Specifications for a tracking system

Prepared for the ModMan Advisory Group

3rd draft, 2 January 2002
This paper is a proposal for a system to help Collaborative Review Groups (CRGs) keep track
of their editorial processes involving reviews, persons, series of events, and the relations
between them. Based on this paper a search should be performed to identify available
software (commercial or not) which fulfils the specifications. If such software cannot be found,
the Collaboration may wish to consider developing its own software.

Background
In 2000, Sonja Henderson carried out a survey among CRGs about their need for a tracking
system. The results were summarised in a report to the SDG in October 2000. At the
ModMan Advisory Group (MAG) meeting in March 2001 it was agreed to restructure the
report so it could form the basis for a set of system specifications. This paper is the third draft
of these specifications. In this version, the paper has been divided into two sections: part 1 is
a general description of the rationale and concepts while part 2 is a working paper describing
a proposed system implementation, which may later be turned into a full set of specifications.
Note: This paper will be revisited once Nancy Owens has identified best practice as part of
the quality improvement project.

Part 1: General overview

What is a tracking system?
A tracking system offers immediate and easy access to information on the production status
of each individual review, from title registration to review publication and updates. It keeps a
comprehensive history of everything that happens during the preparation of the review. This
information can be made available in customised report formats or as search results.
Prompts, reminders and letters can be generated automatically. The Editorial team will be
able to easily monitor revisions, decisions and overdue activities. The data about reviews and
the people involved in the review process will facilitate performance assessment.

Need
A tracking system is needed for the following reasons:

Most of the 'systems' used by review groups to track reviews from title to update stage
are inadequate. As the number of reviews increases, the existing systems do not provide
the information which staff need to monitor progress and identify deficiencies quickly.

Routine and repetitive tasks in the editorial process need to be assessed for
simplification, consolidation or elimination.

Requirements for higher quality reviews and the increase in the number of reviews must
be counterbalanced by improved productivity, especially with static or declining
resources.

Accurate and timely data collection will help staff to manage the preparation of reviews,
while simultaneously providing detailed data to support budget requests.

Objectives

To ensure timely and efficient management of the editorial process.

To provide information to the editorial team to support the operations and decision
making within the Group.

To make more time available for the application of standards and quality control by
reducing the time spent on routine editorial operations.

Benefits for editorial bases

a) Operational/needs analysis of procedures

Data collection on current operations of the Review Groups

Investigation and analysis of the need for functional activities by individual Review
Groups
Determination of minimum procedures in the editorial process
Identification of reports required
Identification of the 'ideal' generic system (i.e. the selected system with the most
desirable features and the fewest undesirable features)

b) Improved operating efficiency

Establishment and surveillance of policies and procedures, e.g.

feedback/performance analysis
Ensure that tasks are performed reliably and economically
Automatic generation of reports and processing of enquiries

Reduction of staff time in preparing management information

Reduced need for error detection and correction

Requirements

Simple to use
Minimise data entry and processing errors
Easily identify the status of a review in the editorial process
Easily identify the status of a person involved in the editorial process
Easily identify tasks which need to be carried out
Generate routine prompting when editorial action is required
Assist in carrying out repetitive tasks

Part 2: Proposed system - working paper

Choice of technology
The system proposed in this paper is assumed to be electronic and database based although
this is not the only option. A file based system where the record for each review or person is
kept in a separate file (or on paper) could also be used. The primary advantage of a database
system is that it allows similar information to be assembled in tables where it can easily be
searched, edited, or deleted. In addition, information can be linked and re-used, so that each
record (e.g. a persons record) only exists once. It also opens up the possibility of linking to
other elements of the Cochrane Information Management System (e.g. CRG modules). In
contrast, the advantage of a file based system is that the information is stored independently
so that deleting one file will not affect the other.
Other choices about technology will also have to be made including on which platforms the
system should run (especially whether it should be web based or a stand-alone application)
and how the system and data will be protected from unauthorised access (e.g. multiple levels
of passwords).

Data elements
While going through the results of the survey for a tracking system (see background), three
main data elements were identified: reviews (including titles, protocols, full reviews, and
various intermediate stages of reviews), persons (including reviewers, coordinators, editors,
advisors etc.), and events (e.g. protocol approved). In general, the system should be able to
re-use as much of this information from existing systems (e.g. RevMan, ModMan) as
possible.

Reviews
Reviews are the central data elements in the system. The term reviews is used in this paper
as a generic term for the documents which the system track, i.e. titles, protocols, full reviews,
and various intermediate stages of reviews. It is not perceived that the system should store
the full reviews (text, tables, etc.) only the information needed for tracking. The system
would need a unique identifier for each review to link to from other records. The unique ID
assigned by RevMan is one possibility - another is the user defined four character review
number. The latter is more comprehensible for a human, but not guaranteed to be unique.
The system could keep copies of reviews records for each major update cycle or major
version of the review as a log, but ideally the same information would also be available by
searching though past events.

