Correspondence
Primary Care in Germany – an International Comparison: Results of a Survery Using a Validated Instrument: In Reply
Dtsch Arztebl Int 2008; 105(5): 84. DOI: 10.3238/arztebl.2008.0084b
; ;
As described, this was a simple random selection. Precisely because of the low response rate, special arrangements were necessary to check the structural identity of the sample and total population. One such expedient was weighting according to relevant influencing variables. In Germany, these were gender, specialty area, and Federal State. Although these were representatively distributed for Germany (see additional table 2), this was not the case in all countries. In the UK, for example, there were considerable structural deviations in terms of gender and region. Since the structural identity of the sample with the total population was not assured from the outset, the weighting was required as described. A further provision was to compare the sample with the total population for two further possible influencing variables, for which no weighting had been performed: age distribution and proportion of physicians in one-person practices (see supplement). Here too there were no relevant differences. As a third step, the multivariate analyses were adjusted according to other potential influencing variables.The definition of significance used in the article is indeed to be understood purely descriptively, as expressly stated in the methodology section. Relevant hypotheses are therefore only generated through the descriptive p values and not statistically proved. Further, confirmatory studies would be necessary to deliver such proof. An adjustment for multiple testing is not useful in a descriptive analysis because the type and number of hypotheses are not established from the outset. The hypothesis that workload could be a source of professional dissatisfaction is also only generated, but not statistically proved, by the results of the survey. DOI: 10.3238/arztebl.2008.0084b
Ulrich Gehrmann
Klaus Koch
Prof. Dr. med. Peter T. Sawicki
Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen
Dillenburger Str. 27
51101 Köln, Germany
koch@iqwig.de
Conflict of interest statement
The authors of all articles declare that no conflict of interest exists according to the guidelines of the International Committee of Medical Journal Editors.
