Presentation on theme: "How to do a literature search Andrew Booth (University of Sheffield) and Mary Dixon- Woods (University of Leicester)"— Presentation transcript:
How to do a literature search Andrew Booth (University of Sheffield) and Mary Dixon- Woods (University of Leicester)
This course will cover: The role of the literature review [10 mins] Types of evidence [5 mins] Formulating answerable questions [10 mins] Sources of evidence [12 mins] Search techniques [30 mins] Optimal search strategies [10 mins] Evaluating your literature searching [10 mins] Evaluation [3 mins]
How to do a literature search The role of the literature review
The role of the literature review Exposes main gaps in knowledge [and] identifies principal areas of dispute and uncertainty (Mays et al, 2001). 2.Helps identify general patterns to findings from multiple examples of research in the same area. 3.Juxtaposing studies with apparently conflicting findings helps explore explanations for discrepancies.
The role of the literature review Helps define your terminology or identify variations in definitions used by researchers or practitioners. 5.Helps to identify appropriate research methodologies. 6.You can also identify validated scales and instruments.
Case study: Attitudes of different professions to handwashing in a delivery suite - 1 A midwife is researching attitudes of different staff to handwashing. She firstly searches the literature to focus the scope of the original question. Although the literature on handwashing is vast she needs to discover whether published research has been conducted specifically in obstetrics and gynaecology settings. Has anyone researched the topic specifically in a delivery suite?
Case study: Attitudes of different professions to handwashing in a delivery suite - 2 Handwashing has many more definitions than she had envisaged. Does handwashing include the use of a handrub? Does it constitute use of water only? What is the minimum duration of the procedure before it is classed as "handwashing"? The literature search enables her to explore different definitions of her main concepts.
Case study: Attitudes of different professions to handwashing in a delivery suite - 3 Are there validated instruments to measure attitudes to handwashing (or towards routine hospital hygiene)? The literature review may inform selection of appropriate outcomes - those employed in previous studies (literary warrant) or those considered appropriate by the relevant clinical community (user warrant). Will she focus singly on attitudes or will she investigate knowledge and/or behaviour?
Q: Any other roles or uses for literature review that you have identified?
How to do a literature search The role of the literature review [10 mins] Types of evidence
Some types of evidence Prediction – Models, case studies (single and multiple), documentary analysis Historical – documentary analysis, case studies, narratives Intervention – experimental studies Exploration – literature review, theory building, consensus processes Attitudes – psychological research Qualitative - using specific qualitative techniques Causation – observational studies (e.g. case control)
Exercise: What types of evidence are relevant to your question?
How to do a literature search The role of the literature review [10 mins] Types of evidence [5 mins] Formulating answerable questions
Translates Aims into achievable and focused tasks Helps to distinguish main from subsidiary questions Helps to identify the likely research designs to answer the research question Helps in constructing the literature search
Why? "Ask a poor question and you will get a poor review. A clear question also helps the reader rapidly assess whether the review is relevant to his or her own…practice". (Counsell, 1997) Clearly framed questions "guide much of the review process including strategies for locating and selecting studies or data, for critically appraising their relevance and validity, and for analysing variation among their results". (Cochrane Handbook)
Some types of question Prediction – What is the likely result of X? Historical – How have we got from A to B? Intervention – Is doing Y better than doing Z? Exploration – What are the possible explanations for A? Attitudes – How do people feel about B? Causation – What are the likely causes of C? Measurement: What is the size of X, how often does it occur etc? Characterisation: how can we understand and specify W?
Q: What type of question is your PhD question?
A structure for formulated questions Health services research uses PATIENT- INTERVENTION-COMPARISON-OUTCOME (PICO) structure Within social sciences research the following may be more appropriate: SETTING POPULATION INTERVENTION COMPARISON EVALUATION
An example of SPICE [Actual PhD example] SETTING – UK Marine Conservation POPULATION – public, private and voluntary sector organisations INTERVENTION - partnership approaches COMPARISON – [Unilateral approaches or other countries approaches] EVALUATION – achievement of strategic marine conservation objectives
An example of SPICE [Actual PhD example] SETTING – UK Television Industry POPULATION – Women INTERVENTION – None COMPARISON – [With men or with other industries] EVALUATION – historical, political, organisational and practical issues
Try one! What is the effect of secondary school headteachers leadership and management strategies on student achievement, attitudes, behaviour and recruitment?
