Right now the missing UDS records are defined as no record in visit and no record in uds. The rationale behind it is that the patients seldomly (only 1%?) refuse to provide the UDS results when he/she visits, and both uds and uds_temp do not contain any negative results.
It is better to have another data set indicating the existence of UDS records, even no positive record for any substance. Or I prefer to expand the data set uds_temp to include all UDS no matter positive or negative (if we have), there might be some records in uds_temp with all substances negative but _was_tempok is 0 (means the temperature is abnormal).
Besides, it is better to have documentation to describe this situation.
Right now the missing UDS records are defined as no record in
visitand no record inuds. The rationale behind it is that the patients seldomly (only 1%?) refuse to provide the UDS results when he/she visits, and bothudsanduds_tempdo not contain any negative results.It is better to have another data set indicating the existence of UDS records, even no positive record for any substance. Or I prefer to expand the data set
uds_tempto include all UDS no matter positive or negative (if we have), there might be some records inuds_tempwith all substances negative but _was_tempok is 0 (means the temperature is abnormal).Besides, it is better to have documentation to describe this situation.