The global diabetes prevalence rate for adults in 2021 was estimated to be >10.5% (n=536.6 million people), and it is expected to increase to 12.2% by 2045, with the associated health care expenditure costing an estimated US $1054 billion [1]. Diabetes-related foot disease (DFD) is a leading cause of disability worldwide [2]. It has a 5-year mortality rate and health care costs comparable with those of cancer [3], and it is associated with a low health-related quality of life [4]. The development of diabetes-related foot ulceration precedes up to 85% of nontraumatic amputations [5]. Diabetes-related complications in the lower limb, including peripheral neuropathy and peripheral arterial disease, typically precede the development of diabetes-related foot ulceration [6]. Collectively, these complications are a leading global cause of disability, hospitalization, and amputation [2]. In Australia, the ongoing effect of colonization has led to systemic racism, enduring disadvantage, trauma, political exclusion, and a health care system that is dismissive of Aboriginal health paradigms and world views [7,8]. This has perpetuated cultural safety deficits in Australian health care systems and contributed to disproportionately high rates of DFD, including a 4-fold increase in the risk of peripheral neuropathy and 5- and 6-fold increases in the risk of foot ulcer and amputation, respectively [9,10].

Evidence-based clinical practice, including the involvement of multidisciplinary teams and well-established care pathways, has been shown to reduce amputation rates in people with DFD by 50%, resulting in a significant reduction in patient and health care burden [11]. However, such care is not readily accessible to all patients [12,13]. It is well established that First Nations Peoples and non-Indigenous people living in rural and remote areas in Australia [14] and around the world [15-18] have worse overall health outcomes than their urban counterparts. This is due partly to the large geographical distance between health care services [19,20]. For First Nations Peoples in Australia, additional factors contributing to these outcomes include a lack of access to care that is culturally safe as well as a lack of workforce capacity, including difficulty in recruiting culturally capable practitioners [20,21]. This contributes to a system of care delivery that is frequently fragmented and inconsistent [22].

The use of app-based mobile health (mHealth) information systems is an emerging area of digital health that is increasingly being used to support the management of chronic conditions, including diabetes, particularly in rural and remote areas and across a variety of health systems [3]. mHealth apps are considered part of public health practice and have been shown to facilitate access to in-person local health services and other digital health technologies such as telehealth [23]. Studies have demonstrated that mHealth apps provide effective health education for lifestyle improvements such as smoking cessation and assist with self-monitoring for chronic disease [24]. A previous systematic review published in 2018 evaluated studies investigating mHealth apps for First Nations populations for any condition and found 9 studies in Australia, with the majority targeting mental health conditions (none for diabetes), with high rates of acceptability and engagement with the mHealth interventions by end users [25]. More broadly, studies in non-Indigenous populations internationally have demonstrated that mHealth apps are effective for improving diabetes-related knowledge and self-care management [26-28]. Because of the growing burden of DFD for First Nations Peoples in Australia and in the general population globally, a collective evaluation of studies investigating the use of mHealth apps designed to support diabetes-related foot health is required to help inform the design and implementation of future mHealth apps and assess their effectiveness for the assessment and management of diabetes.

Therefore, the aim of this systematic review was to identify the features of, and outcomes from using, mHealth apps for the assessment and management of diabetes-related foot health in First Nations Peoples in Australia and non-Indigenous populations globally.

This review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) [29] guidelines and registered with PROSPERO (CRD42022349087). An electronic database search of Ovid MEDLINE, Embase, CINAHL Complete, and Scopus was conducted from database inception to September 8, 2022 (Multimedia Appendix 1). At the same time, an additional search to identify any Australian First Nations–specific articles was conducted owing to the potential for some of this research to be published in First Nations–specific locations or gray literature. This consisted of an electronic database search of PubMed using specific Medical Subject Headings (MeSH) terms and Informit’s Indigenous Collection database (Multimedia Appendix 1). Differences in search terms and how terms were combined were adapted to the requirements of each database. Hand searches of gray literature were also conducted, and the sources included the Journal of the Australian Indigenous HealthInfoNet, Menzies School of Health Research, Services for Australian Rural and Remote Allied Health, and the Australian Institute of Health and Welfare. Reference lists of included studies and review articles were also searched. No language, publication date, or publication status restrictions were used.

The inclusion criteria were studies describing mHealth apps developed for the management and assessment of DFD. The studies had to include some evaluation (qualitative or quantitative) of the mHealth app (which could include usability, acceptability, feasibility, or diabetes-related foot health outcomes). Studies were excluded if they only evaluated the technical design features of an mHealth app; if the mHealth app was designed for automatic detection of wounds or ulcers or for assessing the accuracy and reliability of wound images only; or if the mHealth app incorporated additional equipment such as thermal imaging cameras, pressure mats, activity monitors, pressure monitoring insoles, glucose monitors, smart socks, or foot temperature probes. Studies were also excluded if the mHealth app was delivered or assessed by people without diabetes or was for medical professionals only. The population, intervention, comparison, and outcomes (PICO) elements for this study are included in Multimedia Appendix 2.

