Developing Evidence-Based Algorithms for Negative Pressure Wound Therapy in Adults with Acute and Chronic Wounds: Literature and Expert-based Face Validation Results
Negative pressure wound therapy (NPWT) is used extensively in the management of acute and chronic wounds, but concerns persist about its efficacy, effectiveness, and safety. Available guidelines and algorithms are wound type-specific, not evidence-based, and many lack clearly described relative and absolute contraindications and stop criteria. The purpose of this research was to: 1) develop evidence-based algorithms for the safe use of NPWT in adults with acute and chronic wounds by nonwound expert clinicians, and 2) obtain face validity for the algorithms. Using NPWT meta-analyses and systematic reviews (n = 10), NPWT guidelines of care (n = 12), general evidence-based guidelines of wound care (n = 11), and a framework for transitioning between moisture-retentive and NPWT care (n = 1), a set of three algorithms was developed.
Literature-based validity for each of the 39 discreet algorithm steps/decision points was obtained by reviewing best available evidence from systematic literature reviews (n = 331 publications) and abstraction of all NPWT-relevant publications (n = 182) using the patient-oriented Strength of Recommendation (SORT) taxonomy. Of the 182 NPWT studies abstracted, 25 met criteria for level 1 and 2 evidence but only one general assessment step had both level 1 evidence and an “A” strength of recommendation. Next, an Institutional Review Board-approved, cross-sectional mixed methods survey design face validation pilot study was conducted to solicit comments on, and rate the validity of, the 51 discreet algorithm-related statements, including the 39 decisions/steps. Twelve (12) of the 15 invited interdisciplinary wound experts agreed to participate. The overall algorithm content validity index (CVI) was high (0.96 out of 1). Helpful design suggestions to ensure safe use were made, and participants suggested an examination of commonly used wound definitions in follow-up studies. Results of the literature-based face validation confirm that the evidence base for using NPWT remains limited, especially for chronic wounds, and that safety guidance may be affected by the fact that evidence-based ratings cannot accurately reflect relative or absolute product contraindications because they simply are not included in clinical studies. These findings, the positive expert panel comments, and the high CVI confirm the need for an algorithm with explicit NPWT start-and-stop criteria and suggest that follow-up content and construct validation of these algorithms is warranted.
Potential Conflicts of Interest: This study was supported by a research grant from ConvaTec, Inc (Skillman, NJ). The study was conducted, the data analyzed, and the manuscript written by the authors, who are responsible for the content of the publication. The authors do not have a financial interest in any of the products discussed.
Ever since devices to replace improvised systems for the removal and collection of excess wound exudate1 became commercially available, negative pressure wound therapy (NPWT) systems have been widely used. Initially, only one type of NPWT system was marketed by a single manufacturer, but recently many other NPWT devices have been introduced into the healthcare professional marketplace.
Major impetus for generating a new perspective on NPWT usage are its clinical cost and limited knowledge about its cost effectiveness for the wide variety of clinical indications.2 General evidence to substantiate current indications and contraindications for use is limited and sometimes contradictory. Algorithms providing a step-by-step visual interpretation of some NPWT guidelines have been developed for specific types of wounds, but most are based on expert consensus, have not been validated, and do not include guidance for when NPWT treatment should be discontinued.3-7
The purpose of this research was to: 1) develop an evidence-based algorithm for nonwound expert clinician use of NPWT that would provide stepwise evaluation and assessment strategies for optimal use and strategies for transitioning between moist wound healing and NPWT in adults with acute and chronic wounds; 2) obtain face validity for the algorithms, and ultimately, 3) obtain algorithm content validity.
In the fall 2009, following Food and Drug Administration (FDA) reports about adverse events related to NPWT use,8 a research proposal was constructed by the researchers at the request of industry (ConvaTec, Inc, Skillman, NJ) in preparation for possible grant funding to La Salle University, targeting these issues, including specifically to: 1) develop a set of evidence-based algorithms targeting safe use of NPWT in adults with acute or chronic wounds; 2) test the newly formulated NPWT algorithms for face validity, 3) revise algorithms as needed, and 4) obtain full content validation of the algorithms with possible further revisions as needed.
