dict$name)
## add the variable labels to dataset
study_data_cleaned <- study_data_cleaned %>%
set_variable_labels(
## set the labels with the object defined above
.labels = var_labels,
## do not throw an error if some names dont match
## not all names in the dictionary appear in the dataset
.strict = FALSE)
## it is possible to update individual variables manually too
study_data_cleaned <- study_data_cleaned %>%
set_variable_labels(
## variable name = variable label
age_years = "Age (years)",
age_group = "Age group (years)",
age_months = "Age (months)",
age_group_mon = "Age group (months)",
age_category = "Age category"
)
## add in variables for newly created variables (based on old names)
## e.g. here label the simple vacc variables with the original variable labels
## we use c() to repeat this for the time since vaccination variable too
vacc_var_labels <- setNames(
## add variable labels as a list
as.list(
c(study_data_dict$short_name[study_data_dict$type == "vaccine_card"],
study_data_dict$short_name[study_data_dict$type == "vaccine_card"])),
## name the list with the variable names of interest
c(CARDVARSSIMP, CARDVARSTIMEGRP)
)
## add the variable labels to dataset
study_data_cleaned <- study_data_cleaned %>%
set_variable_labels(
## set the labels with the object defined above
.labels = vacc_var_labels,
## do not throw an error if some names dont match
## not all names in the dictionary appear in the dataset
.strict = FALSE)
## Adding labels for variables with multiple responses are allowed
## In Kobo - multioption variables apear once in the dictionary
## but exports have a character variable with all responses and then a
## binary (1/0) variable for each of the various options
## These are not automatically labelled from the dictionary above
## first create a list of the binary variable names by combining with the
## dictionary defined options for that variable
reason_vars <- study_data_dict_long %>%
filter(type == "reasons") %>%
select(option_name, option_label_english) %>%
mutate(option_name = str_glue("reason_not_all_vacc_{option_name}"))
## add in variables for newly created variables (based on old names)
reason_var_labels <- setNames(
## add variable labels as a list
as.list(reason_vars$option_label_english),
## name the list with the variable names of interest
reason_vars$option_name
)
## add the variable labels to dataset
study_data_cleaned <- study_data_cleaned %>%
set_variable_labels(
## set the labels with the object defined above
.labels = reason_var_labels,
## do not throw an error if some names dont match
## not all names in the dictionary appear in the dataset
.strict = FALSE)
## Drop unused rows -----------------------------------------------------------
## store the cases that you drop so you can describe them (e.g. non-consenting)
dropped <- study_data_cleaned %>%
filter(!consent| village_name == "other")
## use the dropped cases to remove the unused rows from the survey data set
study_data_cleaned <- anti_join(study_data_cleaned, dropped, by = names(dropped))
## Drop columns ----------------------------------------------------------------
## OPTIONAL: This step shows you how you can remove certain variables.
## study_data_cleaned <- select(study_data_cleaned, -c("age_years", "sex"))
## OPTIONAL: if you want to inspect certain variables, you can select these by
## name or column number. This example creates a reduced dataset for the first
## three columns, age_years, and sex.
# study_data_reduced <- select(study_data_cleaned, c(1:3, "age_years", "sex")
## option 1: only keep the first occurrence of duplicate case
study_data_cleaned <- study_data_cleaned %>%
## find duplicates based on case number, sex and age group
## only keep the first occurrence
distinct(uid, sex, age_group, .keep_all = TRUE)
# ## option 2: create flagging variables for duplicates (then use to browse)
#
# study_data_cleaned <- study_data_cleaned %>%
# ## choose which variables to use for finding unique rows
# group_by(uid, sex, age_group) %>%
# mutate(
# ## get the number of times duplicate occurs
# num_dupes = n(),
# duped = if_else(num_dupes > 1 , TRUE, FALSE)
# )
#
# ## browse duplicates based on flagging variables
# study_data_cleaned %>%
# ## only keep rows that are duplicated
# filter(duped) %>%
# ## arrange by variables of interest
# arrange(uid, sex, age_group) %>%
# View()
#
# ## filter duplicates to only keep the row with the earlier entry
# study_data_cleaned %>%
# ## choose which variables to use for finding unique rows
# group_by(uid, sex, age_group) %>%
# ## sort to have the earliest date by person first
# arrange(as.Date(submission_time)) %>%
# ## only keep the earliest row
# slice(1)
## stratified ------------------------------------------------------------------
## create a variable called "surv_weight_strata"
## contains weights for each individual - by age group, sex and health district
study_data_cleaned <- add_weights_strata(x = study_data_cleaned,
p = population_data_age,
surv_weight = "surv_weight_strata",
surv_weight_ID = "surv_weight_ID_strata",
age_group, sex, health_district)
## cluster ---------------------------------------------------------------------
## get the number of people of individuals interviewed per household
## adds a variable with counts of the household (parent) index variable
study_data_cleaned <- study_data_cleaned %>%
add_count(index, name = "interviewed")
## create cluster weights
study_data_cleaned <- add_weights_cluster(x = study_data_cleaned,
cl = cluster_counts,
eligible = number_children,
interviewed = interviewed,
cluster_x = village_name,
cluster_cl = cluster,
household_x = index,
household_cl = households,
surv_weight = "surv_weight_cluster",
surv_weight_ID = "surv_weight_ID_cluster",
ignore_cluster = FALSE,
ignore_household = FALSE)
## stratified and cluster ------------------------------------------------------
## create a survey weight for cluster and strata
study_data_cleaned <- study_data_cleaned %>%
mutate(surv_weight_cluster_strata = surv_weight_strata * surv_weight_cluster)
## simple random ---------------------------------------------------------------
survey_design_simple <- study_data_cleaned %>%
as_survey_design(ids = 1, # 1 for no cluster ids
weights = NULL, # No weight added
strata = NULL # sampling was simple (no strata)
)
## stratified ------------------------------------------------------------------
survey_design_strata <- study_data_cleaned %>%
as_survey_design(ids = 1, # 1 for no cluster ids
weights = surv_weight_strata, # weight variable created above
strata = health_district # sampling was stratified by district
)
## cluster ---------------------------------------------------------------------
survey_design_cluster <- study_data_cleaned %>%
as_survey_design(ids = village_name, # cluster ids
weights = surv_weight_cluster, # weight variable created above
strata = NULL # sampling was simple (no strata)
)
## stratified cluster ----------------------------------------------------------
survey_design <- study_data_cleaned %>%
as_survey_design(ids = village_name, # cluster ids
weights = surv_weight_cluster_strata, # weight variable created above
strata = health_district # sampling was stratified by district
)
# rio::export(study_data_cleaned, here::here("data", str_glue("study_data_cleaned_{Sys.Date()}.xlsx")))
Results
Survey inclusion
## get counts of number of clusters
num_clus <- study_data_cleaned %>%
## trim data to unique clusters
distinct(cluster_number) %>%
## get number of rows (count how many unique)
nrow()
## get counts of number households
num_hh <- study_data_cleaned %>%
## get unique houses by cluster
distinct(cluster_number, household_number) %>%
## get number of rounds (count how many unique)
nrow()
We included r num_hh households across r num_clus clusters in this survey analysis.
