Research

Contact

Dr. Eleonora SchönherrPresse- & ÖffentlichkeitsarbeitPhone: +49 40 42818-269
Mail: presse(at)bnitm.de

Julia HäberleinPresse- & ÖffentlichkeitsarbeitPhone: +49 40 42818-264
Mail: presse(at)bnitm.de

Essential Control Group

Clinical studies occasionally yield conflicting results for the same objective. For example, there have been studies indicating that patients with malaria come down with life-threatening salmonella blood-stream infections both more frequently and less frequently than patients without concomitant malaria. Together with international colleagues, we have shown that such contradictions result from differences in study designs.

Formulas for the statistical analysis of complex study designs
Formulas for the statistical analysis of complex study designs

The group used for comparison (control group) is critical. Comparing the frequency of malaria among patients who have salmonella blood-stream infections to the frequency of malaria among patients hospitalised with any other disease gives a distorted result. The appropriate control group was a group of patients with a similarly severe disease as the salmonella group, for instance with blood-stream infections by bacteria other than salmonellae, because, regardless of malaria, they had the same probability of being treated as inpatients in the hospital. Applying this study design, we found that patients with malaria indeed have twice as frequently salmonella blood-stream infections than those without concomitant malaria. Using the unselected control group yielded exactly the opposite result.


Krumkamp R. et al., Clin Infect Dis 2016, 62 Suppl 1:S83-9

Ralf Krumkamp, Benno Kreuels, Benedikt Hogan, Anna Jaeger, Lisa Reigl, Jürgen May and external cooperation partners (see publication)