Our systematic review will include clinical trials in which any category of digital education intervention was utilized to train health care professionals in chronic wound management. The following study designs will be included: randomized controlled trials (RCTs), cluster RCTs, and quasi-RCTs. Cross-over trials will be excluded due to high risk of contamination as a result of carry-on effect. We will include eligible papers published in any language and in any type of publication, including research articles, abstracts, and conference proceedings.

The populations included in this review will include the following: (1) preregistration students pursuing a degree in any health care-related field, in a university or tertiary institution recognized by relevant governmental or professional bodies or (2) postregistration health care professionals, as referred by the Health and Welfare chapter of the International Standard Classification of Education: Fields of education and training 2013 (ISCED-F 2013) [38]. The health care-related professions eligible for the review include the following: medicine, nursing and midwifery, medical diagnostic and treatment technology, and therapy and rehabilitation. If a study presents data on more than one professional group or includes pre- and postregistration participants, it will be included if they report the results for each subgroup separately. Studies presenting a mixed-group analysis will be excluded. The included studies should present any of the comparisons that are listed in Textbox 1.

This systematic review is one in a series of systematic reviews our group is conducting to assess the use of digital education modalities for pre- and postregistration health professionals’ education and training. Each review of the series addresses a different aspect of digital education, including categories (ie, online, offline, virtual reality, etc) [24,25,27,28,30,31], specific pathologies [26,29,32], or learning theories [33,34]. This systematic review will evaluate the use of digital education interventions for chronic wound management training. We will therefore utilize the same literature search as other reviews already completed or under development. The search strategy includes the databases listed in Textbox 2.

All databases were searched for studies published from January 1990 to August 2018 without any language restriction. We selected the year 1990 as the starting date of our search, since computer usage was limited to basic functions before that time. The search strategy was developed for MEDLINE and was later adapted to the other databases. Multimedia Appendix 1 presents the MEDLINE search strategy. We also searched the International Clinical Trials Registry Platform Search Portal and the Current Controlled Trials metaRegister of Controlled Trials for unpublished clinical trials to try to mitigate publication bias. Finally, we will examine the reference lists of all included studies and relevant systematic reviews, as well as perform a hand search of relevant journals. In the event that data retrieved from the published studies is incomplete or missing, we plan to contact the study authors to request clarification. The search results from all databases have been imported into a single EndNote X8.1 (Clarivate) library and duplicate records were removed.

Two authors will work in parallel to screen titles and abstracts to identify studies for full-text revision. The full-text versions of selected articles will be retrieved and assessed by two reviewers working independently. Reviewers’ individual results at each step of the screening process will be compared; disagreements will be settled between them or through an arbiter if an agreement cannot be reached. The steps of the screening process will be presented in a flow diagram according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement [39], including the reasons for exclusion of papers at the full-text screening stage.

Two reviewers, working independently, will extract the data for all included studies using a Microsoft Excel prepiloted, standardized data recording form used by the group in other digital education systematic reviews. The information to be extracted includes the following: study design and participants’ demographics, type of digital education intervention, method and device used to deliver the intervention, and type of content (eg, images, text or video, and reported outcomes). Disagreements between the authors will be resolved through consensus or consultation with a third review author, considered the arbiter.

The methodological quality of included papers will be assessed in parallel by two authors using the Risk of Bias tool from Cochrane [40]. We will assess the following domains: random sequence generation; allocation sequence concealment; blinding of outcome assessment; completeness of outcome data; selective outcome reporting; and other sources of bias such as baseline imbalance, inappropriate administration of an intervention, or contamination. We will not assess blinding of participants or personnel, as the nature of the intervention precludes blinding. If we include cluster RCTs, the assessment will include the following: recruitment bias, baseline imbalance, loss of clusters, incorrect analysis, and comparability with individually randomized trials. Each parameter will be classified as high, low, or unclear risk of bias, using the words yes, no, or unclear and the colors red, green, and yellow, respectively. We will report the results using a risk-of-bias table or summary as per the Cochrane Handbook for Systematic Reviews of Interventions [41].

We will use Cochrane’s RevMan 5.3, the software used for preparing and maintaining Cochrane Reviews, to analyze the data. To estimate the effect size of the digital education interventions in the primary study, we will first calculate the mean difference and 95% CI if the results are reported as a continuous variable. We will calculate the risk ratio and 95% CI when the study reports the outcome as a dichotomous variable. If the same outcome is reported by more than one study using different measurement tools, we will recalculate mean differences into standardized mean differences. In the event that an RCT presents more than one intervention arm, the relevant digital education arm will be compared with the least-active control arm. If cluster RCTs are included in this review, we will aim to obtain data at the student level. If that is not possible, we will first establish if the original analysis had been adjusted for the effects of clustering and, if so, we will extract and use the reported estimates. Otherwise, we will check for unit of analysis errors and we will attempt to reanalyze the data using the appropriate unit of analysis and account for the intraclass correlation coefficients [42].

When a primary study is missing relevant outcome data, we will attempt to obtain the information by contacting the study authors. If a response is not obtained, we will report it accordingly. We will not impute any missing outcome data. We will, whenever possible, conduct analyses on an intention-to-treat basis.

If the systematic review includes more than 10 studies, we will assess publication bias through a qualitative analysis of the characteristics of included studies using a funnel plot and regression weighted by the inverse of the pooled variance [43].

If the characteristics of the included studies allow, we will attempt to perform a meta-analysis. To proceed to the analysis, we will group the articles according to study design and type of intervention. We will categorize the studies’ outcomes as per Miller’s classification of clinical competence [44] to assess learners’ knowledge and skills in accordance to the type of assessment utilized (eg, if an outcome reported as skill is assessed by a knowledge test, we will consider the outcome as knowledge), independently of the teaching method. Leaners’ attitudes will be divided into cognitive, behavioral, or affective attitudes and analyzed independently [45]. Learners’ satisfaction will be reported in a narrative synthesis.

Before attempting the meta-analysis, we will assess if it is feasible by evaluating the included studies for methodological and statistical heterogeneity. We will assess the characteristics of the forest plot and calculate the I² statistic [37]. If we observe substantial heterogeneity (ie, I² greater than 0.5), we will explore its causes by conducting subgroup analysis. If extensive clinical or methodological heterogeneity is identified, we will not report a meta-analysis, but will instead use a narrative synthesis.

If a meta-analysis is possible, we will use a random-effects model, as it provides a more conservative estimate of effect and it is the preferred method when there is moderate heterogeneity. We will perform separate analyses for interventions among pre- and postregistration health care professionals. We will include the intention-to-treat analysis of the results in the meta-analysis.

To examine the impact of bias on study outcomes and in the results of the meta-analysis, we plan to perform sensitivity analyses. We will exclude studies according to the criteria in Textbox 3.

If the data allow, we plan to conduct subgroup analyses, stratifying the data as described in Textbox 4.

We will produce a narrative synthesis of results, even if a meta-analysis is not possible. The report will include a Summary of Findings table following the Cochrane Handbook for Systematic Reviews of Interventions [41]. The table will outline the main outcomes for each included study and, if a meta-analysis is possible, the table will present the results for each of the primary outcomes, as well as potential adverse effects, if they were reported by the primary studies.