THE COCHRANE COLLABORATION VISITING FELLOWSHIP
Progress report: 17 October 2004 –19
November 2004
Karishma Busgeeth
Trials Search Coordinator
HIV/AIDS Cochrane
Review Group
Table of Contents
Title Page
- Acknowledgements 3
- The Problem Area 4
- Aims of the fellowship 4
- Progress made towards overcoming the barriers to solving the
problem 5
- Outline of plan for implementing the learned strategies 7
- Additional opportunities and achievements, not specific to the
programme 8
- What was not achieved? 8
Appendix 1 9
Appendix 2 10
Appendix 3 11
Appendix 4 12
Appendix 5 13
1.
Acknowledgements
I acknowledge the assistance, help and support of
Nandi Siegfried throughout the fellowship.
Many thanks for the
conversations that clarified my thinking and for the thoughtful and creative
comments.
I
am particularly grateful to Mark Fenton of the Schizophrenia Group for his advice,
guidance and suggestions regarding the HIV/AIDS trials register.
Special
thanks go to Jini Hetherington for ensuring the success of the fellowship.
Two colleagues graciously made time for me,
helped me to consolidate my information
management skills and provided me with the necessary TSC skills. In this regard, I am indebted to Anne
Eisinga and Carol Lefebvre, Information Specialists at the UKCC.
Thank you to the South African Cochrane Centre and
the Cochrane Collaboration Secretariat for their financial support for the
fellowship. Thanks also to the
Co-convenors of the Quality Advisory Group for selecting me as the recipient of the Cochrane Collaboration
Exchange Fellowship.
Last but not least, thank you to the UKCC staff for
being so nice and accommodating.
2.
The Problem Area
The Cochrane HIV/AIDS Review Group (CRG) had not
had a dedicated Trials Search Coordinator (TSC) since 2001. As a result, the register had not been well
maintained. The register was
reference-based and was maintained in ProCite, a bibliographic software
package. The HIV/AIDS CRG believes that
creating a study-based, rather than a reference-based register is more
appropriate, not least because of the practice of multiple publications from
single studies. The UK Cochrane Centre, Schizophrenia Review
Group, others in The Cochrane Collaboration and Update Software have developed
a software tool, MeerKat (MS Access), specifically to support study-based
registers. After visiting the editorial
bases of the Dementia and Cognitive Improvement Group and the Schizophrenia
Group to see how study-based registers were maintained in ProCite and MeerKat
respectively, Karishma Busgeeth (KB) and Nandi Siegfried (NS), the Deputy
Co-ordinating Editor of the HIV/AIDS Review Group, decided to adopt MeerKat, in
particular because of Karishma’s extensive knowledge and experience of MS
Access.
3.
Aims of the fellowship
The aims of the fellowship were to
1. Transfer
all current HIV/AIDS records in CENTRAL (The Cochrane CENTRAL Register of Controlled
Trials) to MeerKat;
2. Acquire
skills to maintain and update the HIV/AIDS trials register with studies
retrieved from searches conducted on PUBMED, EMBASE, AIDSearch and other
databases identified as relevant;
3. Finalise
the EMBASE and AIDSearch comprehensive HIV/AIDS search strategy;
4. Acquire
skills to effectively and efficiently prepare the register for regular
submission to CENTRAL;
5. Code
studies appropriately using a comprehensive coding sheet;
6. Procure
full-text articles for the trials in the register.
4. Progress made
towards overcoming the barriers to solving the problem
(See
Appendix 1)
4.1 Transfer
of all current HIV/AIDS records in CENTRAL to MeerKat
The
Cochrane highly sensitive search strategy was used to download 5952 HIV/AIDS
records from CENTRAL and using the CLIB configuration file, the records were
transferred from CENTRAL to ProCite.
The duplicates were then configured by comparing the fields: title,
author, date of publication and location in work. (See Appendix 2)
The
deduplication process revealed 338 duplicates, 8 triplicates and one
quadruplicate. Since HIV only became an
entity in the early 1980s, NS and KB decided to delete all records before
1980. The number of records left in
ProCite was then 5550.
After
the deduplication process, Mark Fenton (MF), TSC of the Schizophrenia Group,
assisted KB in transferring the records from ProCite to MeerKat. This process involved configuring the
bibliography in ProCite, printing the bibliography in ProCite and finally
defining the references in MeerKat.
(See Appendix 3)
A
screenshot of the HIV/AIDS trials register (SR–HIV) in MeerKat is provided in
Appendix 4. For ease of information
transfer and the awesome responsibility of efficiently coding each and every
reference in the register, it was decided that initially 2004 records would be
transferred to Meerkat followed by 2003 records then 2002 and so on.
