A total of 5,001 participants were included across the studies. The mean age was 46.01 years (SD 10.56), ranging from 18 to 84 years, indicating a predominance of middle-aged adults. Gender was inconsistently reported, with 1,813 participants (36.2%) identified as male and 1,007 (20.1%) as female. Diabetes type was not consistently described; among those with available data, 298 participants (6.0%) had type 1 diabetes and 1,729 (34.6%) had type 2 diabetes. Glycemic control was reported in a subset of studies, with a mean value of 56.91 mmol/mol (SD 14.93). Mean BMI was 30.0 kg/m² (SD 5.33), suggesting most participants were in the obese range. Mean weight was 80.76 kg (SD 15.37), and mean height was 166.38 cm (SD 8.46). This demographics information is presented in Table 2. Demographic variables such as gender and diabetes classification were frequently underreported, limiting subgroup analyses and the ability to assess representativeness across study populations.

Assessment of study origin by country indicated that the Netherlands is identified as the leading contributor, with 20 studies conducted in this setting [15, 27–45]. The United Kingdom [12, 46–53] and Australia [13, 54–60], each contributed 8 studies respectively. The United States contributed with 7 studies [12, 53, 61–65], while Spain has 5 studies [66–70]. Iran [52, 65, 71, 72], China [73–76], and Thailand [77–80] contributed 4 studies each. Germany [38, 81, 82], Malta [83–85], and Sweden [14, 86, 87] each contributed 3 studies. Several countries had 2 studies each, including India [88, 89], Italy [90, 91], Portugal [92, 93], and Egypt [94, 95]. The following countries contributed one study each: Jordan [96], Canada [97], Indonesia [98], Colombia [99], Singapore [100], Israel [101], and Fiji [102].

Four studies were conducted across multiple countries. One study was conducted between the Netherlands and Germany [38], another study was jointly conducted by the United States and the United Kingdom [12], two studies were conducted in collaboration between Iran and the United Kingdom [52, 65], and one study was conducted in United Kingdom and Germany [82]. These overlaps highlight a small but important trend toward international collaboration in technology-enabled diabetes and mental health care research.

Figure 2 presents the geographic distribution of study origins, revealing a clear imbalance between developed and developing countries. Western nations such as the Netherlands, the United Kingdom, and Australia dominate research contributions. In contrast, only a few developing countries are represented, and their contributions remain limited. This underscores the need for greater support and inclusion of developing nations in research within this field.

Figure 3 summarizes studies based on the study designs. The most common design was the cross-sectional study (n=20), capturing data on exposure and outcome at a single time point. This design was used predominantly for prevalence estimates and behavior assessments [14, 15, 31, 37, 39, 41, 43–45, 58, 61, 68, 69, 76, 80, 84, 87, 90, 96, 103]. Randomized controlled trials (RCTs) were used by 12 studies, offering randomization and control groups to test intervention efficacy [27, 38, 46, 47, 49, 56, 62, 66, 86, 93–95]. A total of 14 studies were observational, where outcomes were tracked without intervention, often in real-world or longitudinal settings [12, 28, 29, 32, 40, 42, 55, 59, 67, 70, 73, 81, 85, 89].

Quasi-experimental study designs were utilized by nine studies, providing controlled conditions without randomization and were typically used to assess device or footwear effects across time or conditions [48, 51, 60, 65, 72, 74, 77, 79, 98]. Seven studies utilized qualitative assessments methodologies such as interpretative phenomenological analysis and thematic coding to explore patient perspectives on footwear adherence, usability, and self-management [13, 36, 50, 75, 97, 100, 102].

A group of randomized crossover experimental studies (n=7) used within-subject designs with randomized order of conditions, allowing participants to serve as their own control. This study design was a common approach in biomechanics and gait research [33, 35, 53, 78, 82, 99, 101]. Smaller and preliminary investigations included pilot studies (n=5), which focused on early-stage feasibility and effect trends [34, 52, 71, 88, 91], and feasibility studies (n=2), which tested practical implementation without assessing outcomes [63, 64].

Two studies used a pre-clinical mixed-methods design, combining qualitative and quantitative methodologies data across iterative phases involving usability, validation, and device refinement [57, 92]. One exploratory qualitative study [83] focused on user experience in context, and one retrospective clinical audit [54] reviewed historical clinical data to assess practice patterns across expected standards.

