Information computing technology for analysis of preventive screening

Information computing technology for analysis of preventive screening data in Health Centers Starunova O. A. Federal Research Institute for Health Organization and Informatics, Moscow, Russia

Introduction • Health centers: mass data of preventive screening • Feature 1: big data (? ) • Feature 2: Significant number of incorrect data: – Outliers – Frauds • Need for automated data management and quality control

Materials and methods • Databases: – Federal Information Resource of Health Centers database (4. 78 mln records) – bioimpedance database (2. 35 mln records) – anthropometric database (7. 88 mln records) • Data storage: – SQL database (100 Gb) – Server: R-IT Data. Mill , 24 Tb HD, MS Windows Server 2012, MS SQL Svr Standard 2014 10 Tb • Software – R with libraries: shiny, gamlss, sp

Results HCViewer: software for Health Centers data analysis – Interactive map of Health Centers – Data filtration (outliers & frauds removal) – Data statistics – Percentile curves – Geographic distributions of data

HCViewer interface

HCViewer interface

HCViewer interface

HCViewer interface

HCViewer interface Total number of observations Total number of incorrect observations

HCViewer interface

HCViewer interface 20% of Health Centers produce 80% of incorrect data

HCViewer interface

HCViewer interface

HCViewer interface

To be realized: centile curves construction (gamlss model) Age, years

Data filtration: application of Benford's law 0, 4 0, 35 Part of numbers 0, 3 0, 25 Benford’s Law Normal(0, 1)^10 0, 2 0, 15 0, 1 0, 05 0 1 2 3 4 5 6 7 First significant digit 8 9

Benford’s law residual as a criteria of data base quality % of incorrect data

THANKS FOR YOUR ATTENTION