4.2 Maintaining
and updating of the HIV/AIDS trials register with studies retrieved from
searches conducted on PUBMED, EMBASE and other databases identified as relevant
MF
assisted KB in setting bibliography and configuring files for both PubMed and
EMBASE. These configuration files would
facilitate the transfer of records from PubMed and EMBASE to ProCite. MK also demonstrated the different
functionalities of MeerKat in terms of maintaining, updating and searching the
SR-HIV.
NS
and KB decided on the studies to be included or excluded in the SR-HIV. All
randomized and controlled clinical trials evaluating the efficacy or
effectiveness of prevention or treatment of HIV/AIDS will be included. This
will include those studies which assess sexual behaviour relating to condom use
(non-HIV specific) as no other register is currently capturing these. Studies
specifically excluded are non-HIV Sexually Transmitted Infections (e.g. Herpes)
trials and Phase 1 (safety) trials.
4.3 Finalising
EMBASE and AIDSearch comprehensive HIV/AIDS search strategy
Anne
Eisinga (AE), Information Specialist at the UKCC, provided her assistance in
translating the comprehensive HIV/AIDS search strategy via OVID to
EMBASE.COM. At the South African
Cochrane Centre (SACC), EMBASE is accessed via EMBASE.COM and not via
OVID. Both NS and AE helped to develop
and finalise the comprehensive HIV/AIDS search strategy for AIDSearch.
4.4 Effective
and efficient preparation of the register for regular submission to CENTRAL
Carol
Lefebvre (CL), Information Specialist at the UKCC, and MF provided much insight
into the submission of the SR–HIV to CENTRAL.
During the fellowship, several meetings were arranged between KB and
CL. The points discussed were:
·
TSC role and duties in the Cochrane
Collaboration
·
Identifying RCT/CCT
·
‘Included studies’ in the register
·
Reviewer’s support
·
Copyright legislation
4.5 Coding
studies appropriately using a comprehensive coding sheet
NS
and KB decided on the different domains within the SR-HIV. The domains are as follows:
·
Design: RCT, CCT, Systematic review
·
Intervention: Treatment and Prevention
(these are further expanded into specific interventions)
·
Methods: Allocation concealment,
Blinding, Loss to follow up
·
Outcomes: Morbidity, Mortality
·
ISRCTN
·
Geographical Coordinates of RCT
location (this will allow for mapping of trials)
KB
spent much time during the fellowship coding the references as RCT/CCT (See Appendix
5) for 2004 and 2003. For the year 2004, the number of records coded was 62
while 400 records were coded for 2003.
A quality control carried out by NS on the first 60 records and
thereafter of all records coded as ‘unclear’ or ‘delete’ revealed a 95%
agreement.
4.6
Procurement
of full-text articles for the trials in the register.
NS assisted KB in the procurement of
full-text articles in pdf format from various databases accessed via the UKCC
and the University of Oxford. The total
number of pdf articles obtained was 36.
5. Outline of plan for implementing the learned
strategies
The table below shows an ‘idealized’ monthly schedule
structured for implementing the learned strategies.
|
Time |
Tasks |
|
|
|
1 week |
|
|
|
|
|
|
3 days |
|
|
2 days |
|
Specific aims for the next 6 months are to
1. Code ‘study design’ and possibly
‘intervention’ in SR-HIV for all studies from all years
2. Procure full-text articles for the register
from other sources (e.g. Medical Research Council library, South African
Universities libraries etc)
3. Code the location (geographical coordinates)
of the RCT/CCT for 2003
4. Design and develop the GIS system
5.
Write an article for the CCInfo in March-April
6. Write a paper on the development of the register for publication purposes in Mid-April
7. Prepare an abstract of the development of the register to present at the Melbourne Colloquium
6. Additional
opportunities and achievements, not specific to the programme
·
Attended weekly Clinical Trials Service Unit (CTSU) seminars at the
Nuffield Department of Medicine, University of Oxford. Attended a seminar presented by Prof Harald
Jaffe, new head of the Oxford University Department of Public Health and
Primary Health Care, on the ”Global Challenge of HIV/AIDS”.
·
Presented the HIV/AIDS trials register at the monthly UKCC seminar.
·
Hand searched for RCT/CCT from French records in African Index Medicus
·
Hand searched for RCT/CCT from French Medical Journal ‘Medecine Afrique
Noire’ for the year 1990-1999.
7. What was not
achieved?
·
The pharmaceutical contacts are yet to be established. Due to time constraints, it was decided that
this task would be followed up in South Africa.
·
The comprehensive HIV/AIDS search strategy for LILACS requires further
testing and translating. AE is
currently busy with the testing phase.
·
The content and timescale of AIDSearch and GATEWAY are yet to be
finalized.
