Introduction to Literature Searching


greetings everyone my name is Don Jason
I’m a health informationist in the UC Health Sciences Library and today we’ll
be talking about an introduction to literature searching so today’s course
objectives will be developing a Pico question conducting a literature search
and understanding the different types of research so before I get started I just
want to kind of talk a little bit more about Pico and so I’m sure you all may
have heard of it you may be familiar with it but if you haven’t that’s okay
Pico is a consistent formula that helps you develop answerable and researchable
questions and so it may look deceptively simple but it really is something that
can really help you hone the research process and really help you to evaluate
the evidence so here’s the Pico elements there are population or disease
intervention or variable comparison outcome and time and so with the
population or disease that could be the age gender ethnicity or certain disorder
if the intervention could be a exposure to a disease or risk behavior or
prognostic factor the comparison could be placebo or business-as-usual or
another kind of intervention and the outcome is the risk of disease or
accuracy of the diagnosis and then time and so sometimes you will see a t added
on to Pico and that’s just letting you know that there’s a time factor that’s
associated with it and please keep in mind that there’s not always a
comparison and there’s not always a time factor so here’s an example of a Pico
question in teenage patients with eating disorders how effective is cognitive
behavioral therapy in improving self-esteem so let’s identify
those Pico elements so population is teenage patients with eating disorders
intervention is cognitive behavioral therapy comparison is none and outcome
is improved self-esteem and time is the duration of therapy so it’s important to
note that in addition to having Pico and knowing the P I C O and other elements
there are different categories of Pico questions and so here the different
categories listed there is treatment or therapy questions so a therapeutic
intervention there are prevention questions about preventive measures
diagnosis questions or diagnostic tests and procedures or prognosis – prognosis of
disease or condition etiology or harm questions as well as meaning questions
that give you a better understanding so these are all the different categories
of Pico questions and you may see a few different ones out there but just note
that these are the main categories and people might ask well why do we need to
know those different categories of Pico questions and these categories are
helpful because the type of question that you have determines the
best form of evidence and so if you’re looking for a research and you
determined that you know I have a Pico question it’s a diagnosis question then
you know you’re looking for controlled trials when you go to your databases and
then therapy questions you’ll be looking for double-blind RCT systematic reviews
and meta-analyses so that’s why you need to know and a lot of librarians will use
this in order to help you they’ll say here’s your Pico question you all will
sit down you all understand what type of question you have
and then you’ll be able to go into the databases and know exactly what type of
evidence you’re looking for and it can really help you when you want to break
down the number of sources and results you’re getting you may get thousands of
results but if you’re able to filter a limit to control trials or systematic
reviews you may get less results in the end so this is why knowing what type of
category of Pico question you have is very helpful so the next thing we’re
going to talk about is how to conduct a research literature review so how to
conduct a literature review and so for this part of the presentation we’ll see
what is a literature search define the search question decide on research
topics identify main concepts kind of look at some information sources to
decide on databases search the databases using the boolean operators and learn
how to save your search and document your search so the first set of things
we’ll talk about are heavily dependent on Pico so if you’ve already got a solid
Pico question you’re already half of the way there so what is a literature search
it’s really a comprehensive review of the literature where you’re really
synthesizing and breaking down key trends and finding gaps in a
particular subject area and so one of the best places to get started is to
look at the library and so the library has a UC Summon or a
catalog where you can look and put in basic search terms you can search for
books you can search for articles you can search for databases on specific
topics and you can get to it from our home page and so that’s very important
to start your search there you can just type in a few keywords and see what
comes up and then you can start detailing a more comprehensive search
strategy so you know before you may have just searched online or whatnot but when
you want to do a literature search you really have
to start with the library you really want to start honing down on your
research and coming up with that search strategy so defining your search
question and so do you know exactly what you’re looking for and so this is very
important and so it may seem deceptively simple just like the Pico question but
you really want to start looking at that of your question or your main concepts
and understanding what you really want to get out of that question and it’s
okay to talk your research question over with a colleague a professor a friend a
librarian to really understand what you want to get out of that research and
we find this more and more in the health disciplines but it’s also moving to
other disciplines as well so healthcare is leaving the pack but other
disciplines are close behind with doing this and it really helps you to be
comprehensive helps you to sift through the literature really pull out those
things that are going to be valuable to you and to throw away those things that
may be irrelevant so deciding on your research topic so what is the topic of
your research so for example going back to the example we started with before
eating disorders in cognitive behavioral therapy step two will be establishing
what aspect of the topic you’re particularly interested in and so you’re
really interested in this case and how self-esteem interplays with those two
things and then you want to state your topic as a question and that’s that
