; Reviewing the Literature
Documents
Resources
Learning Center
Upload
Plans & pricing Sign in
Sign Out
Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Reviewing the Literature

VIEWS: 18 PAGES: 37

  • pg 1
									Reviewing the Literature

     Prof. Jimmy Volmink
       PHC Directorate
Faculty of Health Science, UCT
   What is a literature review?
• Account of what is written on a particular
  subject by scholars and researchers
• Purpose: convey current knowledge and
  ideas as well as their strength and
  weaknesses
• May be:
  – part of essay, research report or thesis
  – stand alone exercise
     Annotated bibliography vs.
         literature review
• AB: List of sources each accompanied by
  a brief note (annotation) discussing one or
  more aspects e.g. content, usefulness,
  quality, etc
• LR: Discursive writing – presents themes
  and concepts; integrates ideas and
  concepts
  Objectives of literature review
1. Address a research question or thesis
2. Critically appraise research for validity
3. Synthesize and summarise current
   knowledge
4. Identify controversies in literature
5. Formulate questions for future research
         Information explosion
•   Journals
•   Textbooks
•   Unpublished literature
•   World Wide Web
Reviews… reduce the load!!!
  Reviews: what do we want ?
• A readable summary of ALL the evidence
• Unbiased
• Transparent
• Up-to-date
But is this what we get?
Corticosteroids for prevention of
 respiratory distress syndrome
• Expert review of the literature
  (Robertson, BMJ 1982)
• “The evidence suggests that antenatal
  steroids are of value only in white
  males, and even for them the benefit is
  mainly in those of 30-32 weeks’
  gestation…”
  Corticosteroids for prevention of
   respiratory distress syndrome
• Another view (Crowley, 1989)
• “The benefits of antenatal
  corticosteroids have been established.
  No further trials are necessary with the
  exception of certain specific situations
  (such as pre-eclampsia) or to establish
  other dosages or routes of
  administration.”
       Treatments for Acute MI
• Existing evidence compared with
  recommendations in textbooks and
  journals
• “advice on some life-saving therapies has
  been delayed for more than a decade,
  while other treatments have been
  recommended long after controlled
  research has shown them to be harmful.”
  Antman et al. JAMA 1992; 268: 240-8
          Traditional review
• Qualitative, narrative summaries of
  evidence on a given topic usually written
  by an expert in the field.
• They typically involve informal and
  subjective methods to collect and interpret
  information
  Problems with traditional reviews
    1. Bias in selection of included studies

• Personal (“file drawer”) bias
• Publication bias:
   – Studies with statistically significant results are:
      • More likely to be published
      • More likely to be published in English
      • More likely to be published in journals with high citation
        impact factors
                                        Easterbrooke et al, Lancet 1991
   Problems with traditional reviews
                    2. Indexing bias
       Ongoing


 All
RCTs
                   Unpublished
       Completed
                                 Abstracts,
                                 theses, etc   Non-MEDLINE
                    Published                     journals
                                                 (c. 12000)
                                                   1966+
                                 Journals                       RCTs not
                                               MEDLINE        identified by
                                                 journals      MEDLINE
                                                 (c. 4600)
                                                  1966+          RCTs
                                                              identified by
                                                               MEDLINE
  Problems with traditional reviews
   3. Variable quality of published studies

• Fourth Congress on Peer Review in Biomedical
  Publication :
   “Medical journals are full of serious methodological
     errors “
   “Journal editors are giving no time, energy and
     thought to their craft”
   “Studies are published that reach false conclusions”
                                          BMJ, Sept. 2001:323
    Problems with traditional reviews
                       4. Effect of Study Design
PURPOSE: To assess the evidence of the effectiveness of vitamin
supplementation, specifically vitamins A, C, and E; beta-carotene; folic acid;
antioxidant combinations; and multivitamin supplements, in preventing
cardiovascular disease.

STUDY SELECTION: The researchers selected English-language reports of
randomized trials and cohort studies that assessed vitamin supplementation in
western populations and reported incidence of or death from cardiovascular
events.

CONCLUSIONS: Some good-quality cohort studies have reported an
association between the use of vitamin supplements and lower risk for
cardiovascular disease. Randomized, controlled trials of specific supplements,
however, have failed to demonstrate a consistent or significant effect of any
single vitamin or combination of vitamins on incidence of or death from
cardiovascular disease.

Ann Intern Med. 2003;139:56-70.
Problems with traditional reviews
        5. Effect of Study Size



“ … we still have no clear evidence that
beta-blockers improve long-term survival
after infarction despite almost 20 years of
                clinical trials.”