Appendix 1
Appendix 1 displays the five-weeks fellowship
programme at the UKCC.
|
|
MONDAY
18 |
TUESDAY
19 |
WEDNESDAY
20 |
THURSDAY
21 |
FRIDAY
22 |
|
Morning |
UKCC
orientation/planning |
9AM
Laptop 11AM Carol |
Dedupl
Procite 10AM Mark Fenton |
Transfer
into Meerkat |
Tfer into
Meerkat 11AM Carol |
|
|
|
1PM GC
seminar |
|
12.30
DPH seminar |
|
|
Afternoon |
2PM: AIM
French translation |
Run HIV
strat in Central – Procite |
Mark
Fenton Dedupl Procite |
Develop
register domains |
2PM –
troubleshoot NS (domains) |
|
|
MONDAY
25 |
TUESDAY
26 |
WEDNESDAY
27 |
THURSDAY
28 |
FRIDAY
29 |
|
Morning |
Develop
Embase strategy |
Develop
Embase strategy |
Run
EMBASE 10AM Mark Fenton |
MARK
FENTON |
Register
cleaning |
|
|
|
1PM GC
seminar |
|
|
|
|
Afternoon |
Register
cleaning |
Develop
Embase strategy |
Mark
Fenton Run EMBASE |
|
2PM –
troubleshoot NS |
|
|
MONDAY
1 |
TUESDAY
2 |
WEDNESDAY
3 |
THURSDAY
4 |
FRIDAY
5 |
|
Morning |
Prepare
UKCC talk |
Register
cleaning |
AIDSearch
|
AIDSearch |
Register
cleaning |
|
|
|
1PM GC
seminar |
|
1PM:
UKCC-KB&NS |
|
|
Afternoon |
Register
cleaning |
NS – prepare
UKCC talk |
NS –
mapping and African Jx project |
NS –
mapping and African Jx project |
2PM –
troubleshoot NS |
|
|
MONDAY
8 |
TUESDAY
9 |
WEDNESDAY
10 |
THURSDAY
11 |
FRIDAY
12 |
|
Morning |
Identify
relevant pharmaceuticals |
Identify
relevant pharmaceuticals |
Possible
LSHTM |
LILACS |
Register
cleaning |
|
|
|
1PM GC
seminar |
|
|
|
|
Afternoon |
Register
cleaning |
|
Register
cleaning |
MARK
FENTON? |
2PM –
troubleshoot NS |
|
|
MONDAY
15 |
TUESDAY
16 |
WEDNESDAY
17 |
THURSDAY
18 |
FRIDAY
19 |
|
Morning |
10AM
Mark? Assess progress |
Register
cleaning |
10AM NS
Planning |
Trip
report and feedback QAG |
NS –
feedback report |
|
|
|
1PM GC
seminar |
|
12.30
DPH seminar |
|
|
Afternoon |
Register
cleaning |
Planning
for 6/12 |
Register
cleaning |
Trip
report and feedback QAG |
Trip
report and feedback QAG |
Appendix 2
Appendix 3
Appendix 4
Appendix 5
THE COCHRANE COLLABORATION VISITING FELLOWSHIP
Progress report: 19 December 2004 –
19 May 2005
Karishma Busgeeth
Trials Search Coordinator
HIV/AIDS Cochrane
Review Group
Table of Contents
Title Page
- Terms of reference 3
- Aims for the
past 6 months 3
- Implementation of the learned strategies 4
- Improvement in the problem 5
- The GIS System 6
- Publications 6
1.
Terms of reference
The fellow shall:
a.
submit a second written report to the Co-Convenors of the Quality
Advisory Group within six months of her return to her home entity
b. outline the
implementation of the learned strategies
c. demonstrate
improvement in the problem
2.
Aims for the past six months
The aims were to
1. Code ‘study design’ and possibly
‘intervention’ in SR-HIV[1]
for all studies from all years
2. Procure full-text articles for the register
from other sources (e.g. Medical Research Council library, South African
Universities libraries etc)
3. Code the location (geographical coordinates)
of the RCT/CCT for 2003
4. Design and develop the GIS[2]
system
5. Write an article for the CCInfo in
March-April
6. Write a paper on the development of the register for publication purposes in Mid-April
7. Prepare an abstract of the development of the register to present at the Melbourne Colloquium
3. Implementation of
the learned strategies
3.1 Code ‘study design’ and possibly
‘intervention’ in SR-HIV for all studies from all years
The
following publications have been coded as RCT/CCT or No AB[3]
or Unclear or Delete.
|
YEAR |
Total # of references |
|
1980-1989 |
283 |
|
2003 |
401 |
|
2004 |
304 |
265
PubMed records relating to publications in Oct 2004 - Feb 2005 were downloaded
into ProCite and then imported into the SR-HIV. As of May 2005, the register contains 5819 records published
between 1980-2005, of which 988 have already been coded.
