Searching for Evidence-Based Medicine in Wound Care: An Introduction
Evidence-based medicine can be defined in many ways and many authors use this term in a variety of contexts. However, Sackett's definition1 can easily be applied to the practice of wound care: "Evidence-based medicine (EBM) is the integration of best research evidence with clinical expertise and patient values."
In other words, healthcare practitioners must review the best research evidence (this type of evidence is often provided by scientists who may or may not be directly involved in patient care) and interpret and compare this research material with the current methods of management. This new information then is applied to patient care, keeping in mind that patient-centered concerns are a fundamental part of the plan of care.
The concept of evidence-based medicine includes an organized approach to information that does not exist with more traditional, opinion-based healthcare practice. This organized approach or structured method of analysis allows healthcare practitioners to choose the best interventions for their patients.
Assessing Evidence-Based Literature
One of the biggest hurdles to overcome when looking for evidence-based medicine is developing an approach to searching the literature for valid, reproducible, useful, and relevant studies that provide evidence to change or support the clinician's current practice. Adopting a strategy that can be applied when searching for specific information is valuable. A strategic framework for seeking out evidence-based medicine would be as follows:
1. Ask a question.
a. What information are you looking for?
b. Does it relate to therapy, diagnosis, prognosis, harm or causality?
2. Decide where you are going to search and choose key words.
a. Has this search been done already?
b. If so, where can it be found?
3. How are you going to determine if the results are valid and relevant?
4. Does this new information answer your original question?
5. Apply the information to clinical practice.
6. Evaluate the final outcome on patient care.
The following will demonstrate how the strategy provided above can be used to search the literature for evidence-based medicine that can be integrated into the management of wound care patients. A glossary is provided for the reader's convenience (see "Glossary of Terms").
Ask the question.
Clinical case 1. Mr. J. is a 63-year-old person with diabetes who has a slow-healing neuropathic foot ulcer. The wound care team has experience with cadexomer iodine in treating venous ulcers and wonders if it has been used in diabetic foot ulcers.
The question could be: Is sound clinical evidence available to support the use of cadexomer iodine in a diabetic neuropathic foot ulcer? The key words or phrases in this example would be "cadexomer iodine" and "diabetic foot ulcer."
Decide where to search. PubMed is one of the most accessible search tools for medical literature. A service of the National Library of Medicine (available to anyone with Internet access at www.pubmed.gov2), it allows access to more than 12 million MEDLINE citations dating back to the mid-1960s. A variety of search techniques make looking for information fairly simple. A clinician can perform his/her own search using the following steps that limit the effort to the subject of choice:
1. Go to www.pubmed.gov.
2. Search PubMed for Cadexomer iodine and click "Preview/Index" under the text box.
3. A second page will appear and an additional text box at the bottom of the page will be blank. Enter Diabetic foot ulcer and click "Preview" beside this box.
4. Under "Search and Most Recent Queries, one article is found. Click "1" and the article relative to the query will appear.
By using these advanced search techniques, the final number of articles found is limited. If "cadexomer iodine" only was entered as the key word, 35 articles would have been retrieved (as of June 24, 2003), which would be tedious to review.
Misspelling is a common, easily made mistake. If the initial search word is spelled incorrectly (eg, "cadexemer iodine" instead of "cadexomer iodine"), the search yields 63,901 results because the misspelled word "cadexemer" is not recognized, forcing the database to use only "iodine" as the key word and resulting in many irrelevant search results.
Another method to increasingly streamline search results is to use the built-in filters provided by PubMed. These filters are derived primarily from the work of Haynes and colleagues3 and uses four specific categories: therapy, diagnosis, etiology, and prognosis with two types of emphasis: sensitivity and specificity. Sensitivity allows the researcher to gather the majority of relevant articles on the subject, yielding a broad, all-encompassing retrieval with a high number of results. Specificity offers more precise results with less retrieval; however, not all relevant articles will be offered. These filters can be used when searching under the heading "Clinical Queries." Generally, more articles are retrieved using sensitivity, but when trying to filter a large number of articles, specificity might be a more useful because mostly relevant articles will be retrieved with only a few omissions.
