**3. Results**

Figure 1 illustrates the searching process. Among the 327,405 papers on QOL, 700 papers on interventions to improve QOL of people with diabetes were selected as eligible.

Table 1 shows that in the period 1990–2018, there were a total of 700 papers published about interventions to improve the QOL among people with diabetes. From the first paper counted in 1990, the volume of annual articles increased significantly over time to reach a peak of 67 papers in 2015, before falling to 63 papers by the end of 2018. Papers published in 2005 and 1998 had the highest mean citation rate per year (9.7 and 9.0, respectively). Articles published in 2015 and 2005 had the highest total usage (i.e., the total number of downloads) and mean use rate in the last five years, respectively. In the last six months, the total usage and the mean use rate of papers published in 2018 were higher than those published in other years.

**Figure 1.** Selection of papers.

Overall, 700 papers were published by authors in 61 countries. The United States (U.S.) and the United Kingdom (U.K.) contributed the most publications (322 and 160 papers, respectively), followed by Germany (101 papers), Australia (87 papers), and the Netherlands (85 papers). Only China was the only middle-income country in the top ten nations with the greatest number of publications, while the others were high-income nations. Meanwhile, among the top 20 countries, only China, India, Malaysia, and Iran were middle-income countries.


**Table 1.** General characteristics of publications analyzed.

<sup>1</sup> Total usage: Total number of downloads. <sup>2</sup> Use rate: Total number of downloads/total number of papers.

Figure 2 depicts networks of collaboration among the top 20 countries having the highest volume of publications. The U.K. and the U.S. had the highest amount of collaborations, with 28 and 25 countries, correspondingly. In the U.S., among 322 published papers, there were more than 610 affiliations mentioned. Approximately 80% of them were from U.S. authors (~490 affiliations), 4% were from the U.K. (~23 affiliations), and 2% were from Denmark (~10 affiliations). Similarly, in the U.K., 160 papers were products of authors from 290 affiliations (or organizations), with ~58% from the U.K., 8% from the U.S. and 6% from Australia. These countries were followed by Germany (22 countries), Denmark (18 countries), and the Netherlands (18 countries).

**Figure 2.** Collaboration network between the top 20 countries by the number of publications. The outer rim reflects the volume of collaborations between each country and the other countries in the top 20, showing collaboration among countries. Abbreviation: USA, the United States of America; ENG, England; GER, Germany; AUS, Australia; NET, the Netherlands; CAN, Canada; DEN, Denmark; ITA, Italy; SWE, Sweden; CHI, China; IND, India; SWI, Switzerland, BEL, Belgium; SPA, Spain; MAL, Malaysia; FRA, France; JAP, Japan; SCO, Scotland; IRA, Iran; SIN, Singapore.

To illustrate the scope of the selected studies, we performed a content analysis to evaluate the co-occurrence of the most frequent terms in the abstracts and titles. Figure 3 shows four major clusters emerging from the 271 most common keywords with a co-occurrence rate of at least 15: (1) the yellow cluster refers to clinical trials to test the efficacy and safety of drugs, as well as to control pain and overweight/obesity among people with diabetes (e.g., placebo, efficacy, safety, dose, pain, body weight, obesity, weight loss); (2) the blue cluster indicates trials focusing on diabetes prevention such as lifestyleor behavior-related interventions (e.g., physical activity, exercise training, diet, etc.); (3) the green cluster covers the interventions using insulin-related therapies to control blood glucose level, particularly among patients with type 1 diabetes (e.g., insulin treatment, insulin glargine, etc.); and (4) the red cluster refers to interventions in the community to promote self-care capability (e.g., self-care, practice, self-efficacy, caregivers, etc.), as well as reduce the risk of psychological problems (e.g., depression or distress) among people with diabetes, especially in adolescents and children (e.g., adolescent, child, parent).

**Figure 3.** Co-occurrence of most frequent terms in titles and abstracts. The colors of the nodes indicate principal components of the data structure; the node size was scaled to the keywords' occurrences; the thickness of the lines was drawn based on the strength of the association between two keywords.