As part of the initial process, also in the fall of 2009, an expert consensus meeting with international and national wound care clinicians was held to discuss goals and considerations for NPWT use, goals of moist wound healing, and overall wound management issues for pressure ulcers, diabetic foot ulcers, venous and arterial ulcers, skin grafts/skin substitutes, and complex trauma and surgical wounds. Preliminary feedback was synthesized, and the final formalized consensus guidelines subsequently were published.9
Evidence-Based Algorithm Development
Initial literature review and algorithm framework development. To start the process of algorithm development, a systematic literature search was conducted to identify meta-analyses and systematic reviews of NPWT published between 2005 and 2009 using CINAHL, PubMed, and EMBASE, as well as the Cochrane and Joanna Briggs Institute databases. Ten (10) publications were identified.2,4,10-17 Indications and contraindications were abstracted and references cited in the documents obtained (see Figure 1).
Next, existing NPWT algorithms for specific wound types and guidelines of care were retrieved using the same databases and expanding the publication range to between the years 2003 and 2010. This search initially yielded 12 references.3,5-7,9,18-24
The abstracted information was reviewed and summarized, and a general indications and contraindications framework was developed. Next, the algorithm framework was compared to published evidence-based guidelines and algorithms of patient and wound care25-36 in order to rule out any inconsistencies or discrepancies.
Before initiating the broader literature searches, a spreadsheet was created containing all variables that affect NPWT wound treatment decisions (exudate amount, depth, wound bed characteristics, surrounding skin condition, and other factors)9,26-28,36-40 and that might serve as potential NPWT start/stop criteria. In addition, an existing algorithm (Solutions®, ConvaTec, Skillman, NJ) was reviewed for information affecting transition to moist wound healing.41 The Solutions Algorithms were selected because they are the only validated wound care algorithms for acute and chronic wounds. They are listed in the National Guidelines Clearinghouse.
Wound debridement status, study type, design, sample size, outcomes, complications, and general conclusions or recommendations from the research were abstracted and the quality of the evidence and strength of recommendation recorded.42
Strength-of-evidence ratings. The Strength of Recommendation (SORT) Taxonomy42 was used to abstract and record the level of evidence (study quality) and strength of recommendation in the spreadsheet for all publications reviewed. Study quality is based on patient-oriented evidence (expressed as a number grade of 1 through 3). Strength of recommendation is expressed using letter grades, where A means the recommendation is based on consistent and good quality patient-oriented evidence, B denotes inconsistent or limited-quality patient-oriented evidence, and C is based on consensus, usual practice, opinion, disease-oriented evidence, or case series. All information from previous and subsequent literature searches was reviewed and included in the master spreadsheet.
Ongoing literature review. Systematic literature searches were conducted at the start of the algorithm development process and periodically updated for the next 2 years until September 2011. Using the same databases as those used for the initial search and starting with a 2005 year of publication, the following search terms were used to obtain NPWT publications: negative pressure wound therapy, VAC therapy, negative pressure dressings, topical negative pressure, vacuum-assisted closure, negative pressure therapy, subatmospheric pressure dressing, subatmospheric pressure, and TNP. Older “classic” articles identified in the literature also were included, as were pertinent technology reviews.2,33 Preclinical study results were not included in the evidence database, but publications were reviewed to glean potential information that might affect NPWT indications, usage, and/or transition to a moisture-retentive dressing regimen.43-58
Literature review result. A total of 331 publications, including the literature used to develop the initial framework, was retrieved and reviewed. Of those, 182 were clinical studies and clinically relevant publications on NPWT that were systematically reviewed, graded, and abstracted. This process revealed important trends. For example, many more studies support the use of NPWT for acute than for chronic wounds. Clinically relevant publications and studies on the use of NPWT for acute wound conditions included: mediastinitis,23,55,59-67 surgical myocutaneous flaps,68-70 laparotomy wounds,22,71-76 lower extremity amputation or surgically debrided wounds in persons with diabetes mellitus,77-82 abdominal compartment syndrome or open abdomen,1,10,83-97 skin graft recipient,98-100 skin graft donor sites,48 fractures and trauma,101-108 and other complex acute109 or surgical wounds.110,111
Clinical studies or clinically relevant publications about the use of NPWT on pressure ulcers and other chronic, nonhealing wounds12-15,38,39,80,112-127 were much less available.