original Pico question that we talked about before In teenage patients with
eating disorders how effective is cognitive
therapy in improving self-esteem and so you see how that may seem deceptively
simple but we had to go through a lot of steps to really get at that question
there that’s going to be good and can work well in our databases so then the
next thing you’re going to want to do with identify the main concepts in your
question and so a main concept or just those things that are going to play well
with our databases and return the results that you’re looking for I mean I
can’t tell you how many times someone may be in the computer lab or looking at
the databases or at the library website they have the database on advanced
search and they just type their entire question in there and you know they they
don’t even think that oh you know what why am I not getting anything there has
to be something all eating disorders there has to be something on cognitive
behavioral therapy or self-esteem but the thing is they type the entire
question in there just as it appears instead of pulling out those key
elements so when you have your question it’s very important that you kind of
underline or highlight those key those key elements that are going to produce
results in your database and so this is this also helps you to know what
information you’re going to include what type of things are going to exclude and
so I would recommend for your first few searches write it out on a piece of
paper and actually do the highlighting or do the underlining and as you get
more experience with it you may be able to just kind of eyeball what those main
concepts will be so here’s some information sources so I mentioned that
you can always start with the Health Sciences libraries homepage but there
are lots of other areas that you can work from on our website we have a
popular resources page that has about I’d say 10 or 15 different popular
databases that people use overall we have over 800 different databases and so there’s lots of different ways to get information lots of things
to choose from and given that we have so many databases I would also encourage
you to look at our research guides and our research guides are wonderful pieces
of information that our librarians and information is have created and what
they do is go through our catalog they go through our databases and pull out
the ones that are most relevant for a specific topic like there’s one on
evidence-based medicine one on pharmacy one on nursing research so there’s their
data there are there are research guides on hundreds of different topics so I
would start there kind of see what might work well for your what might work well
for your research and then you can use that guide to get started and there’s
just an example of some of our most popular ones the nursing research guide
and evidence-based medicine guide as well there’s also some other recommended
resources so cinahl has a lot of nursing and allied health research in their
nursing Reference Center plus has that nursing research and also some
continuing education modules the trip database is something that is freely
available online and you’re able to kind of go and put your search terms in it
also allows you to search using Pico so you can put your P I C and O in there and
run your search that way and I love that that database has a pyramid or evidence
pyramid featured in it you will always see with the results is it a high level
of research evidence or a low level of that
and so I recommend that database for people that are just getting started
with research and they have you know a little trouble or they may want extra
help with really appraising the quality of the research and of course PubMed is
always a good resource to know about it’s freely available by the National
Library of Medicine and you’ll always have access to it so something you
certainly want to familiarize yourself with hi my name is Emily Kean I’m the
research and education librarian here at the UC Health Sciences Library I’m going
to start by talking about the differences between keywords and subject
headings keywords are what you are primarily familiar with it’s natural
language or free text lay terminology keywords can be good for new or emerging
concepts however keywords can be slightly problematic because they search
every part of the record in the database unless you specifically tell it not to
so that includes searching things like the title of the journal the author’s
affiliation the full text if you tell it to do that
so you can get a lot of results using keyword searching and they’re not always
as specific as you need them to be if you’re just starting, maybe you’re
formulating your Pico question and you’re doing a literature search to
determine what your question is going to be a keyword search might be a good way
to start and if you’re doing like I said an emerging or a newer topic or a really
specific buzz word or phrase something like kangaroo care that would be good to
do a keyword search subject headings will give you more accurate and more
specific results and subject headings can be called different things so in
some databases that’s referred to as a taxonomy or a thesaurus essentially it’s
a controlled vocabulary it’s a predetermined often a proprietary
listing of headings that are actually applied by a human indexer so someone is
reading all of the articles and determining what they’re about and
assigning or indexing those subject headings some of the most common or
commonly used subject headings are mesh which is medical subject headings which
is used by the MEDLINE data set in databases like PubMed cinahl is these
subject heading vocabulary used by the cumulative index for nursing and allied
health in the cinahl database and then Emtree is used by the Embase
database subject headings are great because they provide more specific
results for broad topics however they can be somewhat problematic if the
indexing hasn’t been done yet so if the subject heading hasn’t been applied to
the article obviously that won’t come up in a subject heading search so we really
strongly recommend that people use a combination of keywords and subject
headings to get the best most effective results so one of the things
that I strongly recommend that is typically somewhat different from how
people normally search is that you actually search for each concept
separately so going back to the Pico question that Don described earlier
we’re going to look at term one term two and term three so these are representing
