methods.qmd

---
title: "Methods"
author: "Jacqueline Grout"
format: html
editor: visual
---

## GP Survey

Technical annex:

Q55 base: Smoking habits - Total responses

Q55_1pct: Smoking habits - % Never smoked

Q55_2pct: Smoking habits - % Former smoker

Q55_3pct: Smoking habits - % Occasional smoker

Q55_4pct: Smoking habits - % Regular smoker

<https://gp-patient.co.uk/downloads/2023/GPPS_2023_Technical_Annex_PUBLIC.pdf>

## CHD synthetic prevalence estimates

CHD synthetic prevalence estimates data has been removed from Fingertips by OHID.

OHID have sent the archived data. This is for 7564 practices from 2015, of which 6297 are current practices.

Need to identify practices with missing data and work out how to get a prevalence for them.

Also have file from previous project - checked data is the same but need to check practices included - maybe there are more?

nearest neighbours to impute chd prevalence

## Census 2021 Ethnicity by LSOA

Some LSOAs that have gp list population assigned to them do not feature in the census data of ethnicity by lsoa. Need to check if these are the changes from 2011 to 2021 census.

Possible solution would be to work out the average proportions for all the other LSOAs for the practice and distribute the gp list population for the lsoas that have no match proportionally across the ethnic groups.

Another option would be where lsoas in 2021 have a parent lsoa in 2011 then apply the average ethnicity % for the child lsoas to the parent (this is when one 2011 lsoa has been split into multiple 2021 lsoas.) Where the lsoas in 2021 have a child lsoa in 2011 then apply the ethnicity % for the 2021 parent to the 2011 children (this is when several lsoas have been combined into one new lsoa)

join lsoa_lookup to census 21eth using the 2021 lsoa code in the lookup dataset.

Have opted for the latter solution.

Imputing data in Census 2021:

<https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationestimates/methodologies/itemeditingandimputationprocessforcensus2021englandandwales>

Variation in age groups for different ethnicities:

<https://www.ethnicity-facts-figures.service.gov.uk/uk-population-by-ethnicity/demographics/age-groups/latest/>

<https://www.ethnicity-facts-figures.service.gov.uk/uk-population-by-ethnicity/demographics/age-groups/latest/#download-the-data>

## Which date should be used for the gp list size dataset?

Data mainly 2022/23 could take April 22 or April 23 file. Will try both. April 2023 is still using census 2011 lsoas.

<https://digital.nhs.uk/data-and-information/publications/statistical/patients-registered-at-a-gp-practice/october-2022>

Have decided to use October 2022 and also downloaded the GP to icb mapping file from the same period and joined to the gp list size to easily create analysis by icb/sub-icb.

Used October 2022 gp list size by age and gender to calculate % of list size aged 45+

## Clustering

### Notes

<https://www.statology.org/k-medoids-in-r/>

https://www.rdocumentation.org/packages/cluster/versions/2.1.6

look at 18:50 in this video: https://www.youtube.com/watch?v=dKXTNxNDpMA

## Mapping new GP practices to old

### To obtain CHD prevalence for as many new practices as possible

link to to the blog

<https://the-strategy-unit.github.io/data_science/blogs/posts/2024-01-17_nearest_neighbour.html>

151 missing and this finds all but 26 using a 1.5km radius - more or less?

## Disparity

<https://www.health.state.mn.us/data/mchs/pubs/raceethn/rankingbyratio20032007.pdf>

## Smoking PCAS

Metric 12 - Exception reporting for Smoking cessation support offered

<https://digital.nhs.uk/data-and-information/publications/statistical/quality-and-outcomes-framework-achievement-prevalence-and-exceptions-data/2022-23>

Donwloaded:

qof-2223-prev-ach-pca-ls-prac.xlsx

## References

<https://journals.sagepub.com/doi/epdf/10.1093/phr/117.3.273>