The included studies represent a methodologically diverse body of evidence, spanning from exploratory, hypothesis-generating research to confirmatory randomized controlled trials. This diversity reflects the evolution of a maturing research field that increasingly integrates clinical effectiveness with user-centered design, implementation science, and contextually informed evaluation approaches.

The studies utilized a diverse set of outcomes to examine the effectiveness of the footwear interventions, as shown in Table 3. Among the studies investigating therapeutic footwear and insoles for people with diabetes, plantar pressure was the most common outcome, reported in 55 studies. Plantar pressure was assessed using peak plantar pressure (PPP) in 40 studies [29, 31, 33, 35, 37–39, 41–45, 47–49, 53, 55, 57, 58, 60, 63, 64, 69, 74, 77–79, 82, 84–86, 88, 90–92, 95, 96, 98, 99, 104], and pressure time integral (PTI) in 13 studies [37, 58, 60, 64, 69, 74, 77–79, 84, 85, 95, 104]. PPP evaluates the maximum pressure in the affected region while PTI quantifies the cumulative pressure applied over time to a specific area. Only one study reported more advanced plantar pressure metrics such as peak pressure gradient (PGrad), peak pressure time curve (PTC), peak pressure map (PMap), and peak pressure time map (PTM) [44], highlighting the limited adoption of these multidimensional approaches in routine clinical or research evaluation.

Adherence was the second most common outcome reported across the studies (n=27). Self-reported adherence was used in 18 studies [13–15, 30, 38, 45, 50, 55, 67, 75, 83, 86, 87, 93, 94, 96, 102, 105] through interviews, questionnaires, or visual analog scale (VAS) ratings. Objective monitoring techniques were applied in eight studies [15, 27, 28, 40, 42, 49, 81, 96], where temperature sensors were used to assess mean daily wear time and step count. In addition, one study [12] employed an activity monitor to capture physical activity data. Together, these measures reflect increasing efforts to quantify real-world device usage accurately.

Gait and balance were evaluated in 20 studies, with 15 studies analyzing gait and postural parameters such as double support time, stride length, sway velocity, and stability indices [32, 36, 39, 52, 61, 65, 69, 71, 72, 76, 89, 92, 93, 101, 105]. Additionally, center of pressure (COP) metrics was reported in five studies [39, 51, 78, 90, 101], assessing foot trajectory, velocity, and foot-off signals as markers of dynamic and static balance.

Patient satisfaction and preferences were reported in 19 studies using a combination of questionnaires like QUEST 2.0, structured interviews, and subjective assessments of fit, comfort, aesthetics, and ease of use [30, 31, 33, 36, 41, 43, 55, 57, 58, 68, 69, 75, 86–88, 92, 100, 101, 105]. Ulcer incidence or recurrence was reported in 13 studies [28, 49, 58, 62, 64, 66, 67, 70, 73, 81, 94, 95, 105], using clinical data to track reulceration frequency, time to recurrence, and number of affected patients.

Other outcomes were reported less frequently. Quality of life and capability were assessed in five studies [47, 92, 93, 97, 101] via validated tools such as EQ-5D-5L and SF-36. Contact area was measured in five studies [76–79, 90], and ambulatory activity was assessed in four studies [15, 32, 38, 91] through stair navigation time, Timed Up and Go (TUG), stride count and using an activity monitor [15, 32, 38, 91]. Ground reaction force [39, 89, 90], plantar temperature [61, 63, 99], and healing rates [38, 59, 94] were each reported in three studies. A small number of studies examined insole compression [31, 96], functional tests [93, 101], adverse events such as amputation or mortality [73, 87], foot measurements [76], and insole aging [86], each appearing in only 1–2 studies.

The use of diverse outcome measures suggests a prevailing research focus on biomechanical parameters, particularly PPP, while also demonstrating increasing attention to adherence monitoring, patient satisfaction, and clinical outcomes such as ulcer recurrence and healing. Notably, recent studies have incorporated adherence as a secondary outcome to evaluate the effectiveness of therapeutic footwear and insoles. The integration of both quantitative and subjective measures reflects a maturing research field with a growing emphasis on real-world effectiveness and user-centered care.