Among the r nrow(dropped) individuals without consent to participate in the survey,
the reasons for refusal are described below.
## using the dataset with dropped individuals
dropped %>%
## only keep reasons for no consent
select(no_consent_reason) %>%
## make NAs a factor level called missing
## to make the proportion of the total, including the missings
mutate(no_consent_reason = fct_explicit_na(no_consent_reason,
na_level = "Missing")) %>%
## get counts and proportions
tbl_summary() %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
## get counts of the number of households per cluster
clustersize <- study_data_cleaned %>%
## trim data to only unique households within each cluster
distinct(cluster_number, household_number) %>%
## count the number of households within each cluster
count(cluster_number) %>%
pull(n)
## get the median number of households per cluster
clustermed <- median(clustersize)
## get the min and max number of households per cluster
## paste these together seperated by a dash
clusterrange <- str_c(range(clustersize), collapse = "--")
## get counts of children per household
## do this by cluster as household IDs are only unique within clusters
hhsize <- study_data_cleaned %>%
count(cluster_number, household_number) %>%
pull(n)
## get median number of children per household
hhmed <- median(hhsize)
## get the min and max number of children per household
## paste these together seperated by a dash
hhrange <- str_c(range(hhsize), collapse = "--")
# get standard deviation
hhsd <- round(sd(hhsize), digits = 1)
The median number of households per cluster was
r clustermed, with a range of r clusterrange. The median number of children
per household was r hhmed (range: r hhrange, standard deviation: r hhsd).
Demographic information
In total we included r nrow(study_data_cleaned) in the survey analysis.
The age break down and a comparison with the source population is shown below.
## counts and props of the study population
ag <- tabyl(study_data_cleaned, age_group, show_na = FALSE) %>%
mutate(n_total = sum(n),
age_group = fct_inorder(age_group))
## counts and props of the source population
propcount <- population_data_age %>%
group_by(age_group) %>%
tally(population) %>%
mutate(percent = n / sum(n))
## bind together the columns of two tables, group by age, and perform a
## binomial test to see if n/total is significantly different from population
## proportion.
## suffix here adds to text to the end of columns in each of the two datasets
left_join(ag, propcount, by = "age_group", suffix = c("", "_pop")) %>%
group_by(age_group) %>%
## broom::tidy(binom.test()) makes a data frame out of the binomial test and
## will add the variables p.value, parameter, conf.low, conf.high, method, and
## alternative. We will only use p.value here. You can include other
## columns if you want to report confidence intervals
mutate(binom = list(broom::tidy(binom.test(n, n_total, percent_pop)))) %>%
unnest(cols = c(binom)) %>% # important for expanding the binom.test data frame
mutate(
across(.cols = contains("percent"),
.fns = ~.x * 100)) %>%
## Adjusting the p-values to correct for false positives
## (because testing multiple age groups). This will only make
## a difference if you have many age categories
mutate(p.value = p.adjust(p.value, method = "holm")) %>%
## Only show p-values over 0.001 (those under report as <0.001)
mutate(p.value = ifelse(p.value < 0.001, "<0.001", as.character(round(p.value, 3)))) %>%
## rename the columns appropriately
select(
"Age group" = age_group,
"Study population (n)" = n,
"Study population (%)" = percent,
"Source population (n)" = n_pop,
"Source population (%)" = percent_pop,
"P-value" = p.value
) %>%
# produce styled output table with auto-adjusted column widths with {flextable}
qflextable() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit") %>%
## set to only show 1 decimal place
colformat_double(digits = 1)
## compute the median age
medage <- median(study_data_cleaned$age_years)
## paste the lower and upper quartile together
iqr <- str_c( # basically copy paste togehter the following
## calculate the 25% and 75% of distribution, with missings removed
quantile(
study_data_cleaned$age_years,
c(0.25, 0.75),
na.rm = TRUE),
## between lower and upper place an en-dash
collapse = "--")
## compute overall sex ratio
sex_ratio <- study_data_cleaned %>%
count(sex) %>%
pivot_wider(names_from = sex, values_from = n) %>%
mutate(ratio = round(Male/Female, digits = 3)) %>%
pull(ratio)
## compute sex ratios by age group
sex_ratio_age <- study_data_cleaned %>%
count(age_group, sex) %>%
pivot_wider(names_from = sex, values_from = n) %>%
mutate(ratio = round(Male/Female, digits = 3)) %>%
select(age_group, ratio)
## sort table by ascending ratio then select the lowest (first)
min_sex_ratio_age <- arrange(sex_ratio_age, ratio) %>% slice(1)
Among the r nrow(study_data_cleaned) surveyed individuals, there were
r fmt_count(study_data_cleaned, sex == "Female") females and
r fmt_count(study_data_cleaned, sex == "Male") males (unweighted). The male to
female ratio was r sex_ratio in the surveyed population. The lowest male to
female ratio was r min_sex_ratio_age$ratio
in the r min_sex_ratio_age$age_group year age group.
The median age of surveyed individuals was r medage years (Q1-Q3 of r iqr
years). Children under five years of age made up
r fmt_count(study_data_cleaned, age_years < 5)of the surveyed individuals.
The highest number of surveyed individuals (unweighted) were in the
r table(study_data_cleaned$age_group) %>% which.max() %>% names()
year age group.
Unweighted age distribution of population by year age group and gender.
# get cross tabulated counts and proportions
study_data_cleaned %>%
tbl_cross(
row = age_group,
col = sex,
percent = "cell") %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
Unweighted age distribution of population by age category (month and year) and gender.
# get cross tabulated counts and proportions
study_data_cleaned %>%
tbl_cross(
row = age_category,
col = sex,
percent = "cell") %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
There were r fmt_count(dropped, is.na(sex)) cases missing information on sex and
r fmt_count(dropped, is.na(age_group)) missing age group.
Unweighted age and gender distribution of household population covered by the survey.
<!-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/// age_pyramid \\
This chunk creates an unweighted (using study_data_cleaned) age/sex pyramid
of your cases.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -->
age_pyramid(study_data_cleaned,
age_group,
split_by = sex,
proportional = TRUE) +
labs(y = "Proportion", x = "Age group (years)") + # change axis labels
theme(legend.position = "bottom", # move legend to bottom
legend.title = element_blank(), # remove title
text = element_text(size = 18) # change text size
)
Unweighted age and gender distribution, stratified by health district,
of household population covered by the survey.
<!-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/// age_pyramid_strata \\
This chunk creates an unweighted (using study_data_cleaned) age/sex pyramid
of your cases, stratified by health district.
