The
primary goal of the COSP Multisite Research Study data repository is to make
valid and reliable data available, in a timely manner, to qualified individuals.
Obviously, data validity, reliability, and timeliness lie first in the
hands (and computers) of the individual study sites.
However, one responsibility of the data repository is to facilitate the
process of assuring the validity and reliability of shared data.
The framework of the process of collecting and sharing data is summarized
in the following points:
v
All data should be submitted electronically to the repository via
the internet. The actual mechanism for submission will be decided as we proceed
and could be either e-mail or copying database files.
Data can also be submitted using FTP (file transfer protocol).
v
Data will be transmitted in a structured format. The appropriate
format will be determined in the coming months, but may take the form of either
a flat ASCII file or a structured database.
v
Data will be sent in IBM PC compatible files.
A complete data set from the individual sites will be sent each time. No
changes to individual records in the repository database will be made.
v
The schedule for data submission will be determined by the
Multi-site research committee and, perhaps independently, by the Cost study
committee. Data receipt dates will
be maintained and made available to appropriate individuals.
v
Quality check algorithms will be developed during the data
definition process. These will be programmed by the coordinating center in a
database language (possibly Visual Basic for Applications, VBA). These
algorithms and programs will be made freely available to individual sites.
Quality checks will be made on the repository data on a regular basis.
The resulting information will be made available to the participants.
v
Basic cross-site comparative information will be generated on a
regular basis. This information will be made readily available to help identify
potential validity or reliability issues.
A listserv (COSP-DATA) has been
created to encourage communication between all individuals involved in the
transmission of data to the repository and its analysis
11.2
Data Management Operations and
Procedures
The
primary focus of the Data Repository Subcommittee Group’s headed by Matthew
Hile, Ph.D., is moving the data collected at the various sites to a centralized
data storage system as easily and efficiently as possible and making that
information available to authorized users. Other, secondary, considerations
involve developing consensus on the methods used in data coding and entry,
developing a simple data entry application, providing feedback and suggestions
for various automated quality assurance reports, and keeping the projects
apprised of the data collection progress. The committee works in conjunction
with the MIS group at the CC who is primarily responsible for providing the
tools necessary to make the repository a reality.
The committee consists of at least one member from
each site. This individual is responsible for the movement data from local sites
to the central database. Additional members can be found at the Coordinating
Centers and Federal government. The committee has met on one occasion (April 12,
1999) and has a current listserv used for rapid communication between the members.
The data repository is currently under development
and three projects are the current focus of CC MIS staff attention: the database
structure, the data entry system, and the system to move information from sites
to the central server.
v
The database
structure defines how the information is to be coded, entered, stored, and
reported. Each individual response to the common protocol has an entry in this
structure which specifies its name, form, and list of valid responses. In this
way the data can be more easily understood and used by its various users. This
process, however, depends on the development of the CP and the items to be
included. For example, a great deal
of discussion in the CC recently culminated in a decision of how to code missing
information. Based on how the final
results are coded, these decisions can more than double the size of the
repository database and the amount of information that must be entered. An example of this occurred when some decisions made during
the final approval of the CP increased the database size by 70%.
These changes mean that much of the work put into the initial data entry
application, used for the CP pilot study, had to be completely redone.
v
The
second project receiving CC MIS attention is the creation of a “heads down” data
entry system. This system was not planned in the initial budgeting. However,
because of its potential utility, the subcommittee group has decided to invest
in its development. A system of
this type is used by data entry operators who read the information coded by
interviewers off of the assessment tools and enter that information into the
computer. A system was developed and used during the pilot testing
process. However, because of the
70% increase in the number of data elements from the pilot CP to the final CP,
much of the work put into the initial data entry application, used for the CP
pilot study, had to be completely redone.
v
Moving
data from
the various sites to the central repository is a major task.
To make this as simple as possible, we have decided to simply have the
encrypted data files sent to the data repository on a quarterly basis.
This procedure will provide clarity of understanding and a process that
should help us identify and overcome any difficulties encountered.
5400
Arsenal Street