Two authors (SS and AS) developed and pilot-tested the electronic searches, with 1 author (AS) conducting the search. Titles, abstracts, and full texts were assessed independently by 2 authors (JG and AS). Disagreements were resolved by consensus and a third author (SS) where necessary. Study data, including publication details (author, year, and location), participant characteristics (age, sex, and diabetes information), sample size, mHealth app features, and results, were extracted by 1 author (AS) and cross-checked by another author (SS). A descriptive synthesis of included study findings was performed.

It was planned to use the Aboriginal and Torres Strait Islander Quality Appraisal Tool to assess studies that included First Nations Peoples from Australia [30]. The tool consists of 14 questions that are used to assess the cultural safety of the study. Answer options include “yes,” “partially,” “no,” and “unclear.” Broadly, questions relate to Community engagement; First Nations leadership and governance; intellectual and cultural rights; and translation of findings to policy, practice, and Community. To assess the potential risk of bias and methodological quality of the included studies, the following tools were used: the revised Cochrane risk-of-bias tool for randomized trials (RoB 2) [31] and the Health Evidence Bulletins Wales checklists for observational, cohort, and qualitative studies [32]. The RoB 2 tool assesses 5 areas of potential bias: randomization, deviations from the intended interventions, missing outcome data, measurement of the outcome, and the selection of the reported result. Each bias domain individually (and each study overall) is determined to display either a low or a high risk of bias or some concerns relating to the risk of bias. The observational study and qualitative study appraisal checklists designed by Health Evidence Bulletins Wales have been developed for a critical appraisal of observational and qualitative studies. These checklists were selected because the tools include a small number of key domains, are simple checklists rather than scales, and were developed using a variety of literature sources. The articles were rated independently by 2 authors (SS and AS), and any disagreements were resolved by consensus and a third author (VC) when necessary. There was no minimum level of quality required for inclusion in this review.

The initial database search resulted in 494 citations (after the removal of duplicates), of which 15 (3%) were appropriate for full-text review (Figure 1). Of these 15 studies, we excluded 9 (60%) for the following reasons: mHealth app only being used in clinic with researchers (n=4, 44%) [33-36], wrong study type (n=1, 11%) [37], wrong study population (n=1, 11%) [38], design and development of the mHealth app only (n=1, 11%) [39], protocol (n=1, 11%) [40], and wound images only (n=1, 11%) [41] (Figure 1). No studies specific to First Nations Peoples from Australia were identified by this review. However, of the initial 494 studies, 6 (1.2%) that assessed mHealth apps for the management of diabetes-related foot health in non-Indigenous populations from Turkey [26,28], Brazil [27], Australia [42], Indonesia [43], and the Philippines [44] were included. The included studies, with sample sizes ranging from 20 to 130 people, included a total of 361 participants (Table 1). Of the 6 studies, 3 (50%) were randomized controlled trials [26,28,43], and the other 3 (50%) used a Delphi [27], qualitative [42], and pretest-posttest [44] methodology, respectively. The follow-up time of the studies ranged from 1 to 6 months. All studies were developed in an academic or university setting, with funding or financial support provided from external organizations for 5 (83%) [27,28,42-44] of the 6 studies.

Foot care education was the main component of all mHealth apps included in this systematic review and covered topics such as daily foot care tasks, appropriate sock and shoe selection, nail care, and identification of peripheral neuropathy signs and symptoms (Table 1). Varied methods of education delivery were used by the studies, including DFD information pages (all studies), cartoon videos [26], test-yourself sections [28], mindfulness and skills coaching [43], 2D animation [27], and weekly support group activities and lectures [44]. Of the 6 mHealth apps, 2 (33%) provided functionality that allowed participants to enter health-related data [28,44], and 1 (17%) included a messaging interface [28].

Our systematic review has demonstrated short-term improvements in self-reported knowledge of, and self-care ability for, diabetes-related foot health with the use of an mHealth app. Specifically, of the 6 studies, 1 (17%) reported high levels of acceptability of the mHealth app content for self-care by people with diabetes and diabetes specialists [27] (Textbox 1), and the remaining 5 (83%) reported that participants had improved diabetes-related knowledge and self-management skills after using their mHealth app [26,28,42-44] (Textbox 1).

