Weight Down Lifestyle

Introduction

Diabetes is one of the biggest global public health concerns affecting an estimated 425 million adults worldwide, and this number is expected to rise to 629 million by 2045 []. This is coupled with a shortage of health care professionals competent in delivering high-quality diabetes care [,]. Enhancing both the size and competencies of health care professionals is a priority and improving health professions education is seen as one of the key strategies to this end []. Digital education, broadly defined as the use of digital technology in education, has been recognized as having the potential to improve health professions education by making it scalable, interactive, personalized, global, and cost-effective [-].

Past systematic reviews on digital education have focused mainly on diabetes self-management education for patients, showing an improvement in patients’ knowledge and outcomes [-]. The effectiveness of digital education interventions for health care professionals on diabetes management is still unknown []. To address this gap, we performed a systematic review to evaluate the effect of digital education on diabetes management on health care professionals’ knowledge, skills, attitudes, competencies, and behaviors, as well as its impact on patient outcomes.

Methods

Systematic Review Guidance

We followed the Cochrane Handbook of Systematic Reviews for our methodology [] and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement for reporting []. For a detailed description of the methodology, please refer to the study by Car et al [].

Data Sources and Searches

This review is part of an evidence-synthesis initiative on digital health professions education, where an extensive search strategy was developed for a series of systematic reviews on different modalities of digital health education for health care professionals (see ) []. The following databases were searched from January 1990 to August 2017:

1. The Cochrane Central Register of Controlled Trials (The Cochrane Library,)
2. MEDLINE (Ovid)
3. EMBASE (Elsevier)
4. PsycINFO (Ovid)
5. Educational Resource Information Centre (ERIC; Ovid)
6. Cumulative Index to Nursing and Allied Health Literature (CINAHL; EBSCO)
7. Web of Science Core Collection (Clarivate analytics).

We included studies in all languages and at all stages of publication. Our search strategy included gray literature sources such as Google scholar, trial registries, theses, dissertations, and academic reports. The citations retrieved from different sources were combined into a single library and screened by 2 authors independently. We also screened references of included papers for potentially eligible studies. Discrepancies and disagreements were resolved through discussion until a consensus was reached.

Study Selection

We included randomized controlled trials (RCTs), cluster RCTs, and quasi-RCTs and excluded cross-over trials due to high likelihood of a carry-over effect in this type of studies []. Studies on pre- or postregistration health care professionals taking part in digital education interventions on diabetes management were considered eligible. We defined health care professionals in line with the Health Field of Education and Training (091) in the International Standard Classification of Education []. Studies on digital education on both type 1 and type 2 diabetes at all educational levels were included.

We defined digital education as any teaching and learning that occurs by means of digital technologies. We considered eligible all digital education modalities, including offline and online education, Serious Gaming and Gamification, Massive Open Online Courses, Virtual Reality Environments, Virtual Patient Simulations, Psychomotor Skills Trainers, and mobile learning. Eligible comparisons were traditional, blended, or another form of digital education intervention on diabetes management. Traditional education was defined as any teaching and learning taking place via nondigital educational material (eg, textbooks) or in-person human interaction (eg, lecture or seminar). Traditional education also included usual learning, for example, usual revisions as well as on-the-job learning without a specific intervention in postregistration health care professionals. Blended education was defined as the act of teaching and learning that combines aspects of traditional and digital education. Eligible primary outcomes measured using any validated and non-validated instruments were knowledge, skills, competencies, attitudes, and satisfaction. Eligible attitudes-related outcomes comprised all attitudes toward patients, new clinical knowledge, skills, and changes to clinical practice.

Eligible secondary outcomes included patient outcomes in studies on postregistration health care professionals (eg, patients’ blood pressure, blood glucose, and blood lipid levels), change in health care professional’s behavior (ie, treatment intensification, defined as an intensity or dose increase of an existing treatment or the addition of a new treatment/class of medication), and economic impact of the intervention.

Data Extraction

In this study, 2 authors independently extracted data from studies using a structured and piloted data extraction form. We extracted information on study design, participants’ demographics, type, content and delivery of digital education, and information pertinent to the intervention. Study authors were contacted in case of unclear or missing information.

