Abstract
Diagnosing atypical parkinsonism can be an error-exposed undertaking in the context of elaborate criteria coupled with time restraints on their comprehensive application. We conducted a retrospective, descriptive study of diagnostic accuracy among physicians at two tertiary neurology centers in Romania and developed an algorithmic tool for comparison purposes. As many as 90 patients qualified for inclusion in the study, with 77 patients actually complying with atypical parkinsonism criteria. Overall, physician-established diagnoses may be incorrect in about one-fourth of cases. The reasons for this finding span a wide range of possibilities, from terminology-related inaccuracies to criteria sophistication. A Boolean-logic algorithmic approach to diagnosis might decrease misdiagnosis rates. These findings prepare the ground for the future refinement of an algorithmic application to be fully validated in a prospective study for the benefit of patients and health professionals alike.