Introduction
The present paper is dedicated to a critical overview of a systematic review “Evaluation of the Clinical Outcomes of Telehealth for Managing Diabetes: A PRISMA-Compliant Meta-Analysis” by Wu et al. (2018) published in a high-quality, open access journal Medicine. The major purpose of this study was to compare existing research evidence on the outcomes of telehealth in diabetes management versus the usual diabetes care. The researchers employed a comprehensive search strategy to locate the final sample of 19 randomized controlled trials (RCTs), which they consequently appraised and analyzed. To report their findings, Wu et al. (2018) utilized the PRISMA framework. PRISMA stands for Preferred Reporting Items for Systematic Reviews and Meta-Analyses and implies the development of a flow diagram reflecting each step that researchers took during the search of sources for their studies. The main purpose of this framework is to ensure the methodological quality and rigor of systematic reviews (Pati & Lorusso, 2017). Thus, by evaluating the methods described in the selected study, it will be possible to identify whether the findings obtained by Wu et al. (2018) are valid and trustworthy.
Research Question
The articulation of the right research question is an important part of locating the literature for systematic reviews. The formulated clinical question helps to narrow down the search by specifying inclusion and exclusion criteria and, consequently, to identify sources that would provide a precise answer on the matter of interest. The PICO model for the development of research questions is considered to be effective in this regard. It requires researchers to choose a particular patient/problem (P), intervention/exposure (I), comparison intervention/exposure (C), and clinical outcome of interest (O) (Eriksen & Frandsen, 2018). This model can direct researchers during the search process and prompt the selection of keywords.
It is apparent that in their systematic review, Wu et al. (2018) employed the PICO framework. The components of their clinical question included adult patients with diabetes (P), telehealth diabetes management (I), and conventional diabetes management (C). The last element in the question, (O), comprised several measures: changes in hemoglobin A1c (HbA1c) as the primary outcome and improvements in patients’ blood pressure, blood lipids, body mass index, and quality of life as the secondary outcomes. Overall, even though Wu et al. (2018) focused on several outcomes at once, their question may be deemed efficient as it clearly identified all the necessary aspects of the clinical problem. This PICO question helped them to obtain an answer that is precise and not too broad.
Search Strategy
Nowadays, researchers can locate necessary information by using a great variety of means, including offline and online libraries, professional consultation, web-based search engines, and so forth. However, when conducting a study on a medical issue, the use of specialized electronic research databases may provide serious advantages since they offer opportunities and tools for a more targeted search of digitalized and indexed materials. Even though a single evidence-based database may store tens of thousands of literature pieces, Grewal, Kataria, and Dhawan (2016) state that researchers must always utilize multiple databases in order to locate the best possible options. The use of additional methods, such as the scanning of reference lists in located articles and the seeking of advice from experts in the field of interest can allow advancing and improving the search outcomes as well.
Taking into account the abovementioned observations, it is valid to say that the search for relevant studies carried out by Wu et al. (2018) was detailed enough. Among the strategies that Wu et al. (2018) mentioned in their methodology section were database searches, searches of related journals, and reference tracking. When employing the former strategy, they used five databases, including MEDLINE, PsycINFO, PubMed, EMBASE, and CINAHL. As noted by Grewal, Kataria, and Dhawan (2016), MEDLINE, PubMed, EMBASE, and CENTRAL are comprise a minimum list of open access databases that must be included in the search. It is possible to say that Wu et al. (2018) fulfilled and even exceeded that basic requirement. By exploring two closed access databases (PsycINFO and CINAHL) as well they could substantially improve the search outcomes.
As the search flow diagram provided in the study reveals, the initial search through databases allowed Wu et al. (2018) to locate 2044 articles. In addition, they managed to identify 43 other articles by using manual search through key journals and reference lists in the relevant literature. Even though the researchers did not state whether they attempted to find non-published materials by contacting professional experts, this overlook does not affect the quality of their search strategy severely. The employed methodology was exhaustive and provided the researchers with a chance to screen a substantial number of the initial record sample.
