Skip to content

Commit

Permalink
report fixes
Browse files Browse the repository at this point in the history
  • Loading branch information
@andrewscolm
andrewscolm committed Feb 28, 2024
1 parent faf99b0 commit 90de294
Showing 1 changed file with 31 additions and 29 deletions.
60 changes: 31 additions & 29 deletions notebooks_r/report.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -169,7 +169,7 @@ All patient counts are rounded to the nearest 5. Percentages may not add to exac
codelist_counts <- read_csv(here::here("output", "simplified_output", "5_group", "tables", "simple_patient_counts_5_group_new_ctv3_registered.csv"))
```

Around `r codelist_counts %>% filter(group=="all")%>%pull(all_filled)/1000000` million patients who have been registered in OpenSAFELY-TPP are identified in all three codelists. CTV3:2020 is the most well-populated with `r codelist_counts %>% filter(group=="all")%>%pull(ethnicity_5_filled)/1000000` million patients having at least one CTV3:2020 recording of ethnicity.
Around `r round(codelist_counts %>% filter(group=="all")%>%pull(all_filled)/1000000,0)` million patients who have been registered in OpenSAFELY-TPP are identified in both codelists. CTV3:2020 is the most well-populated with `r round(codelist_counts %>% filter(group=="all")%>%pull(ethnicity_5_filled)/1000000,0)` million patients having at least one CTV3:2020 recording of ethnicity.

```{r}
# # patient counts
Expand All @@ -196,7 +196,7 @@ new_sus_prop_table <- read_csv(here::here("output", "report_tables", "report_pat
F = "Female",
M = "Male"
),
population = as.character(scales::comma(round(as.numeric(population), 0))),
population = as.character(comma(round(as.numeric(population), 0))),
left_paren = "(",
right_paren = ")"
) %>%
Expand Down Expand Up @@ -228,7 +228,7 @@ ctv3_sus_prop_table <- read_csv(here::here("output", "report_tables", "report_pa
F = "Female",
M = "Male"
),
population = as.character(scales::comma(round(as.numeric(population), 0))),
population = as.character(comma(round(as.numeric(population), 0))),
left_paren = "(",
right_paren = ")"
) %>%
Expand Down Expand Up @@ -259,7 +259,7 @@ new_ctv3_prop_table <- read_csv(here::here("output", "report_tables", "report_pa
F = "Female",
M = "Male"
),
population = as.character(scales::comma(round(as.numeric(population), 0))),
population = as.character(comma(round(as.numeric(population), 0))),
) %>%
filter(`5 CTV3:2020` != "- (-)") %>%
select(group, subgroup, `all filled`, population)
Expand Down Expand Up @@ -554,7 +554,7 @@ SUS5

<br>

```{r, fig.width=25, fig.height=30}
```{r, fig.width=60, fig.height=40}
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 @@ -971,7 +971,7 @@ for (codelist in c("new", "ctv3")) {
replace(., col(.) == row(.) + 1, NA) %>%
mutate(
any_discordant = rowSums(across(where(is.numeric)), na.rm = T) - sum,
any_discordant_perc = round(any_discordant / sum * 100, 0),
any_discordant_perc = comma(round(any_discordant / sum * 100, 1)),
discordantcombined = glue("{any_discordant} ({any_discordant_perc})")
)
Expand Down Expand Up @@ -1012,30 +1012,30 @@ for (codelist in c("new", "ctv3")) {