If you have a stratified survey design this may be useful for visualising
if you have an excess of representation in
either sex or gender in any of your survey strata.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -->
age_pyramid(study_data_cleaned,
age_group,
split_by = sex,
stack_by = health_district,
proportional = TRUE,
pal = c("red", "blue")) +
labs(y = "Proportion", x = "Age group (years)") + # change axis labels
theme(legend.position = "bottom", # move legend to bottom
legend.title = element_blank(), # remove title
text = element_text(size = 18) # change text size
)
Weighted age and gender distribution of household population covered by the survey.
age_pyramid(survey_design,
age_group,
split_by = sex,
proportional = TRUE) +
labs(y = "Proportion", x = "Age group (years)") + # change axis labels
theme(legend.position = "bottom", # move legend to bottom
legend.title = element_blank(), # remove title
text = element_text(size = 18) # change text size
)
Vaccination coverage
Weighted vaccination coverage; accepting equal validity from self-reported and
card-reported vaccination status, and detailing availability of vaccination cards
and health facility records
survey_design %>%
## only keep variables of interest
select(any_vaccine, vaccine_card, hf_records) %>%
## only display one line for each variable (with counts for "Yes")
tbl_svysummary(value = list(everything() ~ "Yes"),
statistic = everything() ~ "{n} ({p}%) [{deff}]") %>%
## add the weighted total
add_n() %>%
## add in confidence intervals
add_ci() %>%
## modify the column headers
modify_header(
list(
n ~ "**Weighted total (N)**",
stat_0 ~ "**Weighted Count (n)**"
)
) %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
Vaccination history verbal
Weighted overall vaccination completeness; based on surveyor assessments of
care-giver verbal accounts and on care-giver's own assessment
survey_design %>%
## only keep variables of interest
select(all_vaccines_survey_woc, all_vaccines_caretaker_woc) %>%
## only display one line for each variable (with counts for "Yes")
tbl_svysummary(value = list(everything() ~ "Yes")) %>%
## add the weighted total
add_n() %>%
## add in confidence intervals
add_ci() %>%
## modify the column headers
modify_header(
list(
n ~ "**Weighted total (N)**",
stat_0 ~ "**Weighted Count (n)**"
)
) %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
Weighted vaccination coverage by disease based on verbal history
survey_design %>%
## only keep variables of interest
select(injection_upper_arm,
scar_present,
poliodrop_woc,
poliodrop_hf_woc,
poliodrop_camp_woc,
polioinjection_woc,
num_polioinjection_woc,
quad_penta_woc,
num_quad_penta_woc,
pcv_woc,num_pcv_woc,
measles_woc,
num_measles_woc,
measles_hf_woc,
measles_camp_woc,
yellow_fever_woc,
num_yellow_fever_woc,
yellow_fever_hf_woc,
yellow_fever_camp_woc,
rotavirus_woc,
num_rotavirus_woc) %>%
## only display one line for each variable (with counts for "Yes")
## also specify for the scar_present variable to show the relevant level
tbl_svysummary(value = list(where(is.factor) ~ "Yes",
scar_present ~ "Scar present")) %>%
## add the weighted total
add_n() %>%
## add in confidence intervals
add_ci() %>%
## modify the column headers
modify_header(
list(
n ~ "**Weighted total (N)**",
stat_0 ~ "**Weighted Count (n)**"
)
) %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
Vaccination history with a card
Weighted overall vaccination completeness; based on surveyor assessments of
vaccination cards
survey_design %>%
## only keep variables of interest
select(all_vaccines_survey_card) %>%
## only display one line for each variable (with counts for "Yes")
tbl_svysummary(value = list(everything() ~ "Yes")) %>%
## add the weighted total
add_n() %>%
## add in confidence intervals
add_ci() %>%
## modify the column headers
modify_header(
list(
n ~ "**Weighted total (N)**",
stat_0 ~ "**Weighted Count (n)**"
)
) %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
Weighted vaccination coverage by vaccine and dose based on vaccination cards
survey_design %>%
## only keep variables of interest
select(all_of(CARDVARSSIMP)) %>%
## automatically only displays one line for each variable (with counts for TRUE)
tbl_svysummary() %>%
## add the weighted total
add_n() %>%
## add in confidence intervals
add_ci() %>%
## modify the column headers
modify_header(
list(
n ~ "**Weighted total (N)**",
stat_0 ~ "**Weighted Count (n)**"
)
) %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
Weighted counts of time (in years) since vaccination by various vaccines
based on vaccination cards
survey_design %>%
## only keep variables of interest
select(all_of(CARDVARSTIMEGRP)) %>%
## automatically only displays one line for each variable (with counts for TRUE)
tbl_svysummary() %>%
## add the weighted total
add_n() %>%
## add in confidence intervals
add_ci() %>%
## modify the column headers
modify_header(
list(
n ~ "**Weighted total (N)**",
stat_0 ~ "**Weighted Count (n)**"
)
) %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
Reasons for not vaccinating
Weighted counts and proportions for reasons not vaccinated, individually, among those not vaccinated based on verbal or card assessment
<!-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/// no_vacc_reason \\
The below chunk creates a weighted table of counts proportions for reasons
not vaccinated.
Note that low counts or short observation times may lead to a confidence interval
that crosses zero (i.e. negative) for ratios. These should be interpreted
as if no deaths or recoded to zero (impossible to have negative deaths).
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -->
survey_design %>%
## filter for those not vaccinated
filter(
all_vaccines_survey_card != "Yes") %>%
select(reason_vars$option_name) %>%
tbl_svysummary(value = list(everything() ~ 1)) %>%
## add in confidence intervals
add_ci() %>%
## modify the column headers
modify_header(
list(
stat_0 ~ "Weighted Count (n) \n (N={round(N, digits = 0)})"
)
) %>%
## make variable names bold
bold_labels() %>%
# change to flextable format
as_flex_table() %>%
# make header text bold (using {flextable})
bold(part = "header") %>%
# make your table fit to the maximum width of the word document
set_table_properties(layout = "autofit")
R4EPI/sitrep documentation built on April 5, 2025, 4:51 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Introduction to this template
This is a template which can be used to create a report from a vaccination coverage survey from the long dictionary.
- There are sections for reading and cleaning data, followed by weighting calculations and then survey analysis.
- For a more detailed explanation of this template, please visit https://r4epis.netlify.com/surveys
-
Feedback and suggestions are welcome at the GitHub issues page
-
Text within <! > will not show in your final document. These comments are used to explain the template. You can delete them if you want.
