Skip to content

Commit

Permalink
Merge pull request #86 from opensafely/notebook-report
Browse files Browse the repository at this point in the history 
fix "Unknown" population
  • Loading branch information
@andrewscolm
andrewscolm authored Feb 28, 2024
2 parents b14de20 + ce89e64 commit 02294b0
Showing 1 changed file with 63 additions and 42 deletions.
105 changes: 63 additions & 42 deletions notebooks_r/report.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -282,7 +282,7 @@ SUS_gt <- combined_prop_table %>%
cols_label(!!!table_cols) %>%
tab_style(
style = list(
cell_fill(color = "gray96")
# cell_fill(color = "gray96")
),
locations = cells_body()
) %>%
Expand All @@ -300,17 +300,17 @@ SUS_gt <- combined_prop_table %>%
table.border.top.color = "transparent",
heading.align = "left",
data_row.padding = px(0)
) %>%
tab_header(
title = md("Table 1: Count of patients with a recorded ethnicity in OpenSAFELY-TPP (proportion of registered TPP population) by clinical and demographic subgroups. All counts are rounded to the nearest 5."),
# ) %>%
# tab_header(
# title = md("Table 1: Count of patients with a recorded ethnicity in OpenSAFELY-TPP (proportion of registered TPP population) by clinical and demographic subgroups. All counts are rounded to the nearest 5."),
)
SUS_gt
```

<br>

```{r, fig.width=25, fig.height=18}
```{r, fig.width=18, fig.height=19}
ctv3_prop_reg <-
read_csv(here::here("output", "sus", "simplified_output", "5_group", "tables", "simple_patient_counts_5_group_ctv3_sus_registered.csv"), col_types = (cols())) %>%
select("group", "subgroup", starts_with("ethnicity_5"), starts_with("any"), "population") %>%
Expand Down Expand Up @@ -521,7 +521,7 @@ SUS5 <- combined_prop_5_table %>%
cols_label(!!!my_cols) %>%
tab_style(
style = list(
cell_fill(color = "gray96")
# cell_fill(color = "gray96")
),
locations = cells_body()
) %>%
Expand All @@ -539,9 +539,9 @@ SUS5 <- combined_prop_5_table %>%
column_labels.border.top.color = "transparent",
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Table 2: Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and clinical and demographic subgroups. All counts are rounded to the nearest 5. "),
# ) %>%
# tab_header(
# title = md("Table 2: Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and clinical and demographic subgroups. All counts are rounded to the nearest 5. "),
) %>%
tab_options(., container.width = 1200) %>%
tab_options(
Expand All @@ -555,7 +555,7 @@ SUS5

<br>