Quality of Primary Studies
Quality assessment is a core step in the process of the systematic literature review. According to Boren and Moxley (2015), basic criteria for quality assessment usually include “appropriateness of study design to the research objective, risk of bias, generalizability, statistical issues, quality of the intervention, and quality of reporting” (p. 59). It is also valid to note that the overall research design employed in primary sources largely defines the quality of findings reported in them.
RCT was the only type of study included by Wu et al. (2018) in their review. Methodologically sound RCTs are considered to provide the finest-quality evidence since they refer to comparative effectiveness research that allows learning which treatment methods and interventions are better than others. To ensure that the selected studies are methodologically sound, Wu et al. (2018) assessed them based on such criteria as random sequence generation and allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attribution bias), and selective reporting (reporting bias). Among them, the categories of performance and detection bias received the lowest scores. However, as Wu et al. (2018) noted, it was impossible for the patients to be blinded to their allocation, due to the very nature of the studied subject (telehealth). Thus, the risk of bias in this regard was high in most of the located RCTs. In all other aspects, the 19 selected studies proved to be methodologically sound.
Wu et al. (2018) also reported the major characteristics of the RCTs. The size of the samples in them ranged from 100 to 1665 participants. All populations comprised adults who, however, lived in distinct areas of Europe and the United States. Importantly, interventions analyzed in the chosen studies had similar goals and included such practices as self-monitoring of blood glucose, electronic or manual data transmission, and feedback. Nevertheless, in five of the RCTs, intervention procedures were not sufficiently described.
It was important for Wu et al. (2018) to consider all differences and similarities in the studies in order to explain variations in findings and detect any methodological deficiencies. For example, evidence in all of the reviewed RCTs revealed that telehealth can be more effective than regular diabetes care in terms of controlling the glycemic index in patients and reducing blood pressure. However, Wu et al. (2018) also found that the degree of effectiveness could be positively correlated with the regularity and overall patterns of data delivery through telehealth, the strength and methods of intervention, and the baseline levels of biological indicators in patients exposed to interventions. Since the analysis of these factors was not in the initial intentions of the selected systematic review, Wu et al. (2018) could not make valid conclusions regarding this matter. However, the conducted investigation allowed providing a high-quality answer to the main clinical question about the effectiveness of telehealth for diabetes management.
Conclusion
Summing up the critique, it is possible to state that the evaluated study is of excellent quality. The researchers utilized a rigorous methodology and recorded in detail all the processes and steps they performed while searching for relevant RCTs and during the quality assessment. Besides that, they acknowledged all possible biases and inconsistencies in methodologies and characteristics of the studies chosen for the review, which helped them to interpret results more critically and realistically. Based on this, the findings of research by Wu et al. (2018) are credible and trustworthy and provide useful clinical data for evidence-based decision-making and practice. Another important contribution of this systematic review is in the identification of gaps in the existing literature about the factors that could be interrelated with the level of telehealth effectiveness. Thus, it prompts ideas for future research and supports further improvement of medical knowledge.
References
Boren, S. A., & Moxley, D. (2015). Systematically reviewing the literature: Building the evidence for health care quality. Missouri Medicine, 112(1), 58-62.
Eriksen, M. B., & Frandsen, T. F. (2018). The impact of patient, intervention, comparison, outcome (PICO) as a search strategy tool on literature search quality: A systematic review. Journal of the Medical Library Association: JMLA, 106(4), 420-431.
Grewal, A., Kataria, H., & Dhawan, I. (2016). Literature search for research planning and identification of research problem. Indian Journal of Anaesthesia, 60(9), 635-639.
Pati, D., & Lorusso, L. N. (2017). How to write a systematic review of the literature. HERD: Health Environments Research & Design Journal, 11(1), 15-30.
Wu, C., Wu, Z., Yang, L., Zhu, W., Zhang, M., Zhu, Q., … Pan, Y. (2018). Evaluation of the clinical outcomes of telehealth for managing diabetes: A PRISMA-compliant meta-analysis. Medicine, 97(43). Web.