Literature reviews, guidelines, and consensus statements on the use of NPWT in various wound care scenarios or articles describing individual case studies or recommendations3-5,7,9,11,18-21,67,119,128-148 were frequently encountered.
Of all the NPWT studies identified, 25 met the SORT criteria for level 1 or 2 evidence. Of those, 12 described study results of acute wounds,62,64,76-78,82,85,90,100,108,110,149 seven described chronic wounds,80,112,120,121,124-126 and six detailed chronic and acute wound38,113,150-153 study results.
In developing the algorithm framework, the researchers identified a need to review other literature sources and obtain evidence level on algorithm components such as patient goals of care. For example, wound healing may not be a goal of a palliative protocol of care.119,154-159 Further literature was reviewed for moist wound healing indications160,161 and wound debridement.162,163 In addition, the need to obtain updated literature on risk factors for delayed wound healing and chronic wound development164 was identified to support the underlying algorithm structure, specifically publications addressing care of venous ulcers,165-172 arterial ulcers,30,31 pressure ulcers,123,173-177 diabetic ulcers178-180 and wound pain management/quality of life, patient education, and nutritional support issues.37,181-187
Algorithm refinement. Using the previously developed framework and observations from the literature, the algorithms were developed to provide a step-by-step approach to NPWT implementation and discontinuation in adults only, but the pediatric literature was reviewed to glean usage criteria that may not have been included in other publications.103,188-191 The first algorithm takes users through the initial patient and wound assessment steps to: 1) develop goals of patient and wound care, and 2) assess patient and wound variables affecting safe and appropriate use of NPWT and potential alternatives (eg, moist wound healing). NPWT contraindications and safety-related stop criteria are highlighted. The next two algorithms help users decide when to continue or stop NPWT when managing acute (algorithm 2) or chronic wounds (algorithm 3) (see Figures 2, 3, 4).
In addition, using information from the literature search results, the algorithms’ structure was derived from several sources, including the validity of decision points and data inherent in previously content and construct validated Solutions® Wound Care Algorithms,27,28,192 consensus conference suggestions for transitioning to moist wound care provided by an expert group,9 and established algorithm development criteria.193
Face Validation Methods and Procedures
Literature-based validation. SORT level of evidence and strength of recommendation tables42 were developed for the 39 discreet decision points/steps in the algorithms. The abstracted information was used to populate the evidence table, containing 39 algorithm components. When evidence gaps were identified in the NPWT or other recommendations, additional literature searches were conducted to identify (in this order): meta-analyses, systematic reviews, validated decision rules, randomized controlled clinical studies, noncontrolled clinical studies, case studies, and consensus documents. If publications with higher levels of evidence were found to support a decision, the literature search for that particular variable was stopped. The detailed literature reviews and evidence development process led to several revisions (five drafts) before the algorithm was presented for external (expert) face validation (see Figures 2, 3, and 4).
Expert face and content validation. The purpose of the face and preliminary content validation was to obtain feedback from a panel of wound experts and to: 1) ascertain if any changes to the algorithm structure or components were needed, 2) ascertain expert rating of the appropriateness/face validity of the components and processes inherent in the NPWT algorithms, and 3) elicit comments about the overall research process and the algorithms. After receiving approval from a university’s Institutional Review Board (IRB) and using a cross-sectional, mixed-methods quantitative mail survey design with a qualitative component, a purposefully selected group of 15 international wound and NPWT experts was invited to provide quantitative ratings and written comments regarding the first formal (draft five) iteration of the new NPWT algorithms. The survey was accompanied by an explanation that the NPWT algorithms were developed for nonexpert users. Participants were asked to respond based on their experience with NPWT. Anonymity and confidentiality were ensured, participants provided written informed consent, and the researchers did not provide financial compensation to the study participants. None of the participants was employed by wound product manufacturers.