those three concepts that he underlined from that question as being the main
concepts for our search we do put the the phrases into I’m using quotation
marks to keep them as a phrase so this will prevent the database from searching
for just eating or disorders this should return results that only refer to eating
disorders as a phrase so we’ll search each of these separately and when we get
a couple points down the road here it’ll make a little more sense why we do those
separately all of our databases allow for a function called boolean logic what
boolean is there are three terms that allow you to combine your search with
different methods so AND OR and NOT are the boolean operators that you will see
in pretty much every subscription database that the library subscribes to
AND is a restrictive combiner that will narrow
your search results or if you combine two concepts with ORs so we talked
earlier about combining a keyword or a subject heading that will broaden the
results and give you more results for a concept NOT you will see this in the
databases but for basic introductory level searches it’s rarely used it’s in
theory it will subtract a concept from a search but it doesn’t always work very
effectively so I don’t recommend that people use that until you get into more
sort of advanced search strategies so what do the boolean operators look like
in the databases the good news is that you don’t actually really have to know a
lot about constructing the searches the databases help you in a lot of ways so
on the left here we have an example from the EBSCOhost cinahl plus with full
text database we have our three separate searches and then the buttons up here we
would just check which ones we want to combine and combine them with AND or OR just by clicking a button it’s that simple and then here in PubMed from the
advanced search screen we have a drop-down that shows the boolean
operators available there so AND OR and NOT in the case of PubMed so why do we
do all of our searches separately and then combine them when we combine these
three separate results we get this cross section in the middle here
so in theory our search results should reflect all three of the main concepts
from our Pico research question on the right here we have an example of what
the search string looks like so this will become important when we talk later
on about documenting our search again you don’t need to know how to type this
out but the placement of the parentheses and the quotation marks the
capitalization of the boolean operators that is all important and luckily these
days the databases will do that for you so why do we do the searches separately
why do we combine them it’s essentially so that we can modify the search as we
need to one of the most difficult things for a lot of people is that searching is
truly an iterative process you have to keep testing and trying and modifying to
get the correct results I very rarely do a search and I get it right the first
time so I’m gonna do things like combine additional synonyms or terms for my
concepts with OR to expand the results if I get too many results I might
introduce another concept maybe another comparison or another outcome and use
that with AND to narrow you can also apply limits to narrow so a limit would
be something like articles from the past five years articles written by a nurse
articles that are a certain type of research which we talked about with the
types of Pico questions so these are all things that I do at the end of a search
process to narrow my results the one thing to point out is to not apply too
many limits at once and I also like to apply my limits one at a time to see how
many articles I’m losing so we keep doing this process over and over and
over again until we get the results that we feel comfortable with so in this
example let’s say we did our three concepts and we didn’t get enough
results we’re going to expand this search by introducing another term that
can be used for our third concept of self esteem in this case self image so
we’ll do a separate search for self image and then we can or that together
with self esteem so this is why I like to do all my searches separately because
it allows for this type of modification so now we’ll go back and combine our new
concepts with AND we’ll get this cross section in theory we should get more
results because we’re pulling from a larger set of iteraturer for our third
concept and you can see where our search string has now changed again the
database will design this for us but that order is important for
documentation we can just copy this and we’ll talk about how to do that a little
more formally here in a minute so saving your search I’m sure you’ve
noticed if you’ve searched a database that when you close out your browser
session unless you specifically save the search the search is gone it’s all
browser-based and things do timeout as well even if you leave the browser open
so we strongly recommend that you save your results there are multiple ways to
do this it’s kind of a matter of personal preference you can save the
results as a text file you could do a screenshot you could print them off
email them to yourself we also like to recommend that you use reference
software for bibliographic management such as refworks or endnote the other
thing that I’m sure you’ve noticed is that the full text of the articles is
linked out separately so you’ll want to look for that UC article linker or
full-text link to obtain the PDF of the articles and you can save those you can
import them into refworks or endnote there are lots of different management
approaches that you can take there is a slight difference between saving and
document your search we’re starting to talk about documentation a lot
earlier because it’s very important for the concept of reproducibility so it’s
becoming best practice to treat a literature search process like you would
a research process so this means documenting where you search for
information what database what day it was what your search terms were if you
try to search a term and it didn’t yield the results that you thought it would
you might want to document that so we advise that people keep track of these
activities as they search and again this is a personal preference you can use a
worksheet you can use a something like the Open Science framework where you
could have a wiki and document that so there’s a lot of different approaches
that you can take but it’s also important too because it’s sometimes very
rare that your search is all conducted in one session so if you start a search