Installing and loading required packages
## hide all code chunks in the output, but show errors knitr::opts_chunk$set(echo = FALSE, error = TRUE, fig.width = 6*1.25, fig.height = 6) # Ensures the package "pacman" is installed if (!require("pacman")) { install.packages("pacman") } # Install and load required packages for this template pacman::p_load( knitr, # create output docs here, # find your files rio, # for importing data janitor, # clean/shape data dplyr, # clean/shape data tidyr, # clean/shape data forcats, # manipulate and rearrange factors stringr, # manipulate texts ggplot2, # create plots and charts apyramid, # plotting age pyramids sitrep, # MSF field epi functions survey, # for survey functions srvyr, # dplyr wrapper for survey package gtsummary, # produce tables flextable, # for styling output tables labelled, # add labels to variables matchmaker, # recode datasets with dictionaries lubridate, # working with dates parsedate # guessing dates ) ## set default text size to 18 for plots ## give classic black/white axes for plots ggplot2::theme_set(theme_classic(base_size = 18))
## generates MSF standard dictionary for Kobo study_data_dict <- msf_dict_survey("Vaccination_long") ## generates MSF standard dictionary for Kobo in long format (for recoding) study_data_dict_long <- msf_dict_survey(disease = "Vaccination_long", compact = FALSE) ## generates a fake dataset for use as an example in this template ## this dataset already has household and individual levels merged study_data_raw <- gen_data(dictionary = "Vaccination_long", varnames = "name", numcases = 1000)
### Read in data --------------------------------------------------------------- ## Excel file ------------------------------------------------------------------ ## read in household data sheet # study_data_hh <- rio::import(here::here("vaccination_survey.xlsx"), # which = "vaccination_survey_erb", na = "") ## read in individual level data sheet # study_data_indiv <- rio::import(here::here("vaccination_survey.xlsx", # which = "r1", na = "") ## Excel file with password ---------------------------------------------------- ## Use this section if your Excel has a password. # install.packages(c("excel.link", "askpass")) # library(excel.link) # study_data_hh <- xl.read.file(here::here("vaccination_survey.xlsx"), # xl.sheet = "vaccination_survey_erb", # password = askpass::askpass(prompt = "please enter file password")) # study_data_indiv <- xl.read.file(here::here("vaccination_survey.xlsx"), # xl.sheet = "r1", # password = askpass::askpass(prompt = "please enter file password"))
## Excel file ------------------------------------------------------------------ ## to read in a specific sheet use "which" # study_data_hh <- rio::import(here::here("vaccination_survey.xlsx"), which = "Sheet1") ## Excel file -- Specific range ------------------------------------------------ ## you can specify a range in an excel sheet. # study_data_hh <- rio::import(here::here("vaccination_survey.xlsx"), range = "B2:J102") ## Excel file with password ---------------------------------------------------- ## use this section if your Excel has a password. # install.packages(c("excel.link", "askpass")) # library(excel.link) # study_data_hh <- xl.read.file(here::here("vaccination_survey.xlsx"), # xl.sheet = "Sheet1", # password = askpass::askpass(prompt = "please enter file password")) ## CSV file -------------------------------------------------------------------- # study_data_hh <- rio::import(here::here("vaccination_survey.csv")) ## Stata data file ------------------------------------------------------------- # study_data_hh <- rio::import(here::here("vaccination_survey.dat"))
## join the individual and household data to form a complete data set #study_data_raw <- left_join(study_data_hh, study_data_indiv, by = c("_index" = "_parent_index"))
## Data dictionary ------------------------------------------------------------- ## read in a kobo data dictionary ## important to specify template is FALSE (otherwise you get the generic dict) # study_data_dict <- msf_dict_survey(name = here::here("vaccination_survey_dict.xlsx"), # template = FALSE) ## look at the dictionary by uncommenting the line below # View(study_data_dict) ## Clean column names ---------------------------------------------------------- ## This step fixes the column names so they are easy to use in R. ## make a copy of your original dataset and name it study_data_cleaned study_data_cleaned <- study_data_raw ## Sometimes you want to rename specific variables ## For example we want to add details to violance_natur_other ## This is because otherwise it will be duplicated after we clean names below ## The formula for this is rename(data, NEW_NAME = OLD_NAME). ## You can add multiple column name recodes by simply separating them with a comma # study_data_cleaned <- rename(study_data_cleaned, # age_in_years = q10_age_yr, # age_in_months = q55_age_mth) ## define clean variable names using clean_names from the janitor package. study_data_cleaned <- janitor::clean_names(study_data_cleaned) ## create a unique identifier by combining indeces of the two levels ## x and y are added on to the end of variables that are duplicated ## (in this case, parent and child index) # study_data_cleaned <- study_data_cleaned %>% # mutate(uid = str_glue("{index}_{index_y}"))
## MSF Survey Dictionary ---------------------------------------------------------- ## generates MSF standard dictionary for Kobo # study_data_dict <- msf_dict_survey("vaccination_long") ## generates MSF standard dictionary for Kobo in long format (for recoding) # study_data_dict_long <- msf_dict_survey(disease = "vaccination_long", compact = FALSE) ## look at the standard dictionary by uncommenting the line below # View(study_data_dict) ## You will need to recode your variables to match the data dictionary. This is ## addressed below. ## Clean column names ---------------------------------------------------------- ## This step fixes the column names so they are easy to use in R. ## make a copy of your orginal dataset and name it study_data_cleaned # study_data_cleaned <- study_data_raw ## define clean variable names using clean_names from the janitor package. # study_data_cleaned <- janitor::clean_names(study_data_cleaned) ## Match column names --------------------------------------------------------- ## This step helps you match your variables to the standard variables. ## This step will require some patience. Courage! ## Use the function msf_dict_rename_helper() to create a template based on the ## standard dictionary. This will copy a rename command like the one above to your ## clipboard. # msf_dict_rename_helper("vaccination_long", varnames = "name") ## Paste the result below and your column names to the matching variable. ## Be careful! You still need to be aware of what each variable means and what ## values it takes. ## If there are any variables that are in the MSF dictionary that are not in ## your data set, then you should comment them out, but be aware that some ## analyses may not run because of this. ## PASTE HERE ## Here is an EXAMPLE for changing a few specific names. function. In this ## example, we have the columns "gender" and "age" that we want to rename as ## "sex" and "age_years". ## The formula for this is rename(data, NEW_NAME = OLD_NAME). # study_data_cleaned <- rename(study_data_cleaned, # sex = gender, # TEXT # age_years = age # INTEGER_POSITIVE # )
## Read data ------------------------------------------------------------------- ## This step reads in your population data from Excel. ## You may need to rename your columns. # population_data <- rio::import(here::here("population.xlsx"), which = "Sheet1") ## repeat preparation steps as appropriate ## Enter counts directly ------------------------------------------------------- ## Below is an example of how to enter population counts by groups. # population_data_age <- gen_population( # groups = c("0-2", "3-14", "15-29", "30-44", "45+"), # counts = c(3600, 18110, 13600, 8080, 6600), # strata = NULL) %>% # rename(age_group = groups, # population = n) ## Create counts from proportions ---------------------------------------------- ## This step helps you estimate sub-group size with proportions. ## You need to replace the total_pop and proportions. You can change the groups ## to fit your needs. ## Here we repeat the steps for two regions (district A and B) then bind the two ## together ## generate population data by age groups in years for district A population_data_age_district_a <- gen_population(total_pop = 10000, # set total population groups = c("0", "1", "2", "3", "4", "5+"), # set groups proportions = c(0.0164, 0.0164, 0.015, 0.015, 0.015, 0.1386), # set proportions for each group strata = c("Male", "Female")) %>% # stratify by gender rename(age_group = groups, # rename columns (NEW NAME = OLD NAME) sex = strata, population = n) %>% mutate(health_district = "district_a") # add a column to identify region ## generate population data by age groups in years for district B population_data_age_district_b <- gen_population(total_pop = 10000, # set total population groups = c("0", "1", "2", "3", "4", "5+"), # set groups proportions = c(0.0164, 0.0164, 0.015, 0.015, 0.015, 0.1386), # set proportions for each group strata = c("Male", "Female")) %>% # stratify by gender rename(age_group = groups, # rename columns (NEW NAME = OLD NAME) sex = strata, population = n) %>% mutate(health_district = "district_b") # add a column to identify region ## bind region population data together to get overall population population_data_age <- bind_rows(population_data_age_district_a, population_data_age_district_b)
cluster_counts <- tibble(cluster = c("Village 1", "Village 2", "Village 3", "Village 4", "Village 5", "Village 6", "Village 7", "Village 8", "Village 9", "Village 10"), households = c(700, 400, 600, 500, 300, 800, 700, 400, 500, 500))