### 5 Group

Overall `r round((1 - (sum(discordant_new_5$any_discordant)/sum(discordant_new_5$sum))) *100,0)`% of the latest 5 group ethnicity matched the most frequent 5 group ethnicity for all codelists. `r round((1 - (discordant_new_5 %>% filter(ethnicity_new_5=="White") %>% select(any_discordant) / discordant_new_5 %>% filter(ethnicity_new_5=="White") %>% select(sum))) *100,0)`% of those with the most recent ethnicity classified as White also had the most frequent ethnicity White for both codelists. Mixed was the least concordant for all both codelists with `r round(((discordant_new_5 %>% filter(ethnicity_new_5=="Mixed") %>% select(any_discordant) / discordant_new_5 %>% filter(ethnicity_new_5=="Mixed") %>% select(sum))) *100,0)`% of those with the most recent ethnicity Mixed also had the most frequent ethnicity Mixed. Of those with latest ethnicity Black `r round(((discordant_new_5 %>% filter(ethnicity_new_5=="Black") %>% select(ethnicity_new_5_white) / discordant_new_5 %>% filter(ethnicity_new_5=="Black") %>% select(sum))) *100,0)`% also had the most frequent ethnicity White.
Overall `r round((1 - (sum(discordant_new_5$any_discordant)/sum(discordant_new_5$sum))) *100,0)`% of the latest 5 group ethnicity matched the most frequent 5 group ethnicity for both codelists. `r round((1 - (discordant_new_5 %>% filter(ethnicity_new_5=="White") %>% select(any_discordant) / discordant_new_5 %>% filter(ethnicity_new_5=="White") %>% select(sum))) *100,0)`% of those with the most recent ethnicity classified as White also had the most frequent ethnicity White for both codelists. Mixed was the least concordant for both codelists with `r round(((discordant_new_5 %>% filter(ethnicity_new_5=="Mixed") %>% select(any_discordant) / discordant_new_5 %>% filter(ethnicity_new_5=="Mixed") %>% select(sum))) *100,0)`% of those with the most recent ethnicity Mixed also had the most frequent ethnicity Mixed. Of those with latest ethnicity Black `r round(((discordant_new_5 %>% filter(ethnicity_new_5=="Black") %>% select(ethnicity_new_5_white) / discordant_new_5 %>% filter(ethnicity_new_5=="Black") %>% select(sum))) *100,0)`% also had the most frequent ethnicity White.




```{r}
latestcommon_new_5
latestcommon_ctv3_5
```
<br>
```{r}
latestcommon_ctv3_5
latestcommon_new_5
```

### 16 group

Expanding to the 16 group the percentage of latest ethnicity that match the most frequent ethnicity falls to `r round((1 - (sum(discordant_new_16$any_discordant)/sum(discordant_new_16$sum))) *100,0)`% for both SNOMED:2022 and CTV3:2020. White British was the most concordant for both SNOMED:2022 and CTV3:2020 with `r round((1 - (discordant_new_16 %>% filter(ethnicity_new_16=="White_British") %>% select(any_discordant) / discordant_new_16 %>% filter(ethnicity_new_16=="White_British") %>% select(sum))) *100,0)`% and `r round((1 - (discordant_ctv3_16 %>% filter(ethnicity_16=="White_British") %>% select(any_discordant) / discordant_ctv3_16 %>% filter(ethnicity_16=="White_British") %>% select(sum))) *100,0)`%, respectively, of those with the most recent ethnicity classified as White British also had the most frequent ethnicity White British. For both SNOMED:2022 and CTV3:2020 Other Black was the least concordant, `r round((1 - (discordant_new_16 %>% filter(ethnicity_new_16=="Other_Black") %>% select(any_discordant) / discordant_new_16 %>% filter(ethnicity_new_16=="Other_Black") %>% select(sum))) *100,0)`% of those with the most recent ethnicity Other Black also had the most frequent ethnicity Other Black.

```{r}
latestcommon_new_16 %>%
latestcommon_ctv3_16 %>%
tab_options(., container.width = 1600)
```
<br>
```{r}
latestcommon_ctv3_16 %>%
latestcommon_new_16 %>%
tab_options(., container.width = 1600)
```

Expand Down Expand Up @@ -1091,8 +1091,8 @@ for (codelist in c("new", "ctv3")) {
),
left_paren = " (",
right_paren = ")",
N = scales::comma(as.numeric(`0`)),
population = scales::comma(as.numeric(population))
N = comma(as.numeric(`0`)),
population = comma(as.numeric(population))
) %>%
arrange(ethnicity, ethnicity_sus_5) %>%
unite("labl", N, left_paren, percentage, right_paren, sep = "", remove = F) %>%
Expand Down Expand Up @@ -1198,24 +1198,24 @@ for (codelist in c("new", "ctv3")) {


```{r}
df_sus_new_cross_table_new
df_sus_new_cross_table_ctv3
```

<br>

```{r, fig.width=12, fig.height=10}
alluvial_new
```{r, fig.width=20, fig.height=15}
# alluvial_new
```

<br>