Instrumentation. The face validation instrument consisted of professionally produced color copies of the three NPWT algorithms and a paper-and-pencil data collection instrument. The latter consisted of: 1) a demographic data sheet (education, clinical practice issues, types of wounds encountered), 2) content validation rating sheets for all general algorithm descriptions and each decision point/step, and 3) open-ended questions related to participants’ perceptions of specific decision points and the overall process.
The validation instrument consisted of 51 discreet statements related to the construct of the algorithm general statements and 39 decision points/steps. Steps that are repeated in each algorithm were only tested once; hence, the number of decision points/steps tested decreased from 20 for algorithm 1 to 14 for algorithm 3.
Each discreet statement was accompanied by a modified 4-point Likert style rating scale: 1 = not relevant/appropriate; 2 = unable to assess relevance/appropriateness without revision; 3 = relevant but needs minor alteration; and 4 = very relevant and appropriate.194 After rating the relevance of the components to safe use of NPWT, participants were asked to write comments about statement omissions, provide suggestions for improvement, present an alternative, and provide references to the literature (if possible).
Data collection and analysis. All data were entered into an Excel Version 2003 spreadsheet (Microsoft Corporation) and uploaded into SPSS version 16.0 (SPSS, Inc, Chicago, IL) for analysis. Summary statistics were calculated for demographic variables. Mean, modal, and content validity index (CVI) scores were analyzed. The CVI is calculated by comparing items rated 3 and 4 (relevant/very relevant) versus items rated 1 and 2 (not relevant/unable to assess relevance).194-196 Scores closer to 1.0 suggest stronger content validity. A contract transcriptionist with a confidentiality agreement transcribed all participants written responses and comments. Using qualitative data reduction techniques, researchers scrutinized the typed comments for themes and subthemes.
Expert validation. Fifteen (15) email invitations were sent and 12 consented (participation rate of 80%). Participants’ average age was 51.6 years (range: 32 – 62, SD 3.81). Seven (7) were women, and all (12) had attended formal wound care classes, obtained wound care certification, and practiced in acute (n = 6), subacute (n = 2), long-term (n = 2), or home care (n = 2). Eight (8) worked in large (>300-bed) and three (3) worked in smaller sized (200- to 999-bed) facilities. All but one were English-speaking only, and all had obtained their basic medical/nursing or healthcare education in the United States; they practiced in eight (8) different US states (see Table 1). The most commonly encountered wounds for which they used NPWT were surgical wounds and pressure ulcers.
Literature-based face validity. The majority of publications available to support the discreet steps in all algorithms consisted of low-level evidence. The percentage of good quality (level 1) evidence ranged from 12% for the acute wound (algorithm 2) to 9% for algorithm 3 (chronic wounds). The percentage of publications with a high strength-of-recommendation level and consistent, good-quality patient-oriented evidence was even lower (1% for acute and 5% for chronic wound algorithms) (see Table 2). Subsequent synthesis of the evidence to support each algorithm step also showed the majority of studies were low-level evidence with a C grade (consensus, usual practice, opinion, or disease-oriented evidence, or case series for studies of diagnosis, treatment, prevention, or screening)42 for strength of recommendation (see Table 3). Three steps/decision points had an A strength of recommendation (see Table 4).
