The keywords identified at the first EBP step are now used as search terms to acquire research evidence relevant to the question by conducting a search of the literature.
The search result is a list of titles, author names and journal details for each result. The practitioner may use this list to decide which articles to obtain in full by viewing the title and Abstract (if provided) of each item on the list and thus filtering out those that are not relevant or not likely to be helpful in answering the clinical question. This stage in the process allows us to choose the best available research evidence that is relevant to our clinical question.
As explained in Step 1 Ask we may formulate clinical questions so that we can provide guidance and advice to a particular patient. If the patient is sitting in front of us, this evidence-based practice process would need to be completed in a short period of time. We need to phrase an answerable question, search for information relevant to the question and appraise that information in an appropriate timeframe, from a matter of minutes to a few days at most depending on the situation.
We may also formulate questions to which we feel we will need an answer at some later point. For example, we may be aware of an advertising campaign claiming that a dietary supplement is an effective treatment for macular degeneration. It seems likely that some patients will ask about this, so we may decide to explore a question on this before the question is asked by the patient. In this way, the answer is ready when the patient asks.
The figure below is a flow chart depicting the Acquire process
It is important when searching for evidence that search terms are referred back to your original PICO (see EBP step 1 [link to step Ask]) question. The process of finding evidence follows three key steps:
Identify terms to fit the PICO question. These search terms can be entered into a relevant database (e.g. Medline or Pubmed [link to these sites]) with appropriate Boolean operators (i.e. AND, OR, NOT) and limits.
Use these databases to look for secondary sources (e.g. reviews of high level research evidence) or alternatively see our resource on syntheses and summaries available in eye care. If no such source relevant to the question is found,
Search in databases for primary sources (individual research studies including trials, experiments and surveys)
Searching for secondary or primary research evidence is a skill. It is important to enter appropriate terms into the database so that the output provides you with a manageable amount of directly relevant research evidence.
The following links provide information and guidance on:
How to conduct a search for primary or secondary evidence via a database such as Medline: http://web.med.unsw.edu.au/infoskills/topic.htm
Searching tips including using boolean operators, refining your search and getting the full article: http://web.med.unsw.edu.au/infoskills/jour5.htm
Finding journal articles using a database with an example of Medline at UNSW: http://web.med.unsw.edu.au/infoskills/jour.htm
The importance of using refereed sources: http://web.med.unsw.edu.au/infoskills/refereed.htm
How to make the most of web-based sources: http://web.med.unsw.edu.au/infoskills/internet.htm
Saving time and energy with EBM resources: http://web.med.unsw.edu.au/infoskills/ebm4.htm
Databases (e.g. Medline, Pubmed, Google Scholar) all differ to some extent in the rules they use to search, the filters they use and the indicators they use to limit or expand the search. We can use our answerable clinical question to select key words to enter into the database “search” box. The keywords may need to be refined and we may need to use certain indicators that are recognized by the database in order to conduct an efficient search. This would be a search that finds all of the articles in the database that are directly relevant to the question, and no articles that are not relevant. A less efficient search may yield thousands of articles, only a fraction of which may be directly relevant, or the opposite problem, with very few or no articles.
Once the search has been completed, we may have several pieces of secondary and/or primary research evidence (i.e. Abstracts of journal articles, from the which the full article may be obtained). Depending on the question, we may have just a few, or may be faced with a list of hundreds or thousands of Abstracts, so we need to decide which of the Abstracts describe research that is sufficiently relevant to our question. Based on the Abstracts we also need to filter out the lower quality research evidence and obtain full articles describing the best quality available research evidence.
Look for the following indicators in the Abstract:
Older information is not necessarily invalid, but more recent research may supersede earlier work since methods have improved and as studies are conducted over time their findings are replicated (suggesting valid findings) or found to not be replicable.
For example, was it conducted in a relevant population? You may have been very careful in your choice of search terms, but many irrelevant Abstracts may be included in your search results. Factors such as population relevance are part of external validity.
The quality of a research study can cover many aspects. At this stage, it is pertinent to determine the source of the information.