```{r}
df_sus_new_cross_table_ctv3
df_sus_new_cross_table_new
```
<br>

```{r, fig.width=12, fig.height=10}
alluvial_new
```{r, fig.width=20, fig.height=15}
# alluvial_new
```

<br>
Expand Down Expand Up @@ -1284,8 +1284,8 @@ for (codelist in c("new", "ctv3")) {
),
left_paren = " (",
right_paren = ")",
N = scales::comma(as.numeric(`0`)),
population = scales::comma(as.numeric(population))
N = comma(as.numeric(`0`)),
population = comma(as.numeric(population))
) %>%
arrange(ethnicity, ethnicity_sus_16) %>%
unite("labl", N, left_paren, percentage, right_paren, sep = "", remove = F) %>%
Expand Down Expand Up @@ -1335,12 +1335,12 @@ for (codelist in c("new", "ctv3")) {
```

```{r}
df_sus_new_cross_table_new
df_sus_new_cross_table_ctv3
```
<br>

```{r}
df_sus_new_cross_table_ctv3
df_sus_new_cross_table_new
```
<br>

Expand Down Expand Up @@ -1368,10 +1368,11 @@ ONS_tab_2001 <- ethnicity_na_2001 %>%
mutate(region = fct_relevel(region, "England")) %>%
arrange(region)
my_cols_ons <- setNames(c("Region", rep(c("CTV3 2020", "CTV3 2020 with SUS data", "SNOMED 2022", "SNOMED 2022 with SUS data", "2021 ONS Census*"), 5)), names(ONS_tab_2001))
my_cols_ons <- setNames(c("", rep(c("CTV3 2020", "CTV3 2020 with SUS data", "SNOMED 2022", "SNOMED 2022 with SUS data", "2021 ONS Census*"), 5)), names(ONS_tab_2001))
ONS_tab_2001 %>%
gt(groupname_col = "region") %>%
gt() %>%
tab_spanner(label = "Region", columns = c(1)) %>%
tab_spanner(label = "Asian", columns = c(2:6)) %>%
tab_spanner(label = "Black", columns = c(7:11)) %>%
tab_spanner(label = "Mixed", columns = c(12:16)) %>%
Expand Down Expand Up @@ -1517,10 +1518,11 @@ ONS_tab_2001 <- ethnicity_na_2001 %>%
mutate(region = fct_relevel(region, "England")) %>%
arrange(region)
my_cols_ons <- setNames(c("Region", rep(c("CTV3 2020", "CTV3 2020 with SUS data", "SNOMED 2022", "SNOMED 2022 with SUS data", "2021 ONS Census*"), 16)), names(ONS_tab_2001))
my_cols_ons <- setNames(c("", rep(c("CTV3 2020", "CTV3 2020 with SUS data", "SNOMED 2022", "SNOMED 2022 with SUS data", "2021 ONS Census*"), 16)), names(ONS_tab_2001))
ONS_tab_2001 %>%
gt(groupname_col = "region") %>%
gt() %>%
tab_spanner(label = "Region", columns = c(1)) %>%
tab_spanner(label = "Indian", columns = c(2:6)) %>%
tab_spanner(label = "Pakistani", columns = c(7:11)) %>%
tab_spanner(label = "Bangladeshi", columns = c(12:16)) %>%
Expand Down Expand Up @@ -1627,7 +1629,7 @@ ons_ethnicity_plot_eng_na

### Region

```{r, fig.width=25, fig.height=20}
```{r, fig.width=50, fig.height=30}
## 16 group ethnicity plot NA removed for Regions
ons_ethnicity_plot_na <- ons_na_removed %>%
filter(region != "England", group == "16") %>%
Expand Down Expand Up @@ -1674,7 +1676,7 @@ This report describes existing methods to derive primary care ethnicity in OpenS

# Technical details

This notebook was run on 2023-02-16. The information below is based on data extracted from the OpenSAFELY-TPP database on 2023-02-16.
This notebook was run on `r Sys.Date()`. The information is based on data extracted from the OpenSAFELY-TPP database on `r date(file.info(here::here("output","extract_5","input_5.feather"))$ctime)`.

If a clinical code appears in the primary care record on multiple dates, the latest date is used.

Expand Down

0 comments on commit 90de294

Please sign in to comment.