## view the first ten rows of data head(study_data_raw, n = 10) ## view your whole dataset interactively (in an excel style format) View(study_data_raw) ## overview of variable types and contents str(study_data_raw) ## get summary: ## mean, median and max values of variables ## counts for categorical variables ## also gives number of NAs summary(study_data_raw) ## view unique values contained in variables ## you can run this for any column -- just replace the column name unique(study_data_raw$sex) ## check for logical date inconsistencies ## for example check vaccine age greater than current age and return corresponding IDs # study_data_raw %>% # filter(age_routine_vacc > age_months) %>% # select("uid") ## use the dfSummary function in combination with view ## note that view is not capitalised with this package # summarytools::dfSummary(linelist_cleaned) %>% # summarytools::view()
## Kobo standard data -------------------------------------------------------- ## If you got your data from Kobo, use this portion of the code. ## If not, comment this section out and use the below. ## make sure all date variables are formatted as dates ## get the names of variables which are dates DATEVARS <- study_data_dict %>% filter(type == "date") %>% filter(name %in% names(study_data_cleaned)) %>% ## filter to match the column names of your data pull(name) # select date vars ## find if there are date variables which are completely empty EMPTY_DATEVARS <- purrr::map(DATEVARS, ~all( is.na(study_data_cleaned[[.x]]) )) %>% unlist() %>% which() ## remove exclude the names of variables which are completely emptys DATEVARS <- DATEVARS[-EMPTY_DATEVARS] ## change to dates ## use the parse_date() function to make a first pass at date variables. ## parse_date() produces a complicated format - so we use as.Date() to simplify study_data_cleaned <- study_data_cleaned %>% mutate( across(.cols = all_of(DATEVARS), .fns = ~parsedate::parse_date(.x) %>% as.Date())) ## Non-Kobo data ------------------------------------------------------------- ## Use this section if you did not have Kobo data. ## use the parse_date() function to make a first pass at date variables. ## parse_date() produces a complicated format - so we use as.Date() to simplify # study_data_cleaned <- study_data_cleaned %>% # mutate( # across(.cols = matches("date|Date"), # .fns = ~parsedate::parse_date(.x) %>% as.Date())) ## Fix wrong dates ------------------------------------------------------------- ## Some dates will be unrealistic or wrong. ## Here is an example of how to manually fix dates. ## Look at your data and edit as needed. ## set specific unrealistic dates to NA # study_data_cleaned <- mutate(study_data_cleaned, # date < as.Date("2017-11-01") ~ as.Date(NA), # date == as.Date("2081-01-01") ~ as.Date("2018-01-01"))
## Age group variables ---------------------------------------------------------- ## This step shows you how to create categorical variables from numeric variables. ## We have some intermediate steps on the way. ## make sure age is an integer study_data_cleaned <- study_data_cleaned %>% mutate(age_years = as.integer(age_years), age_months = as.integer(age_months)) ## create an age group variable (combining those over 5 years) study_data_cleaned <- study_data_cleaned %>% mutate(age_group = factor( if_else(age_years >= 5, "5+", as.character(age_years) ) )) ## create an age group variable by specifying categorical breaks (of years) # study_data_cleaned <- study_data_cleaned %>% # mutate(age_group = age_categories(age_years, # breakers = c(0, 3, 15, 30, 45) # )) ## create age group variable for under 1 years based on months study_data_cleaned <- study_data_cleaned %>% mutate(age_group_mon = age_categories(study_data_cleaned$age_months, breakers = c(0, 6, 9, 12), ceiling = TRUE)) ## alternatively, create an age group variable specify a sequence # study_data_cleaned$age_group <- age_categories(study_data_cleaned$age, # lower = 0, # upper = 100, # by = 10) ## If you already have an age group variable defined, you should manually ## arrange the categories # study_data_cleaned$age_group <- factor(study_data_cleaned$age_group, # c("0-4y", "5-14y", "15-29y", "30-44y", "45+y")) ## to combine different age categories use the following function ## this prioritises the smaller unit, meaning: ## it will order your age_category from smallest to largest unit ## i.e. days first then months then years (depending on which are given) study_data_cleaned <- group_age_categories(study_data_cleaned, years = age_group, months = age_group_mon)
## recode values with matchmaker package (value labels cant be used for analysis) study_data_cleaned <- match_df(study_data_cleaned, study_data_dict_long, from = "option_name", to = "option_label_english", by = "name", order = "option_order_in_set")
## Change a yes/no variable in to TRUE/FALSE ## create a new variable called consent ## where the old one is yes place TRUE in the new one ## (this could also be done with the if_else or case_when function - ## but this option is shorter) study_data_cleaned <- study_data_cleaned %>% mutate(consent = consent == "Yes") ## Recode character variables ## This step shows how to fix misspellings in the geographic region variable. ## Ideally, you want these values to match and population data! # study_data_cleaned <- study_data_cleaned %>% # mutate(village_name = case_when( # village_name == "Valliages 1" ~ "village 1", # village_name == "Village1" ~ "village 1", # village_name == "Town 3" ~ "village 3" # village_name == "Town3" ~ "village 3", # TRUE ~ as.character(village_name)) # )) ## create a character variable based off groups of a different variable study_data_cleaned <- study_data_cleaned %>% mutate(health_district = case_when( cluster_number %in% c(1:5) ~ "district_a", TRUE ~ "district_b" )) ## create a new variable by reordering the levels of a factor study_data_cleaned <- study_data_cleaned %>% mutate(no_consent = fct_relevel(no_consent_reason, "There is no benefit to me and my family", "I do not have time", "I have had a negative experience with a previous survey.", "Other") ) ## create simplified groupings for each of the vaccine status variables ## simplified, TRUE (regardless of date) otherwise FALSE (or NA if missing) ## select the variables from dictionary for vaccines by card CARDVARS <- study_data_dict %>% filter(type == "vaccine_card") %>% filter(name %in% names(study_data_cleaned)) %>% ## filter to match the column names of your data pull(name) # select vars ## create simplified variables study_data_cleaned <- study_data_cleaned %>% mutate( ## for each of the vaccine variables dplyr::across(all_of(CARDVARS), ## if a row contains "Yes" return TRUE ~str_detect(.x, "Yes"), ## rename with original name with "_simple" on the end .names = "{.col}_simple") ) ## create a list of vaccines with simplified variable names CARDVARSSIMP <- str_glue("{CARDVARS}_simple") ## create a list of vaccines with date on the front of variable names CARDVARSDATE <- str_glue("date_{CARDVARS}") ## years since vaccination variables ------------------------------------------- ## adapt date of vaccination and then create a new time_since variable study_data_cleaned <- study_data_cleaned %>% mutate( across( all_of(CARDVARSDATE), ~as.numeric(date - .x), ## rename with original name with "_simple" on the end .names = "{.col}_time" )) ## create a grouping variable for years since vaccination study_data_cleaned <- study_data_cleaned %>% mutate( across( all_of(str_glue("{CARDVARSDATE}_time")), ~age_categories(.x, lower = 0, upper = 5, by = 1) %>% ## remove what comes after a dash (i.e. repeated year) str_remove("-.