Systematic Reviews

1
Systematic Reviews
Effective Literature Searching

• Brenda Thomas
• Information Specialist
• Cochrane Stroke Group

2
Stroke Review Group
Comprehensive and versatile register of
randomised controlled trials (RCTs) and
controlled clinical trials (CCTs)
Prepare, keep up to date and disseminate high
quality systematic reviews in all aspects of
stroke healthcare
3
Role of the Trials Search Co-ordinator
(TSC)/Information Specialist

• Deliver a comprehensive trials identification
  strategy
• Assess and code relevant stroke trials
• Develop and maintain quality and growth of the
  register of randomised trials and controlled
  clinical trials
• Support review authors
• Contribute to the review editorial process

Considerable time and resources have been
committed to maximising trial identification
4
Specialised Register of Stroke Trials As many
potentially relevant Randomised Controlled Trials
(RCTs) and Controlled Clinical Trials (CCTs) as
possible

• Published trials in any language from any year
• Unpublished trials
• Ongoing trials
• Planned trials

All the available evidence
5
The Cochrane Stroke GroupTrials Register
How we identified 13000 stroke trial reports

• Stroke Group Trials Register is one of the most
  advanced in the Collaboration
• All published trial reports included in
• the Cochrane Central Register of Controlled
  Trials in
• The Cochrane Library

6
Methods

• Identify sources of relevant trials
• Develop appropriate search strategies
• Screen bibliographic details for potential
  reports of trials
• Obtain full copies and apply selection criteria
• Arrange translation if required

7
A systematic approach to trial identification

• Electronic databases
• Handsearching
• Personal communication

Multiple overlapping search strategies to
minimise bias
Supported by the considerable resources of The
University of Edinburgh and NHS Scotland e-library
8
Electronic resources

• Major medical bibliographic databases
• Eg MEDLINE and EMBASE
• Topic specific databases
• Non-English language databases
• International trials and research registers
  (ongoing trials)
• The Internet

Regular searches of 27 bibliographic
databases Approx 25,000 references screened
annually
9
Topic Specific Databases

• BIOSIS Previews (Biology and pharmacology)
• AMED (Allied and Complementary Medicine)
• CINAHL (Nursing and Allied Health)
• PsycINFO (Psychology)
• Digital Dissertations (Theses)
• MANTIS (Manual, alternative and natural therapy)
• International Pharmaceutical Abstracts
• Occupational Therapy Journal of Research Index
• Applied Social Sciences Index and Abstracts

10
Clinical Trials Registers

• bibliographic
• Cochrane Central Register of Controlled Trials
  (CENTRAL) (The Cochrane Library)
• The most comprehensive source of reports of
  controlled trials
• Chinese Clinical Trials Register
• The Physiotherapy Evidence Database (PEDro)
• Database of Randomised Controlled Trials in
  Hyperbaric Medicine (DORCTHIM)
• ongoing
• Stroke Trials Directory http//www.strokecenter.or
  g/trials/
• Current Controlled Trials http//www.controlled-tr
  ials.com/
• GlaxoSmithKline Clinical Trials Register
• RehabTrials.org

11
Electronic journals

• Contents pages and abstracts on-line
• Current awareness alerts
• Selected free articles or issues
• Full text (free or subscribed)
• Download pdfs

12
Personal Communication and The Internet

• Trialists and colleagues
• Drug Companies
• Equipment manufacturers
• Websites e.g. company press releases, research
  centre bulletins, research funders
• News alerts eg MEDSCAPE http//www.medscape.com/ne
  urology
• News media (current awareness)
• Conference abstracts on-line

Essential sources of unpublished and ongoing
trials
13
Reference lists

• Trial reports from publications not indexed in
  bibliographic databases
• Early trial reports
• References from grey literature
• Conference abstracts
• books
• theses

14
Review-specific searches

• Electronic searches
• Geographic, eg Chinese Biomedical Literature
• Subject specific, Engineering databases
• Handsearching
• specialist journals and conference proceedings
• Unpublished studies
• Contacting pharmaceutical companies or equipment
  manufacturers

15
Key advantage of searching bibliographic databases

• Can be searched electronically for
• words in the title and abstract
• Standardized indexing terms or controlled
  vocabulary assigned to each record based on the
  full report.

16
Limitations and challenges of electronic
searching for published studies

• Selected journals and journal years indexed
• Conference proceedings and books in selected
  databases only
• A complex tailor-made strategy required for each
  database
• Poor indexing often limits retrieval
• Titles and abstracts (when available) may be
  insufficiently detailed for text word searching
• Free-text searching for full-text electronic
  journals!
• High sensitivity searches result in low precision

17
Handsearching

• Advantages
• All reports of trials in journal
• Articles, editorials, letters, abstracts, news
  items
• Journals, books and conference proceedings not in
  electronic databases
• Recent issues
• Limitations
• Time consuming and expensive
• Large numbers of healthcare journals
• Access to journals
• Non-English language journals
• Subject to human error

18
Is handsearching worthwhile?

• Complete retrospective search of 32 years of
  journal STROKE
• Compared with MEDLINE search of STROKE for same
  publication years
• MEDLINE search identified 44 of 803 stroke trial
  reports identified by handsearching.
• Conference abstracts and letters etc not indexed
  by MEDLINE
• Many of the trials published as conference
  abstracts will not come to full publication

19
Stroke Group general strategymodified for each
database
Highly sensitive TRIALS strategy
STROKE specific strategy
?
Screened for stroke trials
20
Review-specificintervention-based strategy
Highly sensitive Cochrane TRIALS strategy
STROKE specific strategy
?
?
?
Review specific INTERVENTION-based strategy
Screened for additional relevant trials
21
New developments that make life easier

• Increasing availability of electronic journals
• 77 of relevant MEDLINE trial reports in 2006
• Free open access journals and conference
  proceedings
• More detailed reporting of study methods
• Increasing numbers of ongoing trial registers
• Including drug company (eg GlaxoSmithKline,
  Novartis, Roche)
• International trial registration
• WHO International Clinical Trials Registry
  Platform
• Multiple Registers (ISRCTN, NCT, ANZCTR)
• Universal Trial Reference Number
• Easier to track trial reports and link to studies

22
Looking for Research Studies?

• How to search like a professional!