Exercise: Now you try! Use the SPICE framework to formulate your own research question
Scenario – A fistful of doulas? You are working on an ESRC funded project looking at the benefits of a doula (a woman experienced in childbirth who provides continuous physical, emotional, and informational support to the mother before, during and just after childbirth) for low- income mothers.
Our example of SPICE SETTING – the Developed World (with comparable health systems to the UK) POPULATION – Low-income mothers INTERVENTION – Doula (Lay support) COMPARISON – Professional support [or No Support] EVALUATION – Perceived levels of social support, birth outcomes, levels of breastfeeding etc.
How to do a literature search The role of the literature review [10 mins] Types of evidence [5 mins] Formulating answerable questions [10 mins] Sources of evidence
Stages in the literature search process Focus your question Decide on the most appropriate sources Perform a scoping search by: –Dividing your search into a series of concepts –Thinking of alternative terms for each concept –Searching each concept separately –Combining concepts using Boolean logic –Limiting your search Revise your search, as necessary, and replicate in other sources
Sources Electronic databases: –Evidence-based The Cochrane Library, Clinical Evidence, etc. –Subject specific ERIC, EconLit, Medline, PsycLit, Cinahl, British Nursing Index, AgeInfo, Biological Abstracts, etc. –Related disciplines LISA, ASSIA, British Humanities Index, Web of Science (Science, Social Sciences, Arts & Humanities Citation Indexes), etc.
Reflective Exercise Which of the listed databases might be of relevance to this topic? [N.B. We are going to search Web of Science and Ovid Medline/CINAHL]
How to do a literature search The role of the literature review [10 mins] Types of evidence [5 mins] Formulating answerable questions [10 mins] Sources of evidence [12 mins] Search techniques
From: Etext on Health Technology Assessment (HTA) Information Resources. Chapter 4: Searching MEDLINE/PubMed for Health Technology Assessment Information by Viveka Alton and Ione Auston. ta/chapter4.html
Take Home Message – 1 An optimal search will combine natural language and controlled vocabulary approaches
Natural language Uses your own words and searches words & phrases (typically from the title, abstract & keyword fields) to retrieve records Potentially can search any field of a database Can be very precise (but there can be ambiguities e.g. Moderation [exam process] and Moderation [within reasonable limits] ) Some problems: –Plurals: e.g. child or children –Different spellings: e.g. esthetic or aesthetic –Different terminology: e.g. pavement or sidewalk –Prefixes: prenatal, pre natal, pre-natal
Database features to support natural language Truncation (e.g. *, $) used to search for different word stems and word endings –e.g. use comput* to find computer, computers, computed, computing, etc. (But comp* would find compost!) Wild cards (e.g. *, ?) used to search for spelling variants –e.g. use leuk*mia to find leukaemia or leukemia Proximity and adjacency operators (e.g. adj or near) –e.g. motor near2 accidents
Controlled vocabulary Maps variations in terminology to a single approved word or phrase (e.g. Toyota, Rolls Royce, Mini, Ferrari etc. all mapped to Automobiles) Can demonstrate hierarchical or conceptual relationships (e.g. Europe-UK-England- Hampshire-Southampton) May not exist for new terms (e.g. single currency) May not map to a precise term (e.g. teaching techniques for problem based learning)
Database features to support controlled vocabulary A Thesaurus (e.g. MeSH, ERIC) [NB. Opposite direction to Rogets Thesaurus – Many-to-One] Mapping Explode functions See Under, Used For and See Also references
Mapping Vocabulary mapping uses statistical analysis to determine which subject headings (index terms) occur most frequently in documents containing your free text query.