the phone rings you get interrupted or you revisit it several days or weeks
later it’s nice to have that personal record that you can
refer back to in this section we’re going to talk
about the difference between popular magazines and scholarly journals the
differences between research articles and non research articles and then how
to identify and retrieve those research articles in the databases so the
difference between popular magazines and scholarly journals a lot of times this
becomes very obvious over time as you spend more time working with scholarly
resources popular magazines might be something like time or Newsweek it’s a
resource that’s targeted to anyone a lay audience this also means that it may
contain articles written by anyone there may not be a really rigid review process
so anything could be published a lot of times these contain a lot of advertising
they’re getting funding and resources from advertisers they probably
don’t contain original research and almost never are they peer reviewed so
peer review is sort of the distinguishing mark for a scholarly
journal it means that there is an editorial board of content experts that
when an article is submitted they are providing feedback it’s more of
a dialogue so you know that you know in these areas where something is targeted
to an academic audience it contains articles that have been
vetted by a group of their peers so it contains a lot of things that are more
technical in nature they’re more specific they’re reporting original
research there may be advertisements but there aren’t as many and if there are
advertisements it’s for something you know another peer or scholarly endeavor
maybe like a database or a research tool and again the refereed or peer-reviewed
which means that there is a review process by content experts for these
items to be published so once you’ve identified a scholarly resource
typically a journal within that journal there are going to be research articles
and non research so research articles are primarily reporting on research a
lot of times we refer to this as primary research so that would be something like
a randomized controlled trial where there’s a specific
group of patients they’re documenting the method of the research that was used
there are a lot of times discussing the outcomes or the future implications for
that research there are also different types of research like observational
studies when you get into systematic reviews and integrative reviews
meta-analysis and meta synthesis these are examples of secondary research
they’re still research but in these examples the authors are going out into
the literature and assessing and compiling what’s already been published
in a research format by someone else so these are high levels of research but
technically they’re secondary a non research article would be something like
a clinical practice guideline a white paper a position statement from a
well-regarded Association or an organization these are still quality
resources but they’re technically not research similarly we have things like
expert opinions which may even be solicited or requested from the expert
or the professional so again still a valuable source but not research a case
study we do see a lot of book reviews in the journal publications as well so
obviously that is not an example of research so why does this matter if you
remember back when we were talking about the type of Pico question a lot of times
the type of question will determine the type of literature that you’re going to
use to answer that question and as we move up the research pyramid we see
higher levels of evidence so at the bottom here we have things like we just
discussed like background information or expert opinion still quality but the
lowest level of the pyramid and kind of in the middle here we have things that
are considered unfiltered information so those are our case reports our cohort
studies randomized control trials a lot of these as you get higher up the
pyramid these are primary research but they’re still unfiltered the examples of
filtered information are at the top and are the highest quality of evidence so
we have stronger methodology we should have less bias and we should also have
fewer studies so like a systematic review sometimes you’ll see a pyramid that has
the meta analysis or the meta synthesis at the very top so a systematic review
again is assessing the base of the literature a meta analysis would then be
extracting the data from that systematic review and a lot of times those are used
for evidence based changes so when you’re searching in the database how do
you find an example of a type of research so a systematic review or a
meta-analysis a lot of times you can identify it right from the title or the
abstract so in this example we have an article and the subtitle of the article
is telling us this is a meta-analysis of randomized control trials so we know the
type of research article it is it’s secondary research it’s compiling
research from randomized controlled trials so this one is telling
us you know this is reporting about a randomized controlled trial so this
would be a primary research report of a research study that was undertaken if
the article itself doesn’t identify the type of article it is there are
mechanisms within the database where you can search and
filter that out so this is an example of pubmed if we click on article types we
can select randomized controlled trial or systematic review and limit our
search results just to that article type and here’s an example from the cinahl
database cinahl on the advanced search screen has a checkbox for research
article so we can check that it doesn’t always work with a hundred percent
accuracy but it will it’s supposed to filter out things like secondary
research like systematic reviews sometimes they do sneak in so you still need to do an assessment, an evaluation and make sure
is truly a research article there are additionally checkboxes for peer review
randomized controlled trial and evidence-based practice again I don’t
recommend that you do all of these you would want to be sort of selective so
maybe do one at a time and see what’s out there there’s also a drop-down in
EBSCOhost the cinahl interface for clinical queries so again if you have
identified the type of Pico question you can then use the clinical queries to
just look at literature that would answer a therapy question for example

Leave a Reply

Your email address will not be published. Required fields are marked *