A different approach that utilizes these filters within the clinical queries section of PubMED can be demonstrated using the original clinical case example of cadexomer iodine and diabetic foot ulcers:
1. Go to www.pubmed.gov.
2. Click "Clinical Queries" in the left-hand column of the web page.
3. Decide on the category and emphasis:
In the example, "therapy" is the category, and the emphasis will be "specificity."
4. Enter "subject search" Cadexomer iodine and diabetic foot and click "Go."
5. No articles are found.
6. Change the emphasis to "sensitivity" and use the same key words for the search.
7. This search yields one article4:
Apelquist J, Ragnarson Tennvall G. Cavity foot ulcers in diabetic patients: a comparative study of cadexomer iodine ointment and standard treatment. An economic analysis alongside a clinical trial. Acta Derm Venereol. 1996;76(3):231-235.4
This is the same article found in the previous search; however, in many cases where the first search led to many results, this search will narrow down the number of results.
Systematic reviews and studies utilizing a meta-analysis (a specific type of systematic review that includes statistical analysis) can be found under the clinical queries heading. A systematic review is one in which worldwide scientific evidence is summarized and reviewed using explicit methods and critical appraisal with or without specific statistical techniques.5 These are distinct from narrative reviews that combine opinions and evidence. Well-defined guidelines on performing a systematic review are available on-line.6 Systematic reviews also can be found in any of the four domains of therapy, diagnosis, harm, and prognosis.
An example of meta-analytic reviews accessible through PubMed is the Cochrane Database of Systematic Reviews,7 one of the major results of the work of the Cochrane Collaboration, an international organization committed to maintaining and promoting systematic reviews as a method of obtaining evidence-based medicine in healthcare intervention.
A search for a systematic review also can be performed to determine if an article has been published that has already reviewed the literature for a particular topic; thus, reducing the time spent on searching and analyzing. For this type of literature search in the example clinical case, the key words need to be broadened to "diabetic foot ulcer" and "antimicrobial agents" as follows:
1. Click "Clinical Queries" in the left-hand column of the PubMed web page.
2. Check the circle beside" Systematic Reviews."
3. Enter subject search diabetic foot ulcer and antimicrobial agents and click "Go."
4. This yields three publications8-10:
O' Meara SM, Cullum NA, Majid M, Sheldon TA. Systematic review of antimicrobial agents used for chronic wounds. Br J Surg. 2001;Jan;88 (1):4-218
O'Meara S. Cullum N, Majid M, Sheldon T. Systematic reviews of wound care management (3) antimicrobial agents for chronic wounds: (4) diabetic foot ulceration. Health Technol Assess. 2000; 4(21):1-2379
Temple ME, Nahata MC Pharmacotherapy of lower limb diabetic ulcers. J Am Geriatr Soc. 2000 Jul;48(7):822-828.10
Obtaining systematic reviews that relate to the original question reduces the time required for analysis of the evidence-based medicine because much of the work has already been done. The alternative would be to review many articles with varying levels of evidence to determine if the information found is beneficial - ie, valid and relevant.
Deciding if the information is valid and relevant. When searching for information on a specific therapeutic intervention, clinicians often find articles that may or may not be useful in determining whether that specific therapy is supported by evidence-based medicine. The information in these articles should be reviewed to determine if the evidence for using a specific therapy is valid and relevant.
Healthcare professionals should realize that a new dressing or material may work in an individual case study but may not be applicable to a series of patients or case series. Clinicians are always told about a miraculous new treatment that has worked for the patient of Dr. X or Nurse Y. However, other factors involved in the successful treatment are usually unknown, and the patient may have healed despite the treatment and not because of it. Therefore, once a case series demonstrates a trend or improvement with a new treatment, a controlled trial should be designed.