Quantitative data. On a scale of 1 to 4, the mean score for all algorithm component ratings (N = 589) was 3.72 (SD 0.43, range 3.25 – 4.0). The CVI for all algorithms was 0.96 (out of 1). Ratings that were lower (<3.5) but still rated appropriate were found for 1) considering separate goals of wound care for primary intention (sutured) healing or palliative care, 2) proceed with caution only if coagulation is controlled, and 3) assess wound as surgical/acute or chronic wound. Mean scores and CVI results suggested that the expert panel rated the face validity of content for the NPWT algorithms as “strong and appropriate”.
Qualitative data. Themes of positive response and concern targeted several noteworthy findings, including disagreement and lack of uniformity on definitions of acute versus chronic wound; descriptions for primary-, secondary-, and tertiary-intention healing; and “controlled” coagulation. Other suggestions included areas where clarification was needed (palliative care issues), sepsis specific to wound care, control of coagulation issues (how to operationalize), timing issues in changing NPWT dressings, and clarifications regarding NPWT discontinuation and switching therapy (see Table 5). The expert panel’s overall comments were organized into four themes (utility/ease of use, helpful guidance, need for definitions, and references needed). Participants believed the NPWT algorithms were user-friendly and helpful, but they recommended two notable areas requiring revision/addition: references to support the statement about wound size reduction and the addition of a glossary regarding the definitions of acute versus chronic wound and primary, secondary, tertiary intention. Given the experts’ lack of agreement on basic definitions, the authors decided to add another research question for the content validation phase because greater understanding of this component was needed.
Concerns about the safe and appropriate use of NPWT have been raised,8 and currently available guidelines tend to be broad or condition-specific5,6,20,24,147 and therefore difficult to implement in clinical practice. Quantitative analyses and qualitative themes support that the expert panel found the new set of three NPWT algorithms to have strong face validity with some need for revision. Specifically, there was some disagreement and lack of consensus with regard to defining wound care goals related to NPWT for primary intention and palliative care scenarios. Clear literature support of definitions was recommended. Another step that created comments related to the use of NPWT only if coagulation is controlled. Raters thought it was appropriate but needed better explanation of controlled and questioned if the operationalization would differ based on the type of antigoagulant prescribed (coumadin versus low molecular weight heparin versus intravenous heparin) — in other words, how anticoagulated is still “controlled.” Another component with some disagreement was assess wound as surgical/acute or chronic wound. Reviewers called for clear evidence- or literature-based references for “chronic” wound. Some described situations in which acute or chronic wound could be confused.
The small number of problematic, lower-rated steps was relatively unsurprising, given the fact the algorithms were constructed based on an exhaustive systematic literature review, existing evidence-based reviews, expert consensus, and controlled trials — ie, using known criteria for guideline development197 and generally following the Appraisal of Guidelines for Research and Evaluation (AGREE) items and domains for clinical guideline development.198 On the other hand, the confusion over wound terminology among experts was somewhat surprising, but these concerns continue to be identified27,28,192,199 and need to be addressed in future research. Interestingly, the NPWT literature itself may add to the ongoing confusion. For example, among the NPWT studies reviewed, it is not uncommon for research conducted in postamputation (acute) wounds78 to be described as evidence for diabetic foot24 or chronic wound treatment.112
Evidence levels. Although face validity results were strong and confirmed the literature-based recommendations used for algorithm development, the overall level of evidence and strength of evidence for using NPWT was low, confirming the results of previously published systematic reviews using different evidence rating criteria.2,14,15 All evidence-rating criteria, including SORT used for this review,42 have strengths and weaknesses that must be taken into account.200,201 The authors chose the SORT, patient-outcome oriented evidence assessment because 1) the ultimate goal of care, including use of these algorithms, is to improve patient outcomes, and 2) the criteria used make a clear and important distinction between the type of studies needed to obtain evidence for diagnosis, treatment/prevention/screening, and prognosis.
