Abstract Details

Title Multifactorial Prognostic Models for Parkinson Disease: Integrating the Cardiovascular Health Study with Claims from the Centers for Medicare and Medicaid Services

Topic Movement Disorders

Presentation(s) P06 - (-)

Poster/Presentation
Number 104

Objective

Background BACKGROUND: We integrated data from the Cardiovascular Health Study (CHS) and claims from the Centers for Medicare and Medicaid services (CMS) to develop models which predict PD diagnosis.

Design/Methods DESIGN/METHODS: The CHS is a community-based cohort of 5,888 Medicare eligible participants ?65 years old at baseline with over 20 years follow-up linked with CMS claims. Data from both CHS and CMS claims were combined for adjudication of PD status. Over 50 potential predictors of PD diagnosis from CHS data (demographics, balance/walking/fatigue, cognition, depression, health and physiological measures, and serum biochemistry) were evaluated with logistic and bagged tree models. Incident PD was outcome. A separate logistic model for each predictor was adjusted for age, sex and smoking. Those with p<0.1 were combined into a multivariable logistic model. Bagged classification tree analysis, which averages models using multiple combinations of all predictors, was performed with 25 bootstrap samples. ROC curves evaluated classification of incident PD cases and non-PD cases.

Results RESULTS: There were 165 incident PD cases. The multivariable logistic model contained the following terms: age, sex, smoking, perceived weakness, general health, bioelectric impedance, digit symbol substitution test and ECG abnormality. Bagged tree analysis yielded models in which the most common predictors were hematocrit, white blood cell count, bioelectric impedance, interluken-6, smoking and c-reactive protein. The area under the ROC curve, sensitivity and specificity for the multivariable logistic model was 0.65, 56.3% and 67.5%. For the bagged trees it was 0.98, 84.9% and 84.7%.

Conclusions CONCLUSIONS: Epidemiological cohort data integrated with CMS claims identify novel combinations of predictors which may help in the prognostication of PD. Bagged tree results must be cross-validated. Prognostic models may be improved with more thorough ascertainment of PD cases in other cohorts and inclusion of additional predictors which may exist within CMS claims.

Authors/Disclosures
Samay Jain, MD (University of Pittsburgh Medical Center) PRESENTER	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
	No disclosure on file
Thanh Ton, PhD (Precision Health Economics)	No disclosure on file
Douglas L. Arnold, MD, FAAN (Montreal Neurological Institute, McGill Univ)	Dr. Arnold has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Biogen. Dr. Arnold has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for BMS. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Frequency Therapeutics. Dr. Arnold has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Merck. Dr. Arnold has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Novartis. Dr. Arnold has received personal compensation in the range of $10,000-$49,999 for serving as a Consultant for Roche. Dr. Arnold has received personal compensation in the range of $5,000-$9,999 for serving as a Consultant for Sanofi. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Shionogi. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Xfacto communications. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Eli Lilly. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for EMD Serono. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Biohaven. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Find therapeutics. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for GSK. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Idorsia. Dr. Arnold has received personal compensation in the range of $500-$4,999 for serving as a Consultant for Kiniksa. Dr. Arnold has received personal compensation in the range of $100,000-$499,999 for serving as a Consultant for Clario.

��ɫ��

��ɫ��