Medidata AMUG Meeting / Presentation 2013

Having  your  Cake  and  Ea1ng  it  Too  –  Leveraging  RWS  with  

Phase  1  Studies  

Brock Heinz Spaulding Clinical March 19, 2013

About me and Spaulding Clinical Research •  Brock Heinz - Engineering / Innovation; successfully demoted;

introduced to Medidata in 2009; Technology Partner

•  Established in 2007 with a team of experts from pharmaceutical, CRO, clinical practice, and the medical device industries with over 150 years of combined experience; Located in West Bend, WI – 30 miles north of Milwaukee

•  Highly integrated and automated phase I clinical pharmacology

research unit; 155 beds – 96 telemetry

•  Data Management, Biostatistics and Medical Writing Services

•  Full Service Core ECG Lab

•  Medical Device Manufacturer

2  

Phase 1 trials at Spaulding Clinical •  EDC true to the acronym: truly using electronic means for

data capture; not just electronic storage •  Barcode-driven data collection – right subject each time •  Integration of ECG, vitals, labs

•  Rapid data lock is the rule, not the exception •  Data cleaning is a dirty term •  Sponsors always have real time access to study data •  Basic philosophy I’ve helped instill – let computers do what

computers are good at, thus maximizing the human touch •  Computers consistently do the right thing at the right time •  Humans work well with humans – keep the peace with

subjects •  Reducing variability is good science

3  

Phase 1 Trial – at a Glance

4  

What’s the Cake? •  The cake in this context is a paperless and and automated

Phase 1 study seamlessly integrated with the sponsor’s EDC system of choice – Rave

Who’s Cake is it? •  Sponsors who have a desire to maintain study data

throughout the lifecycle of a compound in Rave

5  

How can one have their cake and eat it too? Introducing SCi Rave (pronounced sky) •  The customer is always right. Sponsors understand their data

structures and their processes, data standards come from the top down.

•  Modular system consists of three major components

•  Rave – where are we going? •  ETL – how / what should we pack? •  Interface Engine – how do we get there?

•  Process Overview •  Upload / parse loader file •  Develop / validate ETL •  Schedule and execute transfers

6  

Loader file driven design – begin with end in mind

7  

Data model – driven by Architect Loader

ETL – packing for the trip

•  Extract, Transform, Load •  Incredibly flexible model – layer of indirection can obtain data

from other EDC systems, relational databases, flat files, etc •  Two primary tasks:

–  Identify subjects –  Query study repository(ies?) and populate tables created

by loader parsing process •  Script is uploaded through our web interface and is compiled

and executed on the server •  Validate ETL script!

–  Pre-transferred data can be reviewed in Excel / CSV

9  

Let’s go! Scheduling and executing the transfer

•  Transfer schedule is flexible – can be dictated by sponsor •  Designate transfer schedule

–  Automatic and near real-time –  Ad hoc

•  Transfer schedule set from web interface •  Subjects inserted, form (Item Group) data sent •  Each transaction response from RWS is stored with the Item

Group for which it was sent; available for review in web dashboard

10  

SCi Rave – system architecture

11  

Configuring system

12  

Validating ETL; ad hoc interaction

13  

RWS transactions

14  

Memories of a recent trip

•  3 studies conducted in rapid succession •  3,926 Folders •  18,803 Forms •  54,039 Item Groups (includes log lines) •  322,685 Items •  1,460,485 characters of time-point data •  78,796 HTTPS Requests to RWS •  0 CRFs manually transcribed from the Spaulding Clinical system

into Rave

15  

Final thoughts

I’ve been around web services for a while – this is well done: ODM, interoperable protocols, intuitive response messages, documentation, user group, etc. As leaders in EDC Medidata is refreshingly different than EMR vendors. What’s next? •  I think we’ll seen see an evolution of drug development

process. There will be disruption as there is a push towards personalized medicine. Tighter iterations, richer data.

•  Big Data – sensors; distributed telemetry; integrated data repositories. Mobile integration.

•  Empower innovators – RWS gives us the highway needed. •  Device level integration - today

16  

Spaulding webECGTM

•  Hand-held diagnostic electrocardiograph uploads data to web-based ECG Management System

•  Biometric voice ID eliminates demographic entry

errors

•  Single button allows for malleable user interface

•  Stores up to 5 minutes of 12-lead ECG data •  Automated report available immediately

•  Nearly instant over-reading

Model 1000iQ Electrocardiograph

Integrated data collection

18  

ECG document – demographics via RWS

19  

Automatic transcription

20  

21  

Thank you!