Idiopathic parkinsonism (IP) is a common disorder, conventionally regarded as neurodegenerative. Its cardinal features, poverty and slowness of movement, muscle rigidity, postural abnormality and a characteristic tremor, are associated with loss of dopaminergic neurones in the substantia nigra of the brain. Genetic factors explain only a minority of cases, and a common toxic environmental insult remains elusive. We propose that IP is a systemic disorder resulting from a ubiquitous peripheral infection, and that only the tip of the iceberg comes to diagnosis. There is evidence for inflammatory/immune activation peripherally and in the brain. We have used statistical modelling to explore links with non-specific and specific systemic markers of inflammation/infection in IP probands, and explore whether their partners and siblings have a frank or pre-presentation parkinsonian state. Critical to this approach is continuous objective measures of the facets of IP. Hypotheses on causality and mechanism are based on the statistical models. There is pathological and clinical evidence for direct involvement of the gastrointestinal tract in IP. The candidacy of Helicobacter pylori infection as a trigger event or driving infection is relatively high. We have found that eliminating infection in late parkinsonism with cachexia, a stage usually considered intractable, can result in a U-turn. However, eradication therapy may not provide a complete solution. Persistence of antibody against cytotoxin-associated antigen (CagA), increases the predicted probability of being labelled as having parkinsonism. Evidence for autoimmunity and immunocompromise is used to build schemes for the natural history. We conclude that current classifications of neuropsychiatric disease may not prove the best with respect to defining sub-clinical disease, prophylaxis or halting progression.