Persons
Each person should have one record only and all reference to this person from other records
should be done through linking. For linking, a unique ID number must be assigned to each
person. If a review record, for example, needs to refer to the contact editor of the review, it
only needs to store the unique ID instead of the persons full name. The unique ID for persons
could be based on the four character code in RevMan, or a new system could be introduced.

Events
Events form the link between persons and reviews. If information such as date review sent to
contact editor was stored together with the review record, the system would only capture the
current state of the review not its past and future history. This is because the date would be
overwritten for each new editorial cycle. This could to a certain extent be improved by creating
a new review rec ord for each update cycle, but still this would not be ideal. If, on the other
hand, the date review sent to contact editor was stored with the record for the editor, you
would only be able to tell the last time he had received a review. In order to keep the entire
history of reviews and persons, a system which stores event records separately from review
and person records is needed.

Each event would be linked to one or more reviews and a number of persons involved in the
events, e.g. the reviewer involved in updating a review or the editor involved in peer reviewing
it. Following the links the other way around each review would be linked to a list of events,
both pending and completed and each person would be linked to a list of events in which the
person is involved.
A typical event consists of three parts: 1) Sending something to person(s). 2) Waiting for
feedback on a given date. 3) Receiving feedback. This is reflected by each event record
having three date fields, 'date sent', 'date expected' and 'date completed'. Therefore, there is
no need for events for both 'Title sent for comments' and 'Comment on title received', only the
former. For some events, e.g. Proposal for title received at editorial base, you only have to
use the date completed.
The default events (see appendix 1) should be based on 'best practice'. It should, however,
be possible for CRGs to modify those, for instance if they use different terms than the default,
and even to add their own events. It should be possible to order events and to specify that
one (group of) event(s) must follow another, e.g. review received always follows protocol
received.
It should be possible to specify the action taken when the expected date for an event has
passed without the date completed being filled in. This could be in the form of warnings that
pop-up or automatic emails, or the system could do nothing at all until the user runs a report.
This could be specified individually for each event or each type of events.

Processes
Preferably, it should be possible to group events, e.g. to group the events involved in the
review update cycle. A group of events could be called a process. In the system, a process
would work like a checklist with a certain workflow. When a new process is started, blank
records for the events associated with the process would automatically be created, where the
dates and other information would later be filled in. The default processes should be based on
'best practice'. It should be possible for CRGs to modify default processes if they use
workflows other than the default.

Appendix 1: Default events defined in the system

The default events and processes defined in the system should be based on best practice for
review groups.
Note: This section will be revisited once Nancy Owens has identified best practice as part of
the quality improvement project.
Note: It may be desirable in a later version of this document to divide the events in the table
into processes. This is already done to some extent by sorting the events by status, but in a
real system, a more granular division should be used.
Guide to table:
Status: T=title, P=protocol, R=full review, U=update
For each event it is marked with an 'X' which of the three dates (Date sent, Date expected,
Date completed) that are used for the event.

Reviewer contacted and asked for protocol

P
P
P
P
P
P
P
P
P
P

Protocol received
Protocol sent to contact editor
Protocol copy edited before sending to referee(s)
Protocol sent to statistician
Protocol sent consumer(s)
Protocol sent to TSC
Protocol sent to referee(s)
Edited version of protocol sent to reviewer
Feedback sent to reviewer expecting revised version
Revised version of protocol in response to feedback sent
for approval
Protocol sent for approval
Protocol sent for external copy editing
Protocol approved and returned to reviewer with date
review expected
Protocol included in CRG module
Protocol withdrawn
Protocol deleted

P
P
P
P
P
P

X
X
X

Date completed

Proposal for title received at editorial base

Internal check for duplication of titles
Check for duplication of title on Review Titles Manager
Title sent for comments
Comments from editor sent to reviewer
Title discussed with other group
Title approved
Title entered into Review Titles Manager
Contact editor assigned
Title modified

Date expected

T
T
T
T
T
T
T
T
T
T

Date sent

Status
Name

X
X
X
X

X
X
X
X
X
X
X
X
X
X

X
X
X
X
X
X

X
X
X
X
X
X
X

X
X

X
X

X
X

X
X

X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X

Notes

Reviewer
contacted?
Combine with
next?

Feedback?