                      JRA Mitchell. BMJ 1981;282:1565-70
    The Science of Research
           Synthesis
“If science consisted of nothing but the
   laborious accumulation of facts, it
   would soon come to a standstill...…
   Two processes are at work side by
   side, the reception of new material and
   the digestion and assimilation of the
   old; and both are essential…”
                           Lord Rayleigh 1884
          A Systematic Review
                  How is it different?
“A review in which bias has been reduced by the:
  – systematic identification,
  – appraisal,
  – synthesis
  – statistical aggregation (if relevant)

 of all relevant studies on a specific topic according to a
 predetermined and explicit method.”
                                            Moher et al. Lancet; 354: 1896-900
 Steps in conducting a systematic
              review
1.   Formulate topic/question
2.   Decide upon selection criteria for studies
3.   Determine search strategy
4.   Critically appraise research
5.   Extract data
6.   Synthesise and summarise data
       1. The review question
• Do betablockers compared to placebo
  decrease morbidity and mortality in people
  with hypertension?
• Does the administration of Bacillus
  Calmette-Guerin (BCG) influence the risk
  of tuberculosis and all-cause mortality in
  populations exposed to infection?
• PICO Method
 2. Selection criteria for studies
• Types of participants
  – People receiving BCG vaccination,
    irrespective of age
• Types of interventions
  – BCG vaccination given by any route.
• Types of outcome measures
  – Death
  – Tuberculosis
• Types of studies
  – RCT
         3. Search strategy
• Scope of search
  – Published vs. unpublished vs. grey literature
  – Peer-reviewed vs. non peer-reviewed
  – Electronic vs. paper based sources
• Search should be as comprehensive as
  possible
Ingredients of a comprehensive
        literature search
• Electronic databases (highly sensitive
  strategy)
    –   MEDLINE/PUBMED
    –   EMBASE
    –   Cochrane Databases – CENTRAL
    –   Other e.g. CINAHL, AIDSLINE, LILACS, etc.
•   Handsearching of journals
•   Reference lists
•   Conference proceedings
•   Personal communication
4. Critical appraisal of research
• Different types of clinical research
• Need to understand strengths and
  limitations of various study designs
• Know what aspects of a study to assess to
  determine likelihood of bias
            5. Extract data
• Pre-specify items of information to be
  collected
• Data extraction form
• Independent data extraction
6. Synthesise and summarise data
• Qualitative (narrative)
• Quantitative synthesis (uses statistics)
• Some statistical knowledge will be needed
              Meta-analysis
               Pooling the results
• A statistical procedure that pools the results
  of several independent studies considered
  sufficiently similar (homogeneous)
• Not appropriate if studies are too dissimilar
  (heterogeneous)
• Provides a quantitative summary of the
  overall treatment effect
• Visual display called a “forest plot”
Meta-analysis
and the
Forest Plot
•   Estimate of effect
•   Confidence intervals
•   Weighting
•   Overall estimate of
    effect
                                  Meta-analysis
     Beta-carotene and cardiovascular mortality
Cohort studies
  Male health workers                    USA
  Social insurance, men               Finland
  Social insurance, women             Finland
  Male chemical workers           Switzerland
  Hyperlipidaemic men                    USA
  Nursing home residents                 USA

                     Cohorts combined
Trials
  Male smokers                        Finland
  Skin cancer patients                   USA
  (Ex)-smokers, asbestos workers         USA
  Male physicians                        USA

                          Trials combined

                                                0.1   0.5     0.75     1    1.25     1.5   1.75
                                                            Relative risk (95% CI)
  Acknowledging your sources
• Why bother:
  – Honesty and transparency
  – Show research done to reach conclusions
  – Allow readers to identify and retrieve
    references
• Plagiarism – “taking, using and passing off
  as your own work the ideas or words of
  another”
  – See UCT policy
              Types of Citation
Two main types:
1. Harvard: author-date style
  – …as one author has put it “the darkest days were
    still ahead” (Weston 1988, p.45)
2. Vancouver: footnote/endnote style set by the
   International Committee of Medical Editors
  – …as one author has put it “the darkest days were
    still ahead”[1]
     •   [ ] , ( ) or supercript …ahead”1 all acceptable
       In summary…

Traditional vs. systematic review
          Traditional review
• Qualitative, narrative summaries of
  evidence on a given topic usually written
  by an expert in the field. Typically, involve
  informal and subjective methods to collect
  and interpret information
         Systematic review
“A review in which bias has been reduced
 by the systematic identification, appraisal,
 synthesis, and, if relevant statistical
 aggregation of all relevant studies on a
 specific topic according to a
 predetermined and explicit method.”
              Moher et al. Lancet 1999; 354: 1896-900
               Cochrane review
• A systematic review produced by the Cochrane
  Collaboration (www.cochranelibrary.com)
   – Standardised format
   – Extensive peer review
   – Published electronically on the Cochrane Library (and indexed in
     Medline)
   – Invites comments and criticism
   – Kept up-to-date
• Quality and reporting on average better than other
  systematic reviews ….but not without flaws!
Summary: Types of reviews
    Meta-analyses


              Systematic
                reviews

              All reviews

								
To top