194 Chapter 7 information, and Thuisarts.nl has been shown to be a safe and effective online platform that can inform and reassure patients (51). Assisting General Practitioners Diagnostic uncertainty is a major determinant of inappropriate antimicrobial prescribing. In the case of an RTI this can be addressed through use of the C-reactive protein point-of-care (CRP-POC) test. When a GP is in doubt, a CRP-POC test can be used to discriminate between an uncomplicated versus complicated RTI. Use of this test has proven effective in lowering the number of antimicrobial prescriptions (52-56). Antimicrobial overprescribing for RTIs and overprescribing of macrolides can be tackled using several interventions. For example, GP communication training on RTIs (57), GP education and a feedback session on antimicrobial prescribing were all effective in reducing prescriptions (41, 58-63). Feedback sessions may provide insight concerning the number of antimicrobial prescriptions a GP writes and their impact on antimicrobial resistance, which may in turn encourage a physician to reflect on his or her antimicrobial prescription habits. Large practice size and GPs failing to verify assumptions about a patient wanting an antimicrobial prescription were the main determinants associated with more inappropriate antimicrobial prescribing in chapters 3 and 6. The latter factor is the most likely explanation of higher inappropriate antibiotic prescribing for patients with a Turkish, Surinamese and Dutch-Caribbean background (chapter 6). This illustrates the benefits of efficient communication skills and having sufficient time to communicate with patients. Methodological considerations The outcomes and interpretation of the studies described here should be viewed in the context of the strengths and limitations of each study. The studies described in chapters 3, 4, 5 and 6 used routinely collected healthcare data. In chapter 6, a large healthcare registry was combined with a large registry containing data on socialeconomic determinants. A limitation of this approach is that health records are not primarily designed for research purposes, which can result in missing data as not all required information is systematically recorded. Missing data can lead to registration bias, causing either under - and over-registration. However, as the healthcare registries used in chapter 4, 5 and 6 contained very large amounts of data, any registration
RkJQdWJsaXNoZXIy MTk4NDMw