*") %>% factor(levels = c("0", "1", "2", "3", "4", "5+")), ## rename with original name with "_grp" on the end .names = "{.col}_grp" ) ) ## create a list of vaccines with time group on the of variable names CARDVARSTIMEGRP <- str_glue("{CARDVARSDATE}_time_grp")
## create variable names with labelled ## define a list of names and labels based on the dictionary var_labels <- setNames( ## add variable labels as a list as.list(study_data_dict$short_name), ## name the list with the current variable names study_data_dict$name) ## add the variable labels to dataset study_data_cleaned <- study_data_cleaned %>% set_variable_labels( ## set the labels with the object defined above .labels = var_labels, ## do not throw an error if some names dont match ## not all names in the dictionary appear in the dataset .strict = FALSE) ## it is possible to update individual variables manually too study_data_cleaned <- study_data_cleaned %>% set_variable_labels( ## variable name = variable label age_years = "Age (years)", age_group = "Age group (years)", age_months = "Age (months)", age_group_mon = "Age group (months)", age_category = "Age category" ) ## add in variables for newly created variables (based on old names) ## e.g. here label the simple vacc variables with the original variable labels ## we use c() to repeat this for the time since vaccination variable too vacc_var_labels <- setNames( ## add variable labels as a list as.list( c(study_data_dict$short_name[study_data_dict$type == "vaccine_card"], study_data_dict$short_name[study_data_dict$type == "vaccine_card"])), ## name the list with the variable names of interest c(CARDVARSSIMP, CARDVARSTIMEGRP) ) ## add the variable labels to dataset study_data_cleaned <- study_data_cleaned %>% set_variable_labels( ## set the labels with the object defined above .labels = vacc_var_labels, ## do not throw an error if some names dont match ## not all names in the dictionary appear in the dataset .strict = FALSE) ## Adding labels for variables with multiple responses are allowed ## In Kobo - multioption variables apear once in the dictionary ## but exports have a character variable with all responses and then a ## binary (1/0) variable for each of the various options ## These are not automatically labelled from the dictionary above ## first create a list of the binary variable names by combining with the ## dictionary defined options for that variable reason_vars <- study_data_dict_long %>% filter(type == "reasons") %>% select(option_name, option_label_english) %>% mutate(option_name = str_glue("reason_not_all_vacc_{option_name}")) ## add in variables for newly created variables (based on old names) reason_var_labels <- setNames( ## add variable labels as a list as.list(reason_vars$option_label_english), ## name the list with the variable names of interest reason_vars$option_name ) ## add the variable labels to dataset study_data_cleaned <- study_data_cleaned %>% set_variable_labels( ## set the labels with the object defined above .labels = reason_var_labels, ## do not throw an error if some names dont match ## not all names in the dictionary appear in the dataset .strict = FALSE)
## Drop unused rows ----------------------------------------------------------- ## store the cases that you drop so you can describe them (e.g. non-consenting) dropped <- study_data_cleaned %>% filter(!consent| village_name == "other") ## use the dropped cases to remove the unused rows from the survey data set study_data_cleaned <- anti_join(study_data_cleaned, dropped, by = names(dropped)) ## Drop columns ---------------------------------------------------------------- ## OPTIONAL: This step shows you how you can remove certain variables. ## study_data_cleaned <- select(study_data_cleaned, -c("age_years", "sex")) ## OPTIONAL: if you want to inspect certain variables, you can select these by ## name or column number. This example creates a reduced dataset for the first ## three columns, age_years, and sex. # study_data_reduced <- select(study_data_cleaned, c(1:3, "age_years", "sex")
## option 1: only keep the first occurrence of duplicate case study_data_cleaned <- study_data_cleaned %>% ## find duplicates based on case number, sex and age group ## only keep the first occurrence distinct(uid, sex, age_group, .keep_all = TRUE) # ## option 2: create flagging variables for duplicates (then use to browse) # # study_data_cleaned <- study_data_cleaned %>% # ## choose which variables to use for finding unique rows # group_by(uid, sex, age_group) %>% # mutate( # ## get the number of times duplicate occurs # num_dupes = n(), # duped = if_else(num_dupes > 1 , TRUE, FALSE) # ) # # ## browse duplicates based on flagging variables # study_data_cleaned %>% # ## only keep rows that are duplicated # filter(duped) %>% # ## arrange by variables of interest # arrange(uid, sex, age_group) %>% # View() # # ## filter duplicates to only keep the row with the earlier entry # study_data_cleaned %>% # ## choose which variables to use for finding unique rows # group_by(uid, sex, age_group) %>% # ## sort to have the earliest date by person first # arrange(as.Date(submission_time)) %>% # ## only keep the earliest row # slice(1)
## stratified ------------------------------------------------------------------ ## create a variable called "surv_weight_strata" ## contains weights for each individual - by age group, sex and health district study_data_cleaned <- add_weights_strata(x = study_data_cleaned, p = population_data_age, surv_weight = "surv_weight_strata", surv_weight_ID = "surv_weight_ID_strata", age_group, sex, health_district) ## cluster --------------------------------------------------------------------- ## get the number of people of individuals interviewed per household ## adds a variable with counts of the household (parent) index variable study_data_cleaned <- study_data_cleaned %>% add_count(index, name = "interviewed") ## create cluster weights study_data_cleaned <- add_weights_cluster(x = study_data_cleaned, cl = cluster_counts, eligible = number_children, interviewed = interviewed, cluster_x = village_name, cluster_cl = cluster, household_x = index, household_cl = households, surv_weight = "surv_weight_cluster", surv_weight_ID = "surv_weight_ID_cluster", ignore_cluster = FALSE, ignore_household = FALSE) ## stratified and cluster ------------------------------------------------------ ## create a survey weight for cluster and strata study_data_cleaned <- study_data_cleaned %>% mutate(surv_weight_cluster_strata = surv_weight_strata * surv_weight_cluster)
## simple random --------------------------------------------------------------- survey_design_simple <- study_data_cleaned %>% as_survey_design(ids = 1, # 1 for no cluster ids weights = NULL, # No weight added strata = NULL # sampling was simple (no strata) ) ## stratified ------------------------------------------------------------------ survey_design_strata <- study_data_cleaned %>% as_survey_design(ids = 1, # 1 for no cluster ids weights = surv_weight_strata, # weight variable created above strata = health_district # sampling was stratified by district ) ## cluster --------------------------------------------------------------------- survey_design_cluster <- study_data_cleaned %>% as_survey_design(ids = village_name, # cluster ids weights = surv_weight_cluster, # weight variable created above strata = NULL # sampling was simple (no strata) ) ## stratified cluster ---------------------------------------------------------- survey_design <- study_data_cleaned %>% as_survey_design(ids = village_name, # cluster ids weights = surv_weight_cluster_strata, # weight variable created above strata = health_district # sampling was stratified by district )
# rio::export(study_data_cleaned, here::here("data", str_glue("study_data_cleaned_{Sys.Date()}.xlsx")))
Results
Survey inclusion
## get counts of number of clusters num_clus <- study_data_cleaned %>% ## trim data to unique clusters distinct(cluster_number) %>% ## get number of rows (count how many unique) nrow() ## get counts of number households num_hh <- study_data_cleaned %>% ## get unique houses by cluster distinct(cluster_number, household_number) %>% ## get number of rounds (count how many unique) nrow()
We included r num_hh households across r num_clus clusters in this survey analysis.
