add measure of recording completeness across practices

analysis/practice_completeness.R

@@ -0,0 +1,34 @@
+## Completeness of ethnicity recording by practice
+
+library(arrow)
+library("tidyverse")
+
+
+data <- read_feather(here::here("output", "extract_5", "input_5.feather"))
+
+data_count <- data %>%
+  filter(registered == T) %>%
+  summarise(practice,
+    n = 1,
+    recorded_ethnicity = case_when(
+      !is.na(ethnicity_new_5) ~ 1,
+      TRUE ~ 0
+    )
+  ) %>%
+  group_by(practice) %>%
+  summarise(
+    n = sum(n),
+    n_recorded_ethnicity = sum(recorded_ethnicity),
+    completeness = n_recorded_ethnicity / n * 100
+  ) %>%
+  filter(n > 1000) %>%
+  ungroup() %>%
+  summarise(
+    p05 = quantile(completeness, probs = seq(0, 1, 0.05), na.rm = TRUE)["5%"] %>% round(digits = 1),
+    Q1 = quantile(completeness, na.rm = TRUE)["25%"] %>% round(digits = 1),
+    median = median(completeness, na.rm = TRUE) %>% round(digits = 1),
+    Q3 = quantile(completeness, na.rm = TRUE)["75%"] %>% round(digits = 1),
+    p95 = quantile(completeness, probs = seq(0, 1, 0.05), na.rm = TRUE)["95%"] %>% round(digits = 1)
+  )
+
+write_csv(data_count, here::here("output", "extract_5", "practice_completeness.csv"))

project.yaml

@@ -121,6 +121,14 @@ actions:
       moderately_sensitive:
         html: output/report/report.html
 
+#### reviewer comments
+  practice_completeness:
+    run: r:latest analysis/practice_completeness.R
+    needs: [generate_study_population_5]
+    outputs:
+      moderately_sensitive:
+        csvs: output/extract_5/practice_completeness.csv
+
 #### move released files to output/released
 ###### run locally for manuscript
 ##### local notebook