Data Management Center DMC

1

• Data Management Center (DMC)
• Stan Azen PhD Director
• Carolee Winstein PhD, PT, FAPTA
• Principal Investigator
• James Baurley DMC Representative

2
Overview

• PART I - Data Management Center
• PART II - PTClinResNet Website
• PART III - Development Process
• PART IV - Structure of DMSC Report
• PART V - Project Management
• PART VI - Future Plans

3
PART I - Data Management Center

• Organization of the DMC
• Responsibilities of the DMC

4
Organization of the DMC

• Coordinators
• Stan Azen PhD
• Carolee Winstein PhD
• Samantha Underwood Patricia Pate
• Informatics Team
• James Baurley
• George Martinez
• Mike Hutchinson
• Statistical Analysis Team
• Carolyn Ervin PhD
• Tingting Ge

5
Organization of the DMC

• Data Entry Team
• Chris Hahn
• JoAnne de los Reyes
• Frances Chien
• Karina Kunder
• Jason Villareal

6
Responsibilities of the DMC

• Finalized the four study protocols with regard to
  design issues, sample size requirements,
  statistical analysis methods.
• Developed the PTClinResNet website
• Defined and built the public and secure sections
• Organized the network into a user friendly
  interface

7
Responsibilities of the DMC

• Designed and implemented study databases and
  web-based data entry application
• Developed the randomization procedures
• Developed a prototype template for reporting
  progress and safety information to the Data
  Monitoring Safety Committee (DMSC)
• Statistical analyses, quality control and
  reporting

8
PART II - PTClinResNet Website

• Located at
• http//pt.usc.edu/clinresnet
• Features of
• Public Website
• Secure Website
• Example - Reports Availability
• Document Management
• Example - Request for Documents

9
(No Transcript)
10
Features of Public Website

• Overview of network and study information
• Background and responsibilities of key personnel
• News items, conference information and
  announcements
• Information for potential study participants

11
Features of Secure Website

• Manual of Procedures for each study
• Reports to the Steering Committee and PT
  Foundation
• Minutes of conference calls with Study
  Investigators
• Recruitment Status
• Reports to the Data Monitoring and Safety
  Committee

12
(No Transcript)
13
Document Management System

• Manages PTClinResNet documents.
• Designed to limit read and write access to
  documents based on user groups.
• Interested researchers can request access.
• Currently in development by Statistical
  Consultation and Research Center.

14
(No Transcript)
15
PART III CLINICAL TRIAL DEVELOPMENT PROCESS

• OVERVIEW
• MANUAL OF PROCEDURES (MOP)
• DATA ACQUISITION DESIGN
• IMPLEMENTATION
• DATA ENTRY AND QUALITY CONTROL
• DATA ANALYSIS
• DATABASE STATISTICS

16
Network Diagram
17
(No Transcript)
18
MANUAL OF PROCEDURES (MOP)

• Provides a central document for the procedures of
  a clinical study
• Specific Aims
• Relevant scientific rationale
• Study design and statistical methods
• Procedures (randomization, data management,
  standardization, test administration, protection
  of subjects, etc.)

19
DATA ACQUISITION DESIGN
INSTRUMENTS Designed for accurate and complete
data collection DATA DICTIONARY A data structure
that stores metadata, i.e., a code book
containing information about the data being
collected. The data dictionary includes the
variable name, data type, allowable and missing
codes, value ranges, algorithms, and dataset and
version information.
20
DATA ACQUISITION DESIGN

• Created data collection forms in collaboration
  with investigators
• Developed multi-study forms and study-specific
  forms
• Multi-study forms utilize common data definitions
  (variable names and codes)
• Developed system for creating unique study and
  site-specific patient identification numbers
• Developed allowable and missing coding system
  for all variables

21
INSTRUMENT
22
DATA ACQUISITION DESIGN

• Create Data Dictionaries in collaboration with
  investigators. Data Dictionary fields include
• variable name
• data type (numeric, date, character),
• allowable and missing codes
• range
• field length
• whether the variable is required
• question as it appears on the form
• versioning
• dataset name

23
DATA DICTIONARY
24
DATA DICTIONARY IN SQL
25
IMPLEMENTATION

• Requirements for building physical database
• Final version of data collection forms
• Final version of data dictionary
• Final version of business rules
• Properties of database
• Security and menu-navigation, automated range
  checking, and auditing of users and changes in
  data values
• Training
• Manual for data entry
• Available on website

26
DATA ENTRY APPLICATION
27
System Accessibility

• Designated users for data entry and statistical
  analysis
• Customized security of research datasets,
  studies, and sites
• Research data restricted for blinded evaluators
  prior to trial completion.

