Searching for the evidence: The process involved

The average physician needs to read an estimated 17 articles per day to keep up to date (1). The abundance of information sources, all which have a unique method for storing and accessing information, can make finding relevant information a challenge. Physicians are often unaware of the different approaches required for searching these sources and, thus, may end up with incomplete information on the topic they are researching. As a result, the University of Alberta's Department of Pediatrics has developed a worksheet (www.pediatrics.ualberta.ca/search.pdf [Figure 1]) to assist physicians with conducting a more complete literature search. We will follow the worksheet throughout the present article to demonstrate the process involved in searching. Worksheet developed by the University of Alberta's Department of Pediatrics (www.pediatrics.ualberta.ca/search.pdf) Worksheet developed by the University of Alberta's Department of Pediatrics (www.pediatrics.ualberta.ca/search.pdf) Create a clinical question. Determine whether the question is a background or a foreground question. Choose the domain (eg, therapy or prognosis). Develop search terms according to the key concepts in the well-built question. Decide whether specific study design(s) best answer the question. Determine where to search. Execute the search by using the key concepts and study design(s) if applicable. A mother brings her three-year-old, who has had a cold and now has a fever with pain in his left ear, to your office. You examine the child and discover that he has an obvious, unequivocal left otitis media. The mother is questioning whether antibiotics should be used for her child and you want to find the latest evidence for this clinical problem. Developing a specific question focuses the topic and (hopefully) results in an answer. Table 1 illustrates the breakdown of a well-built clinical question. The acronym ‘PICO’ can be used to easily remember what to include in a well-built question. Elements of a well-built clinical question Based on “Focusing Clinical Questions”, Centre for Evidence Based Medicine, Oxford, United Kingdom. Elements of a well-built clinical question Based on “Focusing Clinical Questions”, Centre for Evidence Based Medicine, Oxford, United Kingdom. Next, determine whether the question is background or foreground. Background questions require general knowledge about a disorder and typically begin with “Who”, “What”, etc (2). These are often asked by new learners or when rare conditions are found in a patient. If the question is a background question, often the best place to look is a textbook. Foreground questions are used to find specific knowledge about how to manage patients with a disorder and the PICO format is used to formulate the question (2). Answers to foreground questions are most often found in evidence-based summaries (eg, ACP Journal Club or Best Evidence) or databases (eg, PubMed, EMBASE). The question in the above case scenario is a foreground question because it asks how to manage otitis media. Subsequently, choose the domain. Most health care articles can be classified into one of four domains: therapy, diagnosis, prognosis and harm/etiology. Detailed definitions of these domains are available at: . The otitis media question deals with therapy. Before starting the search, map out the major concepts identified in the clinical question. This saves time, makes searching more efficient and can be done by using the attached worksheet. The key concepts are: “otitis media and antibiotics and child”. Using MeSH headings (medical subject headings used to index articles according to their major topics) will help to narrow the search. In the MeSH Browser on PubMed, type in each of the concepts and check if the suggested subject headings are appropriate. The search now looks like: “otitis media[MESH] and antibiotics[MESH] and child[MESH]”. After determining the concepts, considering study design can be important. Particular questions may be best answered by a specific study design. Because the current question is a therapy question, systematic reviews, followed by well-conducted randomized controlled trials, are the highest form of evidence. Following this, consider which sources to search. Depending on the question (and the study design), a variety of sources can be used. A resource to determine which sources to search according to domain is available at: . To search for systematic reviews or randomized controlled trials about antibiotics for otitis media, either PubMed or the Cochrane Library can be used. The “Clinical Queries” link in PubMed is useful for quick searches because it uses the appropriate study types to narrow your search according to domain. Choose the “Systematic Review” filter and type in your search strategy “otitis media[MESH] and antibiotics[MESH] and child[MESH]”. Execute the search. Next, use the “Limits” feature and limit the date to 2000 onward because in the otitis media example we are looking for the most recent information. Searching can be a challenging process. The present article will help you to get started with the basics, but more advanced training or guidance from a professional librarian can be a tremendous aid to finding that elusive piece of evidence.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.