Does it come from a peer-reviewed journal? Peer-review means other experts in the field have reviewed the article to ensure the methods used and conclusions reached are reasonable. These sources are preferrable as the information in the article is more likely to be of better quality and have reduced bias. There is a resource on the peer-review process that you may wish to read for further information. To determine if the journal you have found is peer-reviewed, check the website itself under the 'About this journal' section or you can check the list of journals available on this site. If you use a credible tool to find research articles, such as Google Scholar, Medline, PubMed or other research databases you are likely to find more credible, peer-reviewed sources. A basic Google or other web browser search is likely to include websites and other sources of non peer-reviewed evidence.
Appraisal of the evidence is part of the next step, but at this stage it is worth looking for some basic indicators of the quality of research. Is this a case report, providing results in only one or a small number of subjects? Does the Abstract suggest a major methodological flaw such as a lack of control, a lack of masking, or any other factor likely to bias the results? If so, and if you have plenty of better options in your list, you may want to overlook this one. Factors such as these are part of internal validity.
Graham (2007) Finding, retrieving and evaluating journal and web-based information for evidence-based optometry. Clin Exp Optom 90: 4: 244-249.
Broadly speaking, two types of research evidence sources are available, primary and secondary sources.
Primary research evidence is found in journal articles publishing original research. However, there are vast numbers of journal articles published every day. Answering a clinical question by searching and appraising primary research can be done, and may be appropriate to address questions that may arise in future, when we do not have a particular patient waiting for advice.
Secondary research evidence is found in a range of sources, including systematic reviews of randomized controlled trials, such as Cochrane reviews, which are valuable to busy practitioners because they address particular clinical questions. Reviews of this kind discuss and summarise the highest level available primary research evidence relevant to the clinical question in hand. Cochrane reviews are updated biennially, so they remain reasonably up to date, but many questions that may arise in optometric practice are not addressed by a review of this kind.
Sources of secondary and primary research evidence can be depicted in the form of a pyramid, with primary research at its base and forms of secondary research evidence building on this.
Click on each of the sections below to see more description.
The second level provides syntheses of the highest available research evidence relevant to clinical questions. Here, we find systematic reviews of the literature, in which all available evidence relevant to a particular clinical question has been sourced and appraised. The best available evidence (e.g. randomized controlled trials, if the question relates to intervention) is used to address the question. A well known example is the Cochrane Collaboration [link] whose Eyes and Vision Group [link] have published systematic reviews on a range of clinical questions that are relevant to optometry [link: http://eyes.cochrane.org/our-reviews]. This source does not provide reviews on all possible optometric clinical questions, of course, but it is reasonably up to date since the reviewed are updated every two years. The practitioner would be very fortunate to find their particular question addressed here. However, the database is growing, and the Cochrane Collaboration seeks suggestions from practitioners who would like to conduct a systematic review on particular clinical questions [link: http://eyes.cochrane.org/get-involved].
In this layer we can also find discipline-specific database of abstracts describing research that has been appraised and found to be clinically relevant and of acceptable quality. The research is rated by two independent researchers who have been trained to rate research methodology. Such discipline-specific databases include speechBITE [link] (for speech pathologists) and psycBITE [link] (for clinical psychologists). Databases of this kind allow clinicians to access discipline-relevant research that has been appraised in terms of methodological rigor. No eye care-specific database exists at the time of writing.
"Based on Haynes B (2006) Of studies, syntheses, synopses, summaries, and systems: the "5S" evolution of information services for evidence-based health care decisions. ACP Journal Club. 145, with permission. (The American College of Physicians is not responsible for any inaccuracy in deviation from the original figure)."
The Fresno Test was developed to assess knowledge and skills in evidence based medicine and has been successfully modified for use by other health professions including optometry.
The Project communicates with, and disseminates information to a large Collaborative Network of interested or interesting individuals/parties. It currently includes more than 18 groups spanning across Australia, England, Ireland, Scotland, Canada and India. If you are interested in being part of this Network, please do not hesitate to send us an email.
Contact Us