28
Data Entry Process

• Develop log sheet to track data. 
• Includes dates of collection, data entry, data
  checking, data entry corrections.
• Maintains identifier of tracking personnel
• Data
• Form received and logged
• Form filed in locked filing cabinets
• Entered into SCRC data entry system
• Checked by comparing original data form to the
  data completeness report
• Corrected data entered

29
Quality Control Procedures

• Certification of evaluators
• Range and coding checks built into the data entry
  system
• Cross-sectional and longitudinal quality control
  checks at the SAS level
• Data completeness report

30
Data completeness/ quality reports
Example MUSSEL Education Form Data
31
DATA ANALYSIS

• ODBC-Compliant statistical packages (SAS, SPSS,
  STATA) allow real time access to the study data
  and data dictionary.
• Permits powerful control over data using
  Structured Query Language (SQL)

32
DATA FLOW DIAGRAM
33
Database Statistics - June 2005

• 2202 Variables
• 92 Shared
• 398 STEPS specific
• 478 MUSSEL specific
• 557 STOMPS specific
• 677 PEDALS specific
• 62 Datasets

34
PART IV - Structure of DMSC Report
DMSC Reporting

• Summary of Study Design
• Objective
• Subjects
• Sample Size
• Treatments
• Follow-up
• Endpoints Primary Secondary
• Summary of Analytic Plan

35
Structure of DMSC Report
DMSC Reporting

• Summary of Progress and Results
• Example - PEDALS
• Screening Trial Profile
• Total Subject Enrollment by Month of Study
• Primary Reasons for Ineligibility or Refusal
• Summary of Enrollment by Strata
• Baseline Demographics
• Compliance - Intervention
• Summary of Clinical Events

36
Study Profile
PEDALS Screening Trial Profile
37
Primary Reasons for Ineligibility or Refusal
Example - PEDALS
38
Recruitment Status
PEDALS - Total Subject Enrollment by Month of
Study
39
Baseline Characteristics
PEDALS Baseline Demographics
Mean for continuous variables frequency () for
categorical variables
40
Compliance
PEDALS Compliance - INTERVENTION
41
Adverse Events
Summary of Clinical Events
42
PART V - Project Management

• DATA REQUESTS
• PROJECT SCHEDULE
• Example - STOMPS

43
Example of Data Request Form
44
Protocol For Data Requests

• Process for data request
• Data request send to DMC
• Ticket number assigned recorded
• Request - prioritized assigned
• Request resolved
• Notification emailed to requestor

45
Project Schedule

• STOMPS Example

46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
PART VI - Future Plans

• Complete Primary Analysis
• Coordinate and Schedule Secondary Analyses
• Papers Study Specific and DMC
• Implement New Studies
• LEAPS
• ICARE

50
DMC Papers - In Development
DESCRIPTION OF A CLINICAL RESEARCH NETWORK FOR
THE EVALUATION OF PHYSICAL THERAPY
INTERVENTIONS James Baurley, Carolyn Ervin,
Tingting Ge, Stanley Azen, Carolee
Winstein Departments of Preventive Medicine and
Biokinesiology and Physical Therapy University of
Southern California, Los Angeles CA USA
BAYESIAN META-ANALYSIS OF EFFECTS OF STRENGTH
TRAINING INTERVENTIONS ON FUNCTION IN PATIENTS
WITH PHYSICAL DISABILITIES James Baurley, Stanley
Azen, David Conti, Carolee Winstein, Carolyn
Ervin Departments of Preventive Medicine and
Biokinesiology and Physical Therapy University of
Southern California, Los Angeles CA USA
ASSESSMENT OF THE COMPARABILITY OF THE TWO
VERSIONS OF SF-36 IN A PHYSICAL THERAPY CONTEXT
Tingting Ge, Stanley Azen, Carolyn Ervin,
James Baurley, Carolee Winstein Departments of
Preventive Medicine and Biokinesiology and
Physical Therapy University of Southern
California, Los Angeles CA USA
51

• QUESTIONS?

52
Project Schedule

• PEDALS Example

53
(No Transcript)
54
(No Transcript)
55
Project Schedule

• MUSSEL Example

56
(No Transcript)
57
(No Transcript)
58
Project Schedule

• STEPS Example

59
(No Transcript)
60
(No Transcript)
61
Project Schedule

• STOMPS Example

62
(No Transcript)
63
(No Transcript)
64
(No Transcript)