3.2 Procure full-text articles for the register
from other sources
Procurement of full-text articles is ongoing. Nandi Siegfried (NS), based at the UKCC, provided her assistance in the procurement of African trials. A list of 231 references relating to publications in 2003 has been sent to the Medical Research Council (MRC) library.
4. Improvement in
the problem
4.1 Submission to CENTRAL
The first submission to CENTRAL was made in March 2005. Out of the 532 records that were coded, 300 of them were eligible for CENTRAL submission. The SR Comment Report from Elena Glatman stated the following:
1.
Author field:
Some
records have commas as separators between author names within the database;
they were replaced with “//”. Some records contain “and” before the last author
name. It’s OK to have “and” before the study name, but according to the CMP,
chapter 6, Appendix A, the publisher requires author names only, and no “ands”.
2. Issue field:
Some
records are missing “()” around part information in 5 records. Please see
Appendix 3, chapter 3 of the CMP.
4.2 Update of the USCC Master List of Journals
The
Master list of journals searched has been updated. The following lists the
different journals that have been recently hand searched by KB and NS. Hand
searching of non-indexed African medical journals will identify which journals
give a high yield of RCTs.
|
Journal Title |
Year
|
|
Afrique Medicale
(KB) |
1987-1994 |
|
Medecine
D'Afrique Noire (KB) |
1991-1999 & 2001-2003 |
|
African Index
Medicus (NS & KB) |
1993-2000 |
|
Nigerian Journal
of Paediatrics (NS) |
1985-2003 |
|
The Nigerian
Postgraduate Medical Journal (NS) |
1994-1996 |
|
African Journal
of Medical Practice (NS) |
1994-2001 |
|
The Nigerian
Medical Practioner (NS) |
1986-1991 |
4.3 Timeline and Inclusion content of NLM
GATEWAY and AIDSearch
KB has developed a timeline and inclusion content of NLM GATEWAY and AIDSearch. The purpose of this exercise was to determine and compare the content and coverage of AIDSearch and NLM Gateway for the retrieval of relevant HIV/AIDS conference abstracts. The content comparison between AIDSearch and NLM Gateway revealed that NLM Gateway yields the highest number of relevant HIV/AIDS conferences. The coverage comparison showed that AIDSearch is more consistent from year to year in the inclusion of abstracts. Conference abstracts proceedings are available from mid 1980s in AIDSearch and only from the late 1980s in NLM Gateway. Abstracts from the two most important HIV/AIDS conferences, the International AIDS Conference and the Conference on Retroviruses and Opportunistic Infections, are available in both databases and include year 2004 records.
The
findings suggested that a combination of AIDSearch and NLM Gateway searching is
required to retrieve all relevant conference proceeding abstracts.
5. The GIS System
5.1 Design and develop the GIS system
The
purpose of the GIS system is to demonstrate how the adoption of newly emerging
GIS technology facilitates the visualization and retrieval of HIV/AIDS
trials-related information. Mapping of the trials requires the capture of
geographical coordinates and the design and development of a geodatabase using
the mapping software, ArcGIS. The
geocoding functionality within ArcGIS supports the geographical location of the
trials. KB is currently working on the coordinate system.
5.2 Code the location (geographical coordinates)
of the RCT/CCT for 2003
401 records relating to publications in 2003 were downloaded from CENTRAL. Out of the 401 records, 276 of them are RCT/CCT. Coding the location of RCT/CCT requires the geographical coordinates of the country where the trial has taken place. Since, the study setting is not mentioned in most of the abstracts, full-text articles are being procured. To date, the geographical coordinates of 45 RCT/CCT have been established. The citations of the remaining 231 RCT/CCT have been submitted to the MRC library for procurement.
6. Publications
· The following abstracts have been submitted to the Melbourne Colloquium:
o Improving our quality by developing the HIV/AIDS Specialized Register – experiences and lessons from the inaugural Cochrane Collaboration Visiting Fellowship
o Mapping trials - the HIV/AIDS trials register leads the way to GIS Mapping
o Optimal retrieval of HIV/AIDS conference proceeding abstracts – a comparison of databases
·
KB will be
giving a poster presentation titled “The Cochrane Collaboration HIV/AIDS Trials
Register” at the South African AIDS Conference in Durban, South Africa in June
2005.
·
KB will be
writing a brief report for Cochrane News (August 2005) about her time as
Cochrane visiting fellow.
·
KB is
currently busy with a paper on the development of the SR-HIV.