The top 10 clinical trials having the highest number of citations are presented in Table 2. All of them were RCTs, which focused mainly on the effectiveness of different medications, surgery, or behavioral therapies in controlling blood glucose; reducing the negative effects of diabetic complications (such as diabetic neuropathy, obesity, limb ischemia, or perinatal complications); and improving QOL in people with diabetes, especially pregnant women, children, and adolescents. Among the top three papers, the first paper, by Crowther et al., titled "Effect of treatment of gestational diabetes mellitus on pregnancy outcomes," published in 2005 in the *New England Journal of Medicine,* had the highest number of citations and evaluated the efficacy of dietary advice, blood glucose monitoring, and insulin therapy (by comparing the intervention group to a control group receiving only routine care) in treating gestational diabetes in pregnant women to prevent perinatal complications [32]. By using the Short-form 36 (SF-36) instrument, the results of this study showed a significant increase in QOL among participants [32]. The second paper, titled "Gabapentin for the symptomatic treatment of painful neuropathy in patients with diabetes mellitus—A randomized controlled trial," also used SF-36 and showed that gabapentin use significantly improved the QOL of patients with diabetic neuropathy compared to the placebo group [33]. The SF-36 was also used in the third-most cited paper, entitled "Bariatric Surgery versus Intensive Medical Therapy for Diabetes-3-Year Outcomes" [34]. The authors of this study indicated that bariatric surgery significantly enhanced health-related QOL among obese patients with type 2 diabetes compared to those receiving intensive medical therapy only [34].



RCT: Randomized controlled trials, SF-36 = Short-form 36, WHOQOL-BREF = WHO Quality of Life-BREF, MOS = Medical Outcomes Studies; IWQOL-LITE = Impact of Weight on Quality of Life-LITE.

Figure 4 presents the hierarchical clustering of major research disciplines in interventions aiming to improve the QOL of people with type 2 diabetes. The horizontal axis reflects the dissimilarity between clusters, while the vertical axis reveals disciplines of pooled papers. This figure shows that interventions were separated into five major clusters. In the first cluster, the "Pediatrics" grouping is connected to "Endocrinology and Metabolism," suggesting that the majority of interventions for children with diabetes and adolescents were concentrated on using medications to enhance metabolic pathways such as insulin infusion or an insulin pump.

Meanwhile, the second cluster indicated clinical and lifestyle interventions to improve QOL in people with diabetes in general or with multiple comorbidities (such as cardiac disease and mental problems). For example, the "Pharmacology and Pharmacy" grouping was joined with "Medicine, Research and Experiment," which indicated that interventions using a pharmacological approach such as fenofibrate, ranolazine, or other drugs in glycemic control, reduced the damage done by diabetes and its complications and improved QOL. In addition, the "Sport Sciences" discipline was combined with the "Geriatrics and Gerontology" disciplines, showing that the common intervention approaches in older people with diabetes were facilitating physical activity. Similarly, for people with diabetes suffering cardiac illnesses ("Cardiac and Cardiovascular System"), "Surgery" and "Nutrition and Dietetics" were the two most common interventions.

The third cluster reveals clinical interventions to improve the QOL of people with diabetes in primary care settings. These included home- or community-based interventions.

Meanwhile, the fourth cluster indicated clinical interventions to enhance the QOL of people with diabetes suffering from neurological pain.

Finally, the fifth cluster showed public health, health service, health policy, and medical-information-technology-related interventions aiming to improve the QOL of people with diabetes. Notably, this cluster is not close to other groupings in the figure.

**Figure 4.** Dendrogram of coincidence of research areas using WOS classifications.

By using LDA, we categorized the selected interventions under ten major topic headings. The three topics with the highest number of publications included: (1) community-, family-, and eHealth-based interventions to improve self-management and self-efficacy; (2) lifestyle (e.g., physical activity and dietary) interventions in people with diabetes; (3) interventions focused on comorbidities in people with diabetes (Table 3). The first topic was investigated in 187 papers, accounting for 26.7% of total papers. Meanwhile, the second and third topics accounted for 17.7% and 15.3%, respectively.


Figure 5 shows the correlation of publications to the ten topics. Recently, topics 1 and 2 received the greatest attention, with the largest number of publications every year in the three years 2015–2018. In the same period, the number of publications on topics 3 and 4 also increased, while the volume of publications on other topics decreased.

**Figure 5.** Changes in research topic development.