Risk of Bias and Quality of Evidence Assessment

The methodological quality of included RCTs was independently assessed by 2 authors using the Cochrane Risk of Bias Tool []. The risk of bias assessment was piloted between the reviewers, and we contacted study authors in case of any unclear or missing information. We assessed the risk of bias in included RCTs for the following domains: (1) random sequence generation; (2) allocation concealment; (3) blinding of participants to the intervention; (4) blinding of outcome assessment; (5) attrition; (6) selective reporting; and (7) other sources of bias []. Cluster RCTs were assessed using 5 additional domains: (1) recruitment bias; (2) baseline imbalance; (3) loss of clusters; (4) incorrect analysis; and (5) comparability with individually randomized trials [].

Data Synthesis and Analysis

In line with Miller’s classification, a learning model for assessment of clinical competence [], we classified outcomes based on the type of outcome measurement instruments used in the study. For example, multiple-choice questionnaires were classified as assessing knowledge and objective structured clinical examinations as assessing participants’ skills.

Although some studies reported change scores, we presented only postintervention data as those were more commonly reported and to ensure consistency and comparability of findings. Continuous outcomes are presented using mean difference (for outcomes measured using the same measurement tool), standardized mean difference (SMD; for outcomes measured using diverse measurement tools), and 95% CIs. Dichotomous outcomes are presented using risk ratios (RRs) and 95% CIs. As we were unable to identify a clinically meaningful interpretation of effect size in the literature for digital education interventions, we interpreted the effect size using Cohen rule of thumb with SMD greater than or equal to 0.2 representing a small effect, SMD greater than or equal to 0.5 a moderate effect, and SMD greater than or equal to 0.8 a large effect [,]. In studies that reported more than one measure for each outcome, the primary measure, as defined by the primary study authors, was considered.

Heterogeneity and Subgroup Analyses

Heterogeneity was assessed qualitatively using information relating to participants, interventions, controls, and outcomes as well as statistically using the I2 statistic for outcomes allowing for pooled analysis []. Due to substantial methodological, clinical, and statistical heterogeneity (I2>50%), we conducted a narrative synthesis according to type of comparison, that is, (1) digital education versus traditional education, (2) digital education versus blended education, and (3) one digital education type versus another digital education type. Subgroup analyses were not feasible owing to the small number of studies and limited information. We presented the study findings in a forest plot using the random effects model and standardized mean difference as the measurement scales were different and without the pooled estimates.

Results

Included Studies

Our search strategy for a series of systematic reviews focusing on different digital health professions education modalities yielded 30,532 unique references. We removed 459 duplicates, and upon screening of titles and abstracts, the screening excluded 30,050 citations. We identified 23 potentially eligible studies for which we retrieved and screened full texts. Of these, we included 12 studies: 9 RCTs and 3 cluster RCTs, all published in English (). Moreover, 1 study was reported by 3 journal papers [-]. Although presented as a cluster RCT, this study included randomization at the individual, physician level and was therefore considered an RCT. A total of 9 studies were excluded due to ineligible study design (n=3), missing data (n=5), and ineligible participants (n=1; ).

Participant Characteristics

There were 2263 health care professionals in 12 included studies [-]. A third of the studies included less than 50 participants. The study with 3 published reports had 1182 patient records as a measure of clinical outcomes [-]. Only 1 study targeted pediatric patients with type 1 diabetes []. All other studies reporting patient outcomes focused on adult patients with type 2 diabetes. A total of 8 studies focused on doctors [,,,,,-]. Moreover, 1 study each focused on medical students [], pharmacy students [], nurses [], and jointly on doctors, nurses, and dietitians [].

Study Characteristics

A total of 10 studies were conducted in high-income countries including Australia [], the United States [,,,,-], and the United Kingdom [,]. A total of 2 studies were conducted in middle-income countries such as Thailand [] and Brazil [] each.

A total of 6 studies compared digital education with traditional education [,,,,,]. A total of 3 studies compared 2 different methods of digital education interventions [,,], 2 compared blended education with usual education [,], and 1 study with 3 arms compared usual, blended, and digital education []. Only 4 studies reported duration of the intervention lasting from an hour to 2 weeks [,,,].

Various types of modalities were used to deliver the digital education interventions. A total of 3 studies used a Web-based or online portal [,,]; 3 used a scenario-based simulation software [,,]; 1 study each assessed high-fidelity mannequins []; an online game app on the computer []; periodic email reminders on the lecture content []; personal digital assistant–delivered learning materials []; and a computer-based diabetes management program [].