gms | German Medical Science

Background: A recent analysis of prescriptions of benzodiazepines and Z-drugs revealed that the first month of each quarter showed a maximum number of prescriptions and in the following month a minimal number of prescriptions (cf. Figure 1 [Fig. 1]). Possible reasons could be budget or seasonal impacts. Further analysis showed that prescriptions of all available drugs follow nearly the same pattern. Moreover, the main effect with respect to drugs with long application duration is due to public or school holiday shifts (cf. Figure 2 [Fig. 2]). Drugs designated for short application like antibiotics follow other prescription patterns (cf. Figure 3 [Fig. 3]). With respect to these holiday shifts we considered the variations in the daily number of patients and physicians with respect to the same weekday and the weekday pattern with respect to patient age.

Materials and Methods: We analysed prescription data of the statutory health insurance of the German federal state Schleswig-Holstein and the treatment cases (with and without drug prescriptions) for general medicin e. In order to select drug groups we applied the international ATC (anatomic- therapeutic- chemical) system with German specifications provided by the DIMDI Institute. Data aggregation was done using the script language Perl. For the additional calculations we moved to the program system Mathematica by Wolfram Research. Because of superimposing 7 day frequencies, a weekday smoothing was applied (cf. Figure 4 [Fig. 4]). Prescriptions per day were utilized due to the different lengths of the various months.

Results: Highest numbers of prescriptions were most frequently observed at Mondays (+43%) in relation to the mean without weekends and public holidays, the lowest number value is at Fridays (-35%). There were almost the same relations regarding the number of prescriptions and the expenditures (cf. Figure 5 [Fig. 5]). By applying a 7 day smoothing, a curve with minimal values at or nearby Christmas, New Year, Easter, Ascension Day, Pentecost, First of May and school holidays resulted (cf. Figure 6 [Fig. 6]). Between Christmas and New Year there was a decrease of 70% and before and after that time an increase of 20% till 40%. Deviations in the mean number of patients per physician and the mean number of prescriptions followed nearly the same holiday pattern. During summer break the decrease in patients exceeds the decrease of prescriptions, during December and January (with exception of the first week) the increase in patients exceeds the increase of prescriptions (cf. Figure 7 [Fig. 7]). Due to the unequal distribution of the public and school holidays with respect to the months of the year we observed unequal prescriptions per day and number of patients per day in the range of +9% till -12%. When all prescriptions were used as a baseline benzodiazepines had a 5% higher value in January and December, all other differences were less than 3%. A 15-20% increased number of prescriptions of antibiotics were found in February, March and December, a 10-15% lower value from May till August (cf. Figure 8 [Fig. 8]). In the time before and after holidays there were more physicians available, but the increase in patients exceeded this effect (cf. Figure 9 [Fig. 9]). Regarding an age depending pattern of drug prescription with respect to the days of the week there was a low difference between Tuesday and Thursday as well as Wednesday and Friday in the case of patients younger than 50. Above that age there were increased prescription numbers at Tuesday and Wednesday and decreased numbers at Thursday and Friday. Regarding patients older than 80 years a shift of prescriptions was observed from Fridays to Mondays (cf. Figure 10 [Fig. 10]).

Conclusion: Prescription shifts due to the unequal distribution of holidays during the year therefore might be misinterpreted as a seasonal or any other pattern. Our data indicate that the patients had built up stocks of drugs for use during holidays and additionally for reducing waiting time in physician’s office. It is noteworthy that we only looked for variations in general medicine. Thus, it is important to analyse prescription patterns regarding other specialities particularly with respect to specific drug groups.