P
P

Search results sent to reviewer

Post publication comment(s)received

P
R

Post publication comment(s) responded to

Reviewer contacted and asked for review

R
R
R
R
R
R
R
R
R
R
R

R
R
R
R
R

Review received
Review sent to contact editor
Review copy edited before sending to referee(s)
Review sent to statistician
Review sent to consumer(s)
Review sent to TSC
Review sent to referee (s)
Review sent to consumer network for draft synopsis
Edited version of review sent to reviewer
Feedback sent to reviewer expecting revised version
Revised version of review in response to feedback sent
for approval
Review sent for external copy editing
Review approved and returned to reviewer with date
update expected
Review included in CRG module
Review withdrawn
Review deleted
Permission for publication form to reviewer
Post publication comment(s) received

R
R
U

Post publication comment(s) responded to

Date search for update completed
Reviewer contacted and asked for update

Updated review received

U
U
U
U
U
U

U
U

Updated review sent to contact editor

Updated review copy edited before sending to referee(s)
Updated review sent to other people
Edited version of updated reviews sent to reviewer
Feedback sent to reviewer expecting revised version
Revised version of updated review in response to
feedback sent for approval
Updated review sent for external copy editing
Updated review approved and returned to reviewer with
date next update expected
Update review included in CRG Module
Permission for publication forms to reviewer

Post publication comment(s) received

U
U
U
U

Post publication comment(s) responded to

Date search for update completed
Updated review withdrawn
Updated review deleted

R
R

U
U

X
X
X
X

X
X
X
X
X
X
X
X

X
X

X
X

X
Combine with
next?

X
X
X
X
X
X
X

X
X
X
X
X
X
X
X
X
X
X

X
X
X
X
X

Store title(s) of
comment(s)

X
X

X
X
X

X
X
X

X
X

X
X

X
X
X
X
X
X

Combine with
next?
Store
substantive /
nonsubstantive

Feedback?

X
X
X
X
X

Feedback?

X
X

Store title(s) of
comment(s)

For substantive
updates
Store title(s) of
comment(s)

X
X
X
X

Appendix 2: Events related to people

It has not been decided whether these events related to people should be in the tracking
system or they should be referred to the group's contact database.

New reviewers pack sent to reviewer

Other guidelines sent to reviewer
RevMan sent to reviewer
Meetings/workshops/training attended

Appendix 3: Proposed overall database structure

Processes
Reviews

Events

People

Overall database structure. The arrows on the figure are "one to many" connections, i.e.
reviews and people can be linked to multiple events, and processes consist of one or more
events.

Information stored for each review

ID number (computer generated, e.g. from RevMan)
Current status of the document (i.e. title, protocol, review, update)
Review title
Review number (RGC allocated number - 4 characters)
Version publication number, e.g. P2/R6 meaning second protocol version, sixth review
version published (computer generated). Only used if copies of review records are kept for
each update cycle)
Contact reviewer (link to persons table)
Co-reviewers (links to persons table)
Contact editor (link to persons table)
Internal referees (links to persons table)
External referees (links to persons table)
Keywords (CRG defined)
Notes
Date record created
Date record last modified

Information stored for each person

ID number (c omputer generated, e.g. from RevMan)
Name
Address
Email
Role(s)
Date of first contact
Date of last contact
Date no longer active
Keywords (CRG defined)
Notes
Date record created
Date record last modified

Information stored for each event

ID number (computer generated - user may never see this)
Type of event
Description
Date sent
Date expected
History of dates expected
Date completed
Review(s) involved (links to reviews table)
Person(s) involved (links to persons table)

How to react when passed date expected (warning, email, etc. - users choice)
Links to documents associated with event (reports etc.)
Links to events which must be completed before this event
Status (pending, completed, cancelled)
Feedback from event for each person involved in event
Feedback approved? (for each person involved in event)
Additional data needed for each type of event. See table of default events.
Notes
Date record created
Date record last modified

Information stored for each process

ID number (e.g. computer generated - user may never see this)
Type of process
Description
Ordered list of events involved in process

Appendix 4: Examples of questions which the system should be

able to answer
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.

Number of reviews assigned to a particular person?

A list of the titles of reviews assigned to an individual person.
Number of reviews assigned to a particular person for which we have not had any
feedback?
A list of the titles of reviews assigned to a particular person for which we have not had a
reply.
Number of reviews assigned to a particular person for which we have not received the
feedback by our deadline?
A list of the titles of reviews assigned to a particular person for which we have not
received the feedback by our deadline?
Number of reviews in the editorial process.
Titles of reviews in the editorial process.
Number of titles registered reviews with expired date review expected.
Titles of registered reviews with expired date review expected.
Number of updated reviews in last 12 months/calendar year.
Titles of updated reviews in last 12 months/calendar year.
Number of reviews not updated within last 2 years.
Titles of reviews not updated within last 2 years.
Date last review assigned to a particular person.
Number of reviews assigned to a particular individual in any one calendar year.
Number of reviews for which we have added a comcrit.
A list of the titles of reviews for which we have added a comcrit.
Number of reviews for which we have added a comment but have not had a response
from the reviewer.
List the titles of reviews for which we have added a comment but have not had a
response from the reviewer.
Number of reviews for which the reviewer has received all feedback but not returned final
draft for final approval.
List the titles of reviews for which the reviewer has received all feedback but not returned
final draft for final approval.