However, even-though 25 of the 182 NPWT publications (14%) met the initial criteria for level 1 or 2 evidence, only one NPWT-related recommendation had sufficient and consistent evidence for an A strength of recommendation (see Table 4). Many NPWT studies report disease-oriented outcomes. Although interim outcomes (eg, reduced amount of exudate) can be useful, unless they have been shown to directly correlate with a patient outcome, their clinical utility is limited.201
Another important limitation is that some well-established interventions, such as the need to assess a wound before dressing application, have not been studied and are unlikely to be the subject of future research, given potential ethical barriers.202 As a result, their level of evidence and strength of recommendation are low and unlikely to increase using standard strength-of-evidence ratings. Finally, for some recommendations, the collective evidence of the “opposite” action or decision point, instead of the actual step, was strong or a particular type of wound was simply not included in studies. For example, the algorithms guide users toward debridement of wounds containing necrotic tissue and transitioning to a moisture-retentive dressing protocol of care for superficial wounds and those with limited amounts of exudate. Technically, these recommendations could not be supported with high levels of evidence or a strong recommendation, but the collective evidence of the “opposite” was strong because almost all NPWT studies only include wounds that are: 1) full-thickness, 2) highly exudating, or 3) clean/debrided. In other words, the recommendation to not use NPWT on lightly exuding, superficial, or nondebrided wounds is based on the absence of evidence to do so. From a clinical and practical point of view, this is an important limitation when using any type of strength-of-evidence rating system for clinical algorithms. Also, it is possible that some evidence ratings would be stronger if a clinical guideline instead of an algorithm format is used, because, by definition, algorithm steps are more narrowly defined than guideline recommendations.
Implications for Clinical Practice, Education, and Research
The NPWT algorithms were developed with the intention of providing evidence-based guidance on safe, appropriate use of NPWT for acute and chronic wound care (surgical and nonsurgical wounds) in adults. In general, the literature- and expert-based face validation phase supported their preliminary content validity among a small panel of international experts. With revisions incorporated, they will be submitted for larger content validation with a substantially larger group of wound experts. If content validity can be established, and with appropriate education, these algorithms should be easier to implement in clinical practice than currently available sets of guidelines. Designed for nonexpert users, the NPWT algorithms may help patient safety by clearly recognizing safe and potentially unsafe patient care situations, but such thinking awaits further testing.
The research implications for the NPWT algorithms are substantial. Further content validation is needed, and effect on patient outcomes should be tested.
In addition to some of the evidence-level limitations described, any extant research has some degree of limitations, and this face validation study is no exception. Although developed based on the current best available evidence on NPWT and wound expert consensus, the NPWT algorithms must be tested. This face validation phase provides a critical but still preliminary step in establishing the construct validity of these algorithms. As previously stated, there is more evidence for the use of NPWT in acute than in chronic wounds, and the content of the algorithms is based on the science as currently constructed.
Another limitation is the focus of the NPWT algorithms on adults. No effort was made to scrutinize care issues for pediatric use. This is an area crying out for further research.
Finally, face validation is usually performed with a small group of experts in a selected area (eg, wound care). Similar to a pilot study, the instrument or algorithm must subsequently be tested or validated with a substantially larger sample size.
Safety in wound care delivery is a major concern for all clinicians, and it has become a strong focus in the use of NPWT. Algorithms may help clarify and expedite clinical decisions about when to safely start and when to stop using NPWT. This study has confirmed the face validity of a new set of algorithms designed to guide care for NPWT in adults with acute and chronic wounds. The results also suggest a need for further analysis of wound care definitions and terminology.
The authors are grateful for the constructive and helpful comments provided by all wound expert study volunteers.
Dr. Beitz is a Professor of Nursing, School of Nursing and Health Sciences, Philadelphia, PA. Ms. van Rijswijk is Adjunct Faculty, Holy Family University School of Nursing and Allied Health Professions, Philadelphia, PA; and Clinical Editor, Ostomy Wound Management. Please send correspondence to: Janice M. Beitz, Professor of Nursing, WOCNEP Co-Director, School of Nursing and Health Sciences, La Salle University, 1900 W. Olney Avenue, Philadelphia, PA 19141; email:firstname.lastname@example.org.
For the list of references cited in this article, please download the PDF.