derive input files/subsets for clustering analysis. includes separate .rda files containing dfs for each age bin, with 0/1 denoting diagnosis of ltc within given age bin. includes variables specifying droput due to death/age of latest HES as well as age of dementia diagnosis. outputs written to ../../derivatives/

csv containing OR for each individual ltc. also contains cols denoting include/exclude based on OR<1, and subsequent groupings (G, EYE, EAR, ANXST)

reorg for new script ordering. now also only computes LR of single ltcs in the analysis sample only (65 age cutoff). saves analysis sample to derivatives folder, saves exclusions (dementia < 65, death < 65, follow up < 65) to derivatives as well

more reorg...derive age of diagnosis first. then run individual condition LR on the analysis subset vs all data. use predementia diagnosis vs predementia one in determining valid/pre dementia diagnosis

changes to derive ltc diagnoses from death reports, though these are few. will need to edit order of rest of scripts after 3_derive_death_diagnoses.R

added code to derive predementia_diagnosis_of_ltc, in addition to predementia_one/two/five_diagnosis_of_ltc, which is 1 if diagnosis date is on or before dementia diagnosis date

script to derive age at diagnosis for each ltc. also creates .rda file for age bins of 0-55,55-65,65-70,70-75,75-80,80-85 where value=1 if ltc has diagnosis in given age range

create dementia cutoff dates to use in assessing if ltc id pre/post dementia diagnosis. merge dementia diagnosis variables with main ltc diagnosis set

code for merging pcare and ukb alg diagnosis data, dementia dx is 1 if either sources record a diagnosis. earliest date used

adjusted loop to account for rare cases where diagnosis is 1 but no diagnosis date is recorded (n=16). set diagnosis to 0 for these

cycle through ltcs, assign 1 if either pcare or HES diagnosis is 1. set diagnosis date to earliest date

save output to derivatives instead of raw data..its not raw data. add numbering to indicate step in workflow

added code to load ltc pcare data, merge and format, write out to .rda containing diagnosis and date column for ltcs, derived from pcare

script to mine gp records for specificed read codes. outputs a sheet participants_with_LTC containing the id and event dates of the codes. written to run on a specificed split of gp clinical data. must give A or B as input

a matching file to ltc_codes.csv but contains readcodes instead of icd codes for each ltc. to be used in gp records extraction

refining icd codes based on comparisons to published works. dropping some columns im not using (eg painful conditions)

derive a pre dementia  HES diagnosis - where diagnosis = 1 if and only if diagnosis date is 1 year prior to dementia date. use this in derivationg of age at diagnosis and age bin analysis