End of GP registration date
Hi,
I am trying to calculate primary care follow-up time for small sub cohort of the UKB population. The variable (deduct date) is missing for quite a few of the participants and I am unsure how to estimate the date at which their GP records stopped contributing data. There is a graph in the documentation of drops in patients by primary cares system, but that is no good for making assumptions at individual level. Given my subcohort is very small, this has a great impact on the analysis. if “deduct date” is missing should i estimate with last GP encounter for those patients? HES is no good, since the condition i am looking at during follow-up is mainly diagnosed in primary care not inpatient care.
Any ideas would be greatly appreciated, thanks!
Comments
1 comment
Hi Meda,
In general, where there is no deduct_date, it may be likely that the participant was still registered at the practice at the time the data extract was generated. More information about file extract dates (which varied between and within providers) and possible considerations in determining censoring dates is available in Resource 591, section 4.2.
Please be aware however of the caveats around primary care registration data and and its coverage. This is discussed in more detail in the same resource, in section 4.1. Please also note that the UK Biobank study is not designed for researchers to study very small case numbers for their research; you may wish to consider this carefully when planning analyses.
Thanks,
Laura
Please sign in to leave a comment.