Microsoft Word search phase 3 Title Page Amendment

7.
 
Data Management 
The most popular method of developing a data management system is to use remote data 
entry through a web-based interface.  This method was used for the SEARCH Phase 1 and 2 
and will continue to be used in SEARCH Phase 3.  The WFUSM SEARCH Coordinating 
Center (CoC) data management system operates on a web browser user interface, which 
provides an easy to use, platform independent, data entry and retrieval environment.  Data is 
stored centrally at the CoC in a Windows SQL Server data warehouse.  Front-ends were built 
using HTML and ColdFusion middleware was used to integrate the SQL Server data within 
the HTML interface.  While the interface has been developed to be browser independent, 
browsers do not always display information consistently across versions.  Therefore, centers 
are encouraged to use a recent standards-compliant browser such as Firefox, Safari, Opera or 
Internet Explorer. 
In this model, the clinic staff uses a PC to access the study website and enter the data 
directly.  Inconsistencies found during the data entry process are correctable during the 
original data entry session, thereby reducing error.  Since data submitted is automatically 
entered into the central database, any reporting or further edit validation processes are based 
on the most current data.  Email messages and/or reports are generated at the time of data 
entry to alert the project managers of outstanding data entry issues. 
7.1.
 
QUALITY CONTROL 
The WFUSM SEARCH CoC is responsible for developing and implementing quality control 
(QC) procedures.  QC techniques are incorporated into each phase of the study from case 
ascertainment, recruitment and registration of persons with diabetes through data acquisition, 
reading and/or interpretation of the results and their analyses and publication.  The CoC will 
continue to work with the Protocol Oversight Committee reporting to the Steering Committee 
and to the External Scientific Advisory Committee.  The CoC created dynamic QC reports 
for the QC Committee, and will continue to utilize this highly effect method during the 
second phase of the study. 
7.2.
 
DATA ENTRY 
A web-based application for study data entry was implemented for SEARCH in Phase 1 that 
is complemented by a clinic-based application to manage participant contact.  Through use of 
the local application, the clinics will be able to better track the number of participant contacts 
that they have encountered. 
7.3.
 
ELECTRONIC FORMS 
Electronic forms will be used for SEARCH Phase 3.  Online versions of the forms have been 
developed that closely resemble the paper version as possible.  Electronic data is managed 

---

Section 7 - Data Management (Phase 3 - 12/2010) 
Section 7 - Page 2 
 
centrally at the CoC.  After reviewing case report forms (CRF) for gross errors and errors of 
omission, clinic staff begin the web-based entry process. 
After selecting a valid menu option and entering a valid participant ID, the staff member is 
given a list of forms that can be entered or edited.  After the selection is made, the form is 
displayed and any existing information is pre-filled.  Upon completion of entry or edits, the 
user can submit or cancel the form.  If the save option is selected, the appropriate tables are 
updated and any audit information is saved.  With the exception of registration data in the 
tracking database, no electronic data is housed at the individual study centers.  Data will 
reside on the database server at the CoC in a Windows SQL Server database. 
7.3.1.
 
Computerized Edit Checks 
The CoC performs numerous computerized edit checks to ensure data quality.  These 
include, but are not limited to: (a) initial screening of data, using logic and range checks 
that are built into data entry screens; (b) cross-form functional and consistency checks; 
(c) edits assessing the serial integrity of data, particularly in studies with longitudinal data 
collection; and (d) assessing means, ranges, and standard deviations of registry and 
laboratory data. 
All questions are pre-assigned missing values for the purpose of data entry.  The data entry 
screens require a set degree of completeness before a form can be considered finalized.  
Should a form be incomplete, the missing value would be entered into the database.  
Validation checks are applied during the data entry process.  Checks will be programmed 
using JavaScript routines activated as clinic personnel enter data, from each CRF.  
Additional data validation is performed on the server at the time the form is submitted.  
Feedback regarding the status of the form and any missing or inappropriate data is available 
to the data entry person immediately after form submission as well as on cumulative 
reports available online for clinic review.  If it is determined that certain data points are 
truly missing, a separate code is entered to designate this.  This information, along with the 
rationale for this designation is noted in study logs. 
A more sophisticated series of checks is made after data have been entered.  Computer 
edits are performed across forms to detect and correct instances of entry and transcription 
errors that pass the cross-sectional (intra-form) logical and range checks of the data entry 
screens.  Reports of these errors are provided to study coordinators at the clinics for 
verification based on hard-copy records of forms or clinic information.  When errors are 
discovered in the data, special records are kept to certify when and by whom the error 
was discovered, what steps were taken to ascertain the correct information, and when and 
by whom the database was corrected.  The checks of means, ranges, and standard 
deviations allow for detection, retrospectively, of any relative bias in definitions or 
measurements.  While it may not be possible to rectify these biases (post hoc), these edits 
will at least identify variables for which care must be taken in interpreting analyses. 

---

Section 7 - Data Management (Phase 3 - 12/2010) 
Section 7 - Page 3 
 
7.3.2.
 
On-site Monitoring 
Center visits to the SEARCH study centers are conducted based on a timeline developed 
by the SEARCH Protocol Oversight Committee.  Clinics may be selected for extra center 
visits based on concerns arising from CoC and Protocol Oversight Committee contacts or 
via problems noted on monitoring or QC reports.  For example, if the quality of data from 
a particular study centers is poor, the Protocol Oversight Committee may recommend a 
special center visit for of the Center to assist in identification and correction of data 
quality issues.  Center visits 1) provide a means for continual training, retraining and 
reinforcement of standard study procedures; 2) enhance communication between the 
study centers and the CoC; and 3) detect and document the extent of problems in 
implementing the protocol. 
The data collection/entry performance of the clinical centers and laboratory are evaluated 
during periodic center visits.  These center visits include auditing of data collection/entry 
results received by the CoC for randomly selected participants.  The study centers and 
laboratories are sent a list of the randomly selected participants and requested to have the 
clinic records and participant files for these participants available on the day of the audit.  
The auditor brings from the CoC study data on participants from the central database.  A 
direct comparison of these data with the participant records is performed.  Auditors attempt 
to determine whether discrepancies are due to data entry errors, misinterpretation of study 
protocol, or other reasons.  Data collection/data entry completeness is assessed.  Detailed 
audit results and the preliminary report are discussed on the spot with key staff 
investigators and data managers.  A final report is prepared and issued subsequent to each 
center visit.  An audit visit summary is presented to the Protocol Oversight Committee.  
The review of ten participant records at each center visit should be adequate, unless the 
study center has been targeted for more extensive review due to previous problems. 
7.3.3.
 
Clinic Performance Monitoring 
The SEARCH CoC will continue to provide feedback to clinic staff through the internal 
website with regard to clinic performance in ascertainment of IPS, participant