119 A pilot study of an algorithm designed to identify Parkinson's disease in the early, non-motor phase

A number of risk factors have been reported before the onset of overt motor signs indicative of PD. These include anosmia, RBD, constipation, and depression. Smoking and use of caffeine have been suggested to reduce risk. Individually, most factors convey only a modest effect on overall risk, but may summate with considerable predictive power. Extensive literature review was used to consolidate data on factors that may alter individual risk of PD. A computerised questionnaire delivered through a secure Internet website will be used to screen members of the general population for these factors. Integrated statistical software will select individuals with increased likelihood of future PD using an algorithm combining the different factors. Items that have been selected to give an overall predictive score include: age, gender, family history, mood, various symptoms of dysautonomia, and smoking, caffeine, statins and NSAID usage. As well as generating a score, these factors can be validated prospectively, along with other potential factors for which evidence is incomplete. This pilot study is designed to evaluate the accuracy of the algorithm. A cross-sectional analysis will use hyposmia (tested using the UPSIT) and RBD (tested using a validated questionnaire) as surrogates for those at high risk of PD. 1000 individuals aged between 60 and 80 years will be recruited and tested via Internet. The cohort of participants will be followed-up for the future development of motor and non-motor symptoms of PD. Fast-track testing of predictive algorithms could help establish the relative value of each factor alone or in combination before more costly prospective studies are started.