As mentioned on page 
, the following two definitions for a heavy
hypoglycemic episodes are used:
The program calculates the number of hypoglycemic episodes in the following way: First, the latest entry in each quarter and each year is determine for each patient. Now, the last twelve months before those sets are checked for hypoglycemic episodes which are saved in the fields hyp_s and hyp_koma in the table progression (verlauf). If there is no entry in those twelve months, because the patient was examined only once or did not appear in that period of time, the space of time is set to one year. This is done for all patients, and the result is scaled to a bases of ``100 patient years ''.
In this analysis, two different result are computed. They differ in the interpretation of empty hypoglycemic episodes fields. For the first result, it is assumed that empty fields are equal to zero and this has just not been entered into the database . For the second result, it is assumed that empty fields have not been quantified and so the appropriate period of time is ignored. In both cases, the total number of hypoglycemic episodes is the same, but the underlayed period of time usually shorter in the second part. If all the fields about hypoglycemic episodes are not empty, those two results should be equal.
The last information calculated here is the number of stationary receptions due to a ``acute hyperglycemic episodes or ketoacidosis ''. Those are exactly the receptions with a 3 entered in the field reason (grund) in the table. stat1 Again, this number is scaled to a basis of ``100 patients per year''.