Clearly defined levels of evidence do exist.11 Homogeneous systemic reviews (SR) of randomized controlled trials (RCT) are usually at the top of the evidence hierarchy, followed by individual RCT as level 1a and 1b, respectively. A review of a number of randomized controlled trials provides more substantial evidence to support the final outcome than just one (or an individual) randomized controlled trial. This is also true of cohort studies and case studies where multiple studies demonstrating similar outcomes would be more useful than a single or individual study. Homogeneous SR of cohort studies, individual cohort studies, and "outcomes" research are level 2a, 2b, and 2c, respectively. Level 3 includes homogeneous SR of case-control and individual case-control studies. Homogeneous systematic reviews do not mix different types of studies; rather, they review whatever studies are available on that topic with that type of study.
A systematic review that includes RCTs as well as cohort studies would be less valid than a systematic review that includes only RCTs. Case series and poor quality cohort and case control studies are Level 4. Expert opinion without critical appraisal provides a fifth level of evidence.
When reviewing the article by Apelquist et al,4 various criteria can be used to decide if the level of evidence is adequate and valid. As the article is read, asking the following questions will enable critical appraisal of the study:
1. How were the treatments assigned? Were certain patients pre-selected for certain treatments or was assignment random?
2. Were the two treatment groups similar? What variables were used to define the treatment groups?
3. Were the treatment groups similar to the reader's patient or patients?
4. Was a different prognosis noted for the two treatment groups?
5. Were all outcomes reported? Were all adverse effects included for both groups? If a patient did not complete the therapy, what was the reason?
6. Were there any patients that crossed-over in therapy? Why, and was this taken into consideration with the final analysis?
7. Were there any deaths? If so, were these deaths related to the treatment?
8. How was the outcome assessed? Were the investigators blind to the therapy used? Was the administration of the therapy blinded?
9. Were any quality assessments included?
10. Were both statistical and clinical significance included? Did the two differ, or were statistical significance and clinical significance the same?
11. Does the treatment have practical significance? Is it reasonable, affordable, and useful?
12. Was a co-intervention present to confound the study?
13. Was compliance/adherence to treatment protocols assessed?
Does this information answer the original question? The original question was, "Is there good clinical evidence to support the use of cadexomer iodine in a diabetic foot ulcer?" The simple answer is no. However, it was indicated that cadexomer iodine may be a more cost-effective treatment than the tested standard treatment. Apelquist et al4 conclude that cadexomer iodine was not clinically different in terms of healing time compared to their standard of treatment, but the cost of the treatment regimen that included cadexomer iodine was less than the cost of the standard treatment. Based on a search of systemic reviews on this subject, O'Meara and colleagues8 concluded that more studies are needed. These authors reviewed 25 randomized trials and concluded that the routine use of antibiotics in venous leg ulcers or diabetic foot ulcers was not supported by evidence and recommended that more research with larger well designed studies is necessary.
A systematic review also can be analyzed to determine if it is credible and if the results are appropriately interpreted - in other words, is this review an example of evidence-based medicine? This can be determined by asking the following key questions while reading this article1:
1. Are the results of this systematic review of therapy valid?
a. Is it a review of randomized trials?
b. Does it include a methods section?
i. Inclusion/exclusion of relevant trials
ii. Assessment of individual validity
c. Were the results consistent from study to study?
d. Did the authors use individual patient or aggregate data?
Additional questions can be asked about systematic reviews, including the questions previously posed regarding assessing the validity of treatment trials. For example, in the systematic review by O'Meara et al,8 30 trials were included, but only 25 of those were randomized, possibly leading to a weaker evidence base8 in the final analysis, which may cause the reader to discount the information in this systematic review. The University of Alberta EBM Toolkit6 includes a worksheet for evaluating systematic reviews about therapy as well as parts of other worksheets that address prognosis, diagnosis, and harm. These worksheets are useful and easy to follow. Similar worksheets used for analysis of systematic reviews, clinical trials, and interpretation of medical literature can be accessed at a number of the other web sites listed in Table 1.