23
Increasing numbers of resources available

• Need a focused approach using high quality
  information resources
• Increased use of electronic sources
• Efficient search strategies

24
Where do we start?

• Search question?
• Purpose of search?
• How much information do we need?
• Where do we look?
• How much time do we have?
• How much money do we have (if any!)?

25
Sources of Bias

• Publication bias
• negative results are often not published
• Level of publication bias
• abstracts and letters may more often report
  negative results
• Reference/citation bias
• Cited work is more likely to be positive
• Electronic database bias?
• Need to search multiple databases
• Language bias

Aim to minimize bias
26
Bibliographic DatabasesMEDLINE

• Major general medical database
• Indexes articles from gt5000 journals (2100 not
  EMBASE)
• 37 languages
• gt16 million citations from 1950
• All articles indexed using Medical Subject
  Headings (MeSH) to describe each article

Available on Ovid SP through University or NHS
e-Library
27
Bibliographic DatabasesEMBASE

• Major general medical database
• Indexes articles from 4800 journals (1700 not
  MEDLINE)
• 30 languages
• gt11 million citations from 1974
• All articles indexed using Emtree Headings to
  describe each article (Different from MeSH)
• Pharmacological studies indexed in greater depth
  than MEDLINE

Available on Ovid SP through University or NHS
e-Library
Both databases required for a comprehensive search
28
MEDLINE Recordsearchable fields

• Authors
• Title
• Journal title
• MeSH Subject Headings
• Abstract
• Publication Type
• Language

29
MEDLINEDeveloping a Search Strategy

• Use a combination of MeSH terms and free text
  for
• Disease under investigation
• Intervention
• Study design (methodological filters)
• Use Publication Type field to focus the type of
  research (eg randomised controlled trial)
• Apply other logical limiters (eg age group,
  human, time period)

Ask your local medical librarian for help!
30
Stem Cell Transplantation for ischemic
stroke MEDLINE (Ovid)
1. cerebrovascular disorders/ or basal ganglia
cerebrovascular disease/ or exp brain ischemia/
or carotid artery diseases/ or carotid artery
thrombosis/ or intracranial arterial diseases/ or
cerebral arterial diseases/ or exp "intracranial
embolism and thrombosis"/ or exp stroke/ 2.
(isch?emi adj6 (stroke or apoplex or cerebral
vasc or cerebrovasc or cva or attack)).tw. 3.
((brain or cerebr or cerebell or vertebrobasil
or hemispher or intracran or intracerebral or
infratentorial or supratentorial or middle
cerebr or mca or anterior circulation) adj5
(isch?emi or infarct or thrombo or emboli or
occlus or hypoxi)).tw. 4. 1 or 2 or 3 5. cell
transplantation/ or stem cell transplantation/ or
cord blood stem cell transplantation/ or
hematopoietic stem cell transplantation/ or
mesenchymal stem cell transplantation/ or
peripheral blood stem cell transplantation/ 6.
stem cells/ or adult stem cells/ or embryonic
stem cells/ or fetal stem cells/ or fibroblasts/
or hematopoietic stem cells/ or myeloid
progenitor cells/ or erythroid progenitor cells/
or mesenchymal stem cells/ or multipotent stem
cells/ or exp myoblasts/ or pluripotent stem
cells/ or totipotent stem cells/ or tumor stem
cells/ 7. exp cells/tr 8. ((stem or progenitor or
embryo or fetal or foetal or umbilical or bone
marrow or cord blood) adj5 (cell or
cells)).tw. 9. (cell adj5 (transplant or
graft)).tw. 10. (fibroblast or
myoblast).tw. 11. cell transplantation.jn. 12. 5
or 6 or 7 or 8 or 9 or 10 or 11 13. 4 and 12 14.
limit 13 to humans
31
Search results

• Review retrieved citations
• Modify search strategy
• New text word terms
• Synonyms, related terms, variant spellings
• Check MeSH headings
• Check that search strategy identified known
  references

Gradually build a comprehensive strategy
32
Features of reference management software

• Personal reference database
• organise references
• create bibliographies
• Imports references from various databases
• Automatic formatting of papers and bibliographies
• Reference Manager, EndNote
• Locally installed or networked
• EUCS provides half-day courses on Reference
  Manager and EndNote in almost every month of the
  year

33
ConclusionsBe systematic but realistic!

• Clearly define search question
• Identify relevant available sources
• Prepare search strategy(ies)
• Conduct and modify search
• Consider reference management software
• Translation?
• Correctly document and report the search methods

34
Essential Reading
Cochrane Handbook for Systematic Reviews of
Interventions Chapter 6 Searching for studies

• Main sources of potential studies
• Planning the search process
• Designing and carrying out the search strategies
• Managing references
• Correctly documenting and reporting the search
  process

http//www.cochrane.org/resources/handbook/