Thesaurus A controlled vocabulary arranged in a known (e.g. alphabetical or hierarchical) order containing: –Preferred terms (keywords) –non - preferred terms (synonyms...) –Semantic Relations (Broad Term, Narrow Term, Related Term) –Scope notes
What practical problems would you encounter if you used the following search terms? Postqualifying (2 problems) Labour support (3) PMT (2) Stroke (2) Public school (3) Aids (2) Adolecent (2)
Take Home Message – 2 Getting unexpectedly few results can be worse than getting no results at all – satisfied but inept!
Explosions Pregnancy Childbirth Childbirth, Premature Vaginal Birth Vaginal Birth After Cesarean Labor Cervix Dilatation and Effacement Labor Pain Labor Presentation Labor Stages Uterine Contraction broad search specific search exploded
Boolean - OR DOULAS OR LABOUR SUPPORT DOULAS LABOUR SUPPORT Use to combine like terms or terms within the same concept
Boolean - AND CHILDBIRTH CONTINUOUS SUPPORT CHILDBIRTH AND CONTINUOUS SUPPORT Use to combine together different concepts
Boolean - NOT DOULA NOT MIDWIFE DOULA MIDWIFE Use to exclude terms from your search
Using Boolean in your strategy POPULATIONINTERVENTIO N EVALUATION CHILDBIRTH OR LABOUR DOULAS OR LAY SUPPORT ATTITUDES, VIEWS, OPINIONS AND
Citation searching (Web of Science feature) Find a key reference from (at least) last 5-10 years and follow all references that cite it
Other search techniques Reference chaining: Follow up references from reference lists of relevant articles Hand searching: Identify key journals in your field and browse them cover to cover Relevance feedback: Look at subject indexing for a key reference and use to modify your search terms (Also See Related Records features)
Related records EITHER on the basis of frequency of common subject headings (MEDLINE) OR on the basis of frequency of common citations (Web of Science)
How to do a literature search The role of the literature review [10 mins] Types of evidence [5 mins] Formulating answerable questions [10 mins] Sources of evidence [12 mins] Search techniques [30 mins] Optimal search strategies
What is an optimal search strategy? optimal permutations of search terms found in the titles, abstracts or the subject indexing of relevant articles that have been demonstrated to have a high correlation with study quality pre-prepared search strategies, previously referred to as search filters, quality filters, hedges or optimal search strategies… developed (and usually tested) for use with particular databases and/or search interfaces to retrieve specific types of evidence, study design or …information more effectively
Example 1 – Qualitative Research Qualitative Research/ OR Questionnaires/ OR exp Attitude/ findings OR interview$ OR Interviews/ OR qualitative
Example 2 – User Views Exp Attitude/ Attitude$ OR (user$ adj view$) OR opinion$
How to do a literature search The role of the literature review [10 mins] Types of evidence [5 mins] Formulating answerable questions [10 mins] Sources of evidence [12 mins] Search techniques [30 mins] Optimal search strategies [10 mins] Evaluating your literature searching Evaluation [3 mins]
Evaluating your search strategy RelevantNot Relevant RetrievedUse to generate additional search terms Eliminate terms with poor yield Not RetrievedUse reference lists from key articles to identify these No further action required
Take Home Message 3: Judge not by what you have retrieved but by what you may have missed!
And finally … Documenting a search –Helps to avoid duplication, allows replication in future –e.g. date of search, sources searched, no. of hits, details of strategy, etc. Reference management –Reference Manager, EndNote, etc.
Sources of help Your local University library Database help pages Web-based teaching materials Reading List
How to do a literature search The role of the literature review [10 mins] Formulating answerable questions [5 mins] Types of evidence [10 mins] Sources of evidence [12 mins] Search techniques [30 mins] Optimal search strategies [10 mins] Evaluating your literature searching [10 mins] Evaluation
Identify one thing that you will do when you get back to work as a result of todays workshop Identify one aspect of todays workshop that you will need to explore further.