```{r, fig.width=60, fig.height=40}
```{r, fig.width=30, fig.height=27}
new_prop_reg_cat <-
read_csv(here::here("output", "sus", "simplified_output", "5_group", "tables", "simple_patient_counts_categories_5_group_new_sus_registered.csv"), col_types = (cols())) %>%
rename_with(~ sub("ethnicity_", "", .), contains("ethnicity_")) %>%
Expand Down Expand Up @@ -800,7 +800,7 @@ SUS16 <- combined_prop_16_table %>%
cols_label(!!!my_cols) %>%
tab_style(
style = list(
cell_fill(color = "gray96")
# cell_fill(color = "gray96")
),
locations = cells_body()
) %>%
Expand All @@ -818,9 +818,9 @@ SUS16 <- combined_prop_16_table %>%
column_labels.border.top.color = "transparent",
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and clinical and demographic subgroups. All counts are rounded to the nearest 5."),
# ) %>%
# tab_header(
# title = md("Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and clinical and demographic subgroups. All counts are rounded to the nearest 5."),
) %>%
tab_options(., container.width = 3000) %>%
tab_options(
Expand Down Expand Up @@ -871,9 +871,9 @@ for (codelist in c("new", "ctv3")) {
column_labels.border.top.color = "transparent",
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Table 4: Count of patients with at least one recording of each ethnicity (proportion of latest ethnicity)."),
# ) %>%
# tab_header(
# title = md("Table 4: Count of patients with at least one recording of each ethnicity (proportion of latest ethnicity)."),
) %>%
tab_options(
data_row.padding = px(0)
Expand Down Expand Up @@ -932,9 +932,9 @@ for (codelist in c("new", "ctv3")) {
column_labels.border.top.color = "transparent",
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Table 4: Count of patients with at least one recording of each ethnicity (proportion of latest ethnicity)."),
# ) %>%
# tab_header(
# title = md("Table 4: Count of patients with at least one recording of each ethnicity (proportion of latest ethnicity)."),
) %>%
tab_options(
data_row.padding = px(0)
Expand Down Expand Up @@ -1045,6 +1045,7 @@ latestcommon_new_16 %>%
```{r}
for (codelist in c("new", "ctv3")) {
ifelse(codelist == "new", codelist_name <- "SNOMED:2022", codelist_name <- "CTV3:2020")
ifelse(codelist == "new", codelist_path <- glue("ethnicity_new_5"), codelist_path <- glue("ethnicity_5"))
df_sus_new_cross <- read_csv(here::here("output", "sus", "simplified_output", "5_group", "tables", glue("simple_{codelist}_sus_crosstab_long_registered.csv"))) %>%
rename_with(~"ethnicity", contains("ethnicity_") & !contains("sus"))
Expand All @@ -1062,6 +1063,13 @@ for (codelist in c("new", "ctv3")) {
White_anydiff = White_any - White,
Other_anydiff = Other_any - Other,
)
population <- read_csv(here::here("output", "sus", "simplified_output", "5_group", "tables", glue("simple_patient_counts_5_group_{codelist}_sus_registered.csv")), col_types = (cols())) %>%
filter(group == "all") %>%
summarise(
ethnicity = "Unknown",
population = population - !!as.name(glue("{codelist_path}_filled"))
)
ethnicity_cat_pivot <- ethnicity_cat %>%
pivot_longer(
Expand All @@ -1074,7 +1082,8 @@ for (codelist in c("new", "ctv3")) {
summarise(
ethnicity,
population = n
)
) %>%
bind_rows(population)
df_sus_new_cross_perc <- df_sus_new_cross %>%
Expand Down Expand Up @@ -1109,7 +1118,7 @@ for (codelist in c("new", "ctv3")) {
cols_label(!!!my_cols) %>%
tab_style(
style = list(
cell_fill(color = "gray96")
# cell_fill(color = "gray96")
),
locations = cells_body()
) %>%
Expand All @@ -1127,9 +1136,9 @@ for (codelist in c("new", "ctv3")) {
column_labels.border.top.color = "transparent",
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Table 5: Count of patients with a recorded ethnicity in Secondary Care by ethnicity group (proportion of Primary Care population). All counts are rounded to the nearest 5. "),
# ) %>%
# tab_header(
# title = md("Table 5: Count of patients with a recorded ethnicity in Secondary Care by ethnicity group (proportion of Primary Care population). All counts are rounded to the nearest 5. "),
) %>%
tab_options(
data_row.padding = px(0)
Expand Down Expand Up @@ -1202,7 +1211,7 @@ df_sus_new_cross_table_ctv3