A simpler method is to limit searches to those articles that have already been critically appraised. These articles can be found in evidence-based journals such as ACP Journal Club or Evidence Based Medicine. Evidence-based guidelines and reviews such as those found in the Cochrane Library as part of the Cochrane Collaboration also can be utilized. However, in wound care, the answer to the original question may not be found in these sources or the healthcare practitioner may not have access to these sources. Ultimately, the healthcare practitioner must evaluate the evidence.
Applying the information to clinical practice. Once the medical literature has been reviewed and found to be relevant and valid, the next step is to apply that information to clinical practice.
Often, using an existing model is beneficial. Many models demonstrate the adaptation of research material into healthcare practice, including the Ottawa Model of Research Use (see Figure 1).12
In this example, the information found through a search is first analyzed to determine if it fits the criteria of evidence-based medicine. Once this has been established, healthcare practitioners ("potential adopters") will introduce this information into their clinical practice using various transfer strategies. One such strategy is via clinical practice guidelines, which also have stringent criteria to determine validity and implementation13,14 However, other methods can be used. A method of feedback must be included within each model that allows further critical evaluation of outcomes. The implementation of transferring research information into clinical practice falls within the realm of Knowledge Translation,15 which is well beyond the scope of this article.
Evaluating the final outcome on patient care. Once information has been adopted into clinical use, it is essential that outcomes be evaluated. This step can be further analyzed within a framework using the Appraisal of Guidelines for REsearch and Evaluation (AGREE) Instrument, a step-by-step process designed to allow healthcare practitioners to evaluate the quality and usefulness of clinical practice guidelines in a structured and balanced manner.16
Clinical practice guidelines also need to be constructed so they reflect the evidence base as well as efficiency and effectiveness within the healthcare system. The AGREE instrument is a generic tool designed primarily to assist guideline developers and users in the methodological quality of clinical practice guidelines (www.agreecollaboration.org).16 This instrument has six subscales. Each statement within the sub-scales is graded as strongly agree (four points), somewhat agree (three points), somewhat disagree (two points), and strongly disagree (one point). Each subscale contains a variable number of questions with maximum scores: scope and purpose (12), stakeholders (16), rigor (28), clarity (16), applicability (12) and editorial independence (8). The overall maximum score is 92.
The most important score, which correlates with the overall score for the guidelines in general, is the rigor score. A minimum of three raters should independently use the appraisal tool with the average score used to determine the applicability of guidelines. The raters are asked if they would recommend these guidelines in practice (strongly recommend, recommend with provisos or alterations, would not recommend, or unsure). Guidelines and evidence-based systematic reviews can be found through a number of websites listed in Table 1. Figures 2a and 2b list attributes of guidelines that affect adoption and a summary of effective strategies.
The Canadian Association of Wound Care (CAWC) supports the use of the AGREE instrument and has "chosen [this tool] to assist clinicians to analyze/evaluate protocols and guidelines, and ensure positive outcomes in keeping with "best wound care practice."17 Copies of this tool are available at the CAWC website at www.cawc.net.
Much information is available on evidence-based medicine, including numerous useful books, publications, journal articles, and websites that provide tutorials, guidelines, and toolkits for evidence based medicine. However, searching for evidence-based medicine in wound care is still in the developmental stages. Much of the work in wound care has been left to interpretation by wound care teams who must interpret clinical trial results and systematic reviews and apply this information to practice guidelines. With time, evidence-based medicine will become an essential part of wound care.
Healthcare providers need a method to individually access and evaluate the evidence base. Although generally, a deficiency in adoption of guidelines exists,14 implementation strategies need to overcome the barriers and must be both practice and community based. Creating a toolkit for the successful integration of evidence-based medicine into practice is an important priority. Individuals must understand the material so that when faced with a surprise or a new problem, they have the tools to assess the evidence and interpret guidelines for practice within their constituencies. Evaluating the literature for evidence-based medicine and introducing this information into patient care will help maintain and promote best practice methods in wound care.