Among the r nrow(dropped) individuals without consent to participate in the survey,
the reasons for refusal are described below.
## using the dataset with dropped individuals dropped %>% ## only keep reasons for no consent select(no_consent_reason) %>% ## make NAs a factor level called missing ## to make the proportion of the total, including the missings mutate(no_consent_reason = fct_explicit_na(no_consent_reason, na_level = "Missing")) %>% ## get counts and proportions tbl_summary() %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
## get counts of the number of households per cluster clustersize <- study_data_cleaned %>% ## trim data to only unique households within each cluster distinct(cluster_number, household_number) %>% ## count the number of households within each cluster count(cluster_number) %>% pull(n) ## get the median number of households per cluster clustermed <- median(clustersize) ## get the min and max number of households per cluster ## paste these together seperated by a dash clusterrange <- str_c(range(clustersize), collapse = "--") ## get counts of children per household ## do this by cluster as household IDs are only unique within clusters hhsize <- study_data_cleaned %>% count(cluster_number, household_number) %>% pull(n) ## get median number of children per household hhmed <- median(hhsize) ## get the min and max number of children per household ## paste these together seperated by a dash hhrange <- str_c(range(hhsize), collapse = "--") # get standard deviation hhsd <- round(sd(hhsize), digits = 1)
The median number of households per cluster was
r clustermed, with a range of r clusterrange. The median number of children
per household was r hhmed (range: r hhrange, standard deviation: r hhsd).
Demographic information
In total we included r nrow(study_data_cleaned) in the survey analysis.
The age break down and a comparison with the source population is shown below.
## counts and props of the study population ag <- tabyl(study_data_cleaned, age_group, show_na = FALSE) %>% mutate(n_total = sum(n), age_group = fct_inorder(age_group)) ## counts and props of the source population propcount <- population_data_age %>% group_by(age_group) %>% tally(population) %>% mutate(percent = n / sum(n)) ## bind together the columns of two tables, group by age, and perform a ## binomial test to see if n/total is significantly different from population ## proportion. ## suffix here adds to text to the end of columns in each of the two datasets left_join(ag, propcount, by = "age_group", suffix = c("", "_pop")) %>% group_by(age_group) %>% ## broom::tidy(binom.test()) makes a data frame out of the binomial test and ## will add the variables p.value, parameter, conf.low, conf.high, method, and ## alternative. We will only use p.value here. You can include other ## columns if you want to report confidence intervals mutate(binom = list(broom::tidy(binom.test(n, n_total, percent_pop)))) %>% unnest(cols = c(binom)) %>% # important for expanding the binom.test data frame mutate( across(.cols = contains("percent"), .fns = ~.x * 100)) %>% ## Adjusting the p-values to correct for false positives ## (because testing multiple age groups). This will only make ## a difference if you have many age categories mutate(p.value = p.adjust(p.value, method = "holm")) %>% ## Only show p-values over 0.001 (those under report as <0.001) mutate(p.value = ifelse(p.value < 0.001, "<0.001", as.character(round(p.value, 3)))) %>% ## rename the columns appropriately select( "Age group" = age_group, "Study population (n)" = n, "Study population (%)" = percent, "Source population (n)" = n_pop, "Source population (%)" = percent_pop, "P-value" = p.value ) %>% # produce styled output table with auto-adjusted column widths with {flextable} qflextable() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit") %>% ## set to only show 1 decimal place colformat_double(digits = 1)
## compute the median age medage <- median(study_data_cleaned$age_years) ## paste the lower and upper quartile together iqr <- str_c( # basically copy paste togehter the following ## calculate the 25% and 75% of distribution, with missings removed quantile( study_data_cleaned$age_years, c(0.25, 0.75), na.rm = TRUE), ## between lower and upper place an en-dash collapse = "--") ## compute overall sex ratio sex_ratio <- study_data_cleaned %>% count(sex) %>% pivot_wider(names_from = sex, values_from = n) %>% mutate(ratio = round(Male/Female, digits = 3)) %>% pull(ratio) ## compute sex ratios by age group sex_ratio_age <- study_data_cleaned %>% count(age_group, sex) %>% pivot_wider(names_from = sex, values_from = n) %>% mutate(ratio = round(Male/Female, digits = 3)) %>% select(age_group, ratio) ## sort table by ascending ratio then select the lowest (first) min_sex_ratio_age <- arrange(sex_ratio_age, ratio) %>% slice(1)
Among the r nrow(study_data_cleaned) surveyed individuals, there were
r fmt_count(study_data_cleaned, sex == "Female") females and
r fmt_count(study_data_cleaned, sex == "Male") males (unweighted). The male to
female ratio was r sex_ratio in the surveyed population. The lowest male to
female ratio was r min_sex_ratio_age$ratio
in the r min_sex_ratio_age$age_group year age group.
The median age of surveyed individuals was r medage years (Q1-Q3 of r iqr
years). Children under five years of age made up
r fmt_count(study_data_cleaned, age_years < 5)of the surveyed individuals.
The highest number of surveyed individuals (unweighted) were in the
r table(study_data_cleaned$age_group) %>% which.max() %>% names()
year age group.