<br>

```{r, fig.width=20, fig.height=15}
```{r, fig.width=15, fig.height=10}
alluvial_new
```

Expand All @@ -1213,7 +1222,7 @@ df_sus_new_cross_table_new
```
<br>

```{r, fig.width=20, fig.height=15}
```{r, fig.width=15, fig.height=10}
alluvial_new
```

Expand All @@ -1223,9 +1232,18 @@ alluvial_new
```{r}
for (codelist in c("new", "ctv3")) {
ifelse(codelist == "new", codelist_name <- "SNOMED:2022", codelist_name <- "CTV3:2020")
ifelse(codelist == "new", codelist_path <- glue("ethnicity_new_16"), codelist_path <- glue("ethnicity_16"))
df_sus_new_cross <- read_csv(here::here("output", "sus", "simplified_output", "16_group", "tables", glue("simple_{codelist}_sus_crosstab_long_registered.csv"))) %>%
rename_with(~"ethnicity", contains("ethnicity_") & !contains("sus"))
population <- read_csv(here::here("output", "sus", "simplified_output", "16_group", "tables", glue("simple_patient_counts_16_group_{codelist}_sus_registered.csv")), col_types = (cols())) %>%
filter(group == "all") %>%
summarise(
ethnicity = "Unknown",
population = population - !!as.name(glue("{codelist_path}_filled"))
)
cols <- c("White_British", "White_Irish", "Other_White", "White_and_Black_Caribbean", "White_and_Black_African", "White_and_Asian", "Other_Mixed", "Indian", "Pakistani", "Bangladeshi", "Other_Asian", "Caribbean", "African", "Other_Black", "Chinese", "Any_other_ethnic_group")
ethnicity_cat <-
Expand Down Expand Up @@ -1254,6 +1272,8 @@ for (codelist in c("new", "ctv3")) {
Any_other_ethnic_group_anydiff = Any_other_ethnic_group_any - Any_other_ethnic_group
)
ethnicity_cat_pivot <- ethnicity_cat %>%
pivot_longer(
cols = c(contains("_")),
Expand All @@ -1265,7 +1285,8 @@ for (codelist in c("new", "ctv3")) {
summarise(
ethnicity,
population = n
)
) %>%
bind_rows(population)
df_sus_new_cross_perc <- df_sus_new_cross %>%
Expand Down Expand Up @@ -1301,7 +1322,7 @@ for (codelist in c("new", "ctv3")) {
cols_label(!!!my_cols) %>%
tab_style(
style = list(
cell_fill(color = "gray96")
# cell_fill(color = "gray96")
),
locations = cells_body()
) %>%
Expand All @@ -1319,9 +1340,9 @@ for (codelist in c("new", "ctv3")) {
column_labels.border.top.color = "transparent",
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Table 5: Count of patients with a recorded ethnicity in Secondary Care by ethnicity group (proportion of Primary Care population). All counts are rounded to the nearest 5. "),
# ) %>%
# tab_header(
# title = md("Table 5: Count of patients with a recorded ethnicity in Secondary Care by ethnicity group (proportion of Primary Care population). All counts are rounded to the nearest 5. "),
) %>%
tab_options(
data_row.padding = px(0)
Expand Down Expand Up @@ -1380,7 +1401,7 @@ ONS_tab_2001 %>%
cols_label(!!!my_cols_ons) %>%
tab_style(
style = list(
cell_fill(color = "gray96")
# cell_fill(color = "gray96")
),
locations = cells_body()
) %>%
Expand All @@ -1398,9 +1419,9 @@ ONS_tab_2001 %>%
column_labels.border.top.color = "transparent",
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Table 7: Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and 2021 ONS Census counts [amended to 2001 grouping] (proportion of 2021 ONS Census population). All counts are rounded to the nearest 5. "),
# ) %>%
# tab_header(
# title = md("Table 7: Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and 2021 ONS Census counts [amended to 2001 grouping] (proportion of 2021 ONS Census population). All counts are rounded to the nearest 5. "),
) %>%
tab_options(
data_row.padding = px(0)
Expand Down Expand Up @@ -1467,7 +1488,7 @@ ons_ethnicity_plot_eng_na

### Region

```{r, fig.width=14, fig.height=16}
```{r, fig.width=18, fig.height=16}
## 5 group ethnicity plot NA removed for Regions
ons_ethnicity_plot_na <- ons_na_removed %>%
filter(region != "England", group == "5") %>%
Expand Down Expand Up @@ -1542,7 +1563,7 @@ ONS_tab_2001 %>%
cols_label(!!!my_cols_ons) %>%
tab_style(
style = list(
cell_fill(color = "gray96")
# cell_fill(color = "gray96")
),
locations = cells_body()
) %>%
Expand All @@ -1561,9 +1582,9 @@ ONS_tab_2001 %>%
table.border.top.color = "transparent",
heading.align = "left"
) %>%
tab_header(
title = md("Table 7: Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and 2021 ONS Census counts [amended to 2001 grouping] (proportion of 2021 ONS Census population). All counts are rounded to the nearest 5. "),
) %>%
# tab_header(
# title = md("Table 7: Count of patients with a recorded ethnicity in OpenSAFELY TPP by ethnicity group (proportion of registered TPP population) and 2021 ONS Census counts [amended to 2001 grouping] (proportion of 2021 ONS Census population). All counts are rounded to the nearest 5. "),
# ) %>%
tab_options(
data_row.padding = px(0)
) %>%
Expand Down Expand Up @@ -1630,7 +1651,7 @@ ons_ethnicity_plot_eng_na

### Region

```{r, fig.width=14, fig.height=34}
```{r, fig.width=19, fig.height=34}
## 16 group ethnicity plot NA removed for Regions
ons_ethnicity_plot_na <-
ons_na_removed %>%
Expand Down

0 comments on commit 02294b0

Please sign in to comment.