The ICD 10 (G00-05) search according to the methods described above yielded a total of 73 cases (ICD-10 database). Electronic search of discharge summaries for the terms “meningitis”, “encephalitis”, “enzephalitis”, “myelitis”, “encephalomyelitis”, and “enzephalomyelitis” yielded a total of 902 cases (clinical database). The clinical database and the ICD-10 database were merged and duplicate entries and multiple hospitalizations were again deleted. Fig. 1 provides an overview of the merging process. The diagnostic labels according

to the diagnoses listed in the discharge summary yielded Quizartinib supplier the following distribution of unique and overlapping diagnoses (Fig. 2) Applying the Brighton Collaboration algorithms yielded a distribution, which was considerably less complex ( Fig. 3). A total number of 108 cases were ruled out entirely. Diagnostic labels and BC levels of diagnostic

certainty were compared. Overall rates of agreement (ORA), positive percent agreement (PPA) and negative percent agreement (NPA) were calculated for each level of diagnostic certainty. Table 1 demonstrates AZD2281 molecular weight that ORA ranged from of 77 to 98% for ENC, MYE, and ADEM. Again, as expected for a confirmatory test, levels of positive percent agreement (PPA) were lower than values for negative percent agreement (NPA). The comparison of ASM showed 67% ORA in Level 1, but a significantly lower value at Level 2 (38%), reflecting the overlap with cases of bacterial meningitis (see Section 3.5.2). Point estimates

and 95% confidence intervals were constructed, using because the total sample size for which comparative assessments were available (n = 255) for all calculations. Table 2 shows the results for ASM, BM, ENC, MYE, and ADEM for any level of diagnostic certainty. In most instances, NPA was higher than PPA, which is consistent with a confirmatory test rather than a screening tool, as reported previously in the evaluation of BC definitions [35] and [36]. As mentioned previously, cases of BM were included as negative controls and tested against the BC definition for ASM. As expected, we found significantly lower levels of agreement between a clinical case of BM and the BC category of ASM. Of the 140 cases with an exclusive clinical diagnosis of aseptic meningitis, 96 (68.6%) fulfilled the BC definition for ASM, 44 cases did not fulfill the definition for ASM. In 39 of these discordant cases, no documented gram stain report was available upon chart review. A negative gram stain is a major criterion and required for any level of diagnostic certainty in the Brighton Collaboration definition of ASM.