Unweighted age distribution of population by year age group and gender.
# get cross tabulated counts and proportions study_data_cleaned %>% tbl_cross( row = age_group, col = sex, percent = "cell") %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
Unweighted age distribution of population by age category (month and year) and gender.
# get cross tabulated counts and proportions study_data_cleaned %>% tbl_cross( row = age_category, col = sex, percent = "cell") %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
There were r fmt_count(dropped, is.na(sex)) cases missing information on sex and
r fmt_count(dropped, is.na(age_group)) missing age group.
Unweighted age and gender distribution of household population covered by the survey. <!-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /// age_pyramid \\
This chunk creates an unweighted (using study_data_cleaned) age/sex pyramid of your cases. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -->
age_pyramid(study_data_cleaned, age_group, split_by = sex, proportional = TRUE) + labs(y = "Proportion", x = "Age group (years)") + # change axis labels theme(legend.position = "bottom", # move legend to bottom legend.title = element_blank(), # remove title text = element_text(size = 18) # change text size )
Unweighted age and gender distribution, stratified by health district, of household population covered by the survey. <!-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /// age_pyramid_strata \\
This chunk creates an unweighted (using study_data_cleaned) age/sex pyramid of your cases, stratified by health district.
If you have a stratified survey design this may be useful for visualising if you have an excess of representation in either sex or gender in any of your survey strata. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -->
age_pyramid(study_data_cleaned, age_group, split_by = sex, stack_by = health_district, proportional = TRUE, pal = c("red", "blue")) + labs(y = "Proportion", x = "Age group (years)") + # change axis labels theme(legend.position = "bottom", # move legend to bottom legend.title = element_blank(), # remove title text = element_text(size = 18) # change text size )
Weighted age and gender distribution of household population covered by the survey.
age_pyramid(survey_design, age_group, split_by = sex, proportional = TRUE) + labs(y = "Proportion", x = "Age group (years)") + # change axis labels theme(legend.position = "bottom", # move legend to bottom legend.title = element_blank(), # remove title text = element_text(size = 18) # change text size )
Vaccination coverage
Weighted vaccination coverage; accepting equal validity from self-reported and card-reported vaccination status, and detailing availability of vaccination cards and health facility records
survey_design %>% ## only keep variables of interest select(any_vaccine, vaccine_card, hf_records) %>% ## only display one line for each variable (with counts for "Yes") tbl_svysummary(value = list(everything() ~ "Yes"), statistic = everything() ~ "{n} ({p}%) [{deff}]") %>% ## add the weighted total add_n() %>% ## add in confidence intervals add_ci() %>% ## modify the column headers modify_header( list( n ~ "**Weighted total (N)**", stat_0 ~ "**Weighted Count (n)**" ) ) %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
Vaccination history verbal
Weighted overall vaccination completeness; based on surveyor assessments of care-giver verbal accounts and on care-giver's own assessment
survey_design %>% ## only keep variables of interest select(all_vaccines_survey_woc, all_vaccines_caretaker_woc) %>% ## only display one line for each variable (with counts for "Yes") tbl_svysummary(value = list(everything() ~ "Yes")) %>% ## add the weighted total add_n() %>% ## add in confidence intervals add_ci() %>% ## modify the column headers modify_header( list( n ~ "**Weighted total (N)**", stat_0 ~ "**Weighted Count (n)**" ) ) %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
Weighted vaccination coverage by disease based on verbal history
survey_design %>% ## only keep variables of interest select(injection_upper_arm, scar_present, poliodrop_woc, poliodrop_hf_woc, poliodrop_camp_woc, polioinjection_woc, num_polioinjection_woc, quad_penta_woc, num_quad_penta_woc, pcv_woc,num_pcv_woc, measles_woc, num_measles_woc, measles_hf_woc, measles_camp_woc, yellow_fever_woc, num_yellow_fever_woc, yellow_fever_hf_woc, yellow_fever_camp_woc, rotavirus_woc, num_rotavirus_woc) %>% ## only display one line for each variable (with counts for "Yes") ## also specify for the scar_present variable to show the relevant level tbl_svysummary(value = list(where(is.factor) ~ "Yes", scar_present ~ "Scar present")) %>% ## add the weighted total add_n() %>% ## add in confidence intervals add_ci() %>% ## modify the column headers modify_header( list( n ~ "**Weighted total (N)**", stat_0 ~ "**Weighted Count (n)**" ) ) %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
Vaccination history with a card
Weighted overall vaccination completeness; based on surveyor assessments of vaccination cards
survey_design %>% ## only keep variables of interest select(all_vaccines_survey_card) %>% ## only display one line for each variable (with counts for "Yes") tbl_svysummary(value = list(everything() ~ "Yes")) %>% ## add the weighted total add_n() %>% ## add in confidence intervals add_ci() %>% ## modify the column headers modify_header( list( n ~ "**Weighted total (N)**", stat_0 ~ "**Weighted Count (n)**" ) ) %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
Weighted vaccination coverage by vaccine and dose based on vaccination cards
survey_design %>% ## only keep variables of interest select(all_of(CARDVARSSIMP)) %>% ## automatically only displays one line for each variable (with counts for TRUE) tbl_svysummary() %>% ## add the weighted total add_n() %>% ## add in confidence intervals add_ci() %>% ## modify the column headers modify_header( list( n ~ "**Weighted total (N)**", stat_0 ~ "**Weighted Count (n)**" ) ) %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
Weighted counts of time (in years) since vaccination by various vaccines based on vaccination cards
survey_design %>% ## only keep variables of interest select(all_of(CARDVARSTIMEGRP)) %>% ## automatically only displays one line for each variable (with counts for TRUE) tbl_svysummary() %>% ## add the weighted total add_n() %>% ## add in confidence intervals add_ci() %>% ## modify the column headers modify_header( list( n ~ "**Weighted total (N)**", stat_0 ~ "**Weighted Count (n)**" ) ) %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
Reasons for not vaccinating
Weighted counts and proportions for reasons not vaccinated, individually, among those not vaccinated based on verbal or card assessment <!-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ /// no_vacc_reason \\
The below chunk creates a weighted table of counts proportions for reasons not vaccinated.
Note that low counts or short observation times may lead to a confidence interval that crosses zero (i.e. negative) for ratios. These should be interpreted as if no deaths or recoded to zero (impossible to have negative deaths). ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ -->
survey_design %>% ## filter for those not vaccinated filter( all_vaccines_survey_card != "Yes") %>% select(reason_vars$option_name) %>% tbl_svysummary(value = list(everything() ~ 1)) %>% ## add in confidence intervals add_ci() %>% ## modify the column headers modify_header( list( stat_0 ~ "Weighted Count (n) \n (N={round(N, digits = 0)})" ) ) %>% ## make variable names bold bold_labels() %>% # change to flextable format as_flex_table() %>% # make header text bold (using {flextable}) bold(part = "header") %>% # make your table fit to the maximum width of the word document set_table_properties(layout = "autofit")
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.