As no study in First Nations Peoples from Australia was identified, the Aboriginal and Torres Strait Islander Quality Appraisal Tool was not used. Of the included studies, the randomized controlled trials (3/6, 50%) displayed a risk of bias of some concerns, as assessed by the RoB 2 tool (Table S1 in Multimedia Appendix 3). The risk of bias owing to randomization was rated as some concerns mainly because of a lack of data regarding whether the allocation sequence was concealed until after enrollment. All studies used random sequence generation and reported no baseline differences between groups (except for the study by Kilic and Karadağ [28], where previous diabetic foot education was higher in the control group). The risk of bias owing to deviations from the intended interventions was rated low in all studies. Only the study by Kilic and Karadağ [28] could be considered to have some bias as a result of missing outcome data, with >5% of the participants lost to follow-up and no analyses conducted to compare those lost to follow-up with the full sample. None of the participants in the trial intervention groups were blinded to their allocation because the mHealth app was installed on their mobile phones. This raises some concerns with certain outcome measures because the knowledge of the intervention may have influenced the participants’ self-reported outcomes (diabetes self-management and self-efficacy questionnaires as well as diabetes-related foot knowledge). However, 2 (67%) of the 3 randomized controlled trials also reported additional outcome measures (clinical outcomes) that were not likely to be affected by the participants’ knowledge of the intervention received [28,43].

The Health Evidence Bulletins Wales checklist was used to assess a qualitative study [42], a pilot study [44], and a validation study [27] (Table S2 in Multimedia Appendix 3). All studies (n=3) provided detailed information regarding study population, aims, and outcomes. Although the qualitative study [42] stated that the interviews were coded using thematic analysis, minimal information regarding the analysis was presented. No details were provided regarding the researchers’ role or background, whether the interview questions were predetermined, why both interviews and focus groups were used, how many people attended the focus groups, how themes and concepts were identified in the data, and whether the participants received the interview data for confirmation and feedback. The length of follow-up within the studies (<12 wk) meant that the longer-term impact of the mHealth apps on diabetes-related knowledge and self-care management is unknown. Because of these limitations, the findings of this study should be viewed with caution.

No study investigating an mHealth app for the assessment and management of diabetes-related foot health in First Nations Peoples from Australia was identified. Of the 6 included studies in non-Indigenous populations, 1 (17%) reported high levels of acceptability of the mHealth app content for self-care by people with diabetes and diabetes specialists [27]. The remaining studies (5/6, 83%) found that participants’ self-reported diabetes-related knowledge and self-management skills improved after the use of their mHealth app [26,28,42-44]. This suggests that, over the short term (<6 mo), the use of mHealth apps by patients with diabetes results in improved diabetes-related knowledge and self-care ability. However, further research is required to investigate whether these findings translate to the prevention of DFD and whether there is an improvement in outcomes such as a reduction in ulcer healing time or the rates of amputation.

In Australia, reducing the burden of DFD in First Nations Peoples, particularly those living in rural and remote areas, is a national priority [45]. This requires a multifaceted approach that navigates the obstacles of geographical distance and internet connectivity, facilitates effective communication among health professionals, integrates with existing services, engages the patients in effective preventive care strategies, and better aligns rural and remote service delivery with Australian best practice DFD prevention and management guidelines [46,47]. Although this systematic review did not find any studies using mHealth apps for the assessment and management of diabetes-related foot health in First Nations Peoples in Australia, mHealth apps have been used in working with these populations for other health conditions, including mental health and weight management, and may be a complementary method to help reduce the disproportionate burden of DFD experienced by First Nations Peoples [25]. However, the acceptability, design, implementation, and evaluation of mHealth apps must be Community-led, embrace a co-design approach, and feature a participatory action research approach [48]. mHealth apps must be culturally safe and therefore must actively engage with, and privilege, First Nations epistemologies and ontologies [49,50] (eg, locally produced mHealth strategies that privilege First Nations voices, incorporate culturally appropriate content as judged by local Communities, and promote engagement through alternative mediums to text such as graphics and animations that have been demonstrated to have the best chance of successful implementation [25]). It is through these approaches that First Nations Communities are empowered to take control and have ownership of their own health and well-being [48,51].

In non-Indigenous populations globally, our review has demonstrated short-term improvements in self-reported knowledge of, and self-care ability for, diabetes-related foot health with the use of an mHealth app. However, the impact of these interventions on adherence to routine foot health monitoring by health professionals and DFD occurrence and outcomes have not been evaluated. This is a key recommendation from international guidelines for the prevention and management of DFD [46]. Nevertheless, mHealth apps have been used to improve clinical health outcomes for a number of chronic conditions, including depression, obesity, alcohol dependence, and diabetes [25,52]. Previous studies, including those in patients with diabetes, have focused on mHealth apps that aim to improve glycemic control, with the outcomes from a meta-analysis supporting their use in people with type 1 diabetes [53]. These studies used features similar to those of the mHealth apps developed by the studies in this systematic review for improving knowledge and self-care management of DFD, including patient education resources [26-28,42-44], the ability for patients to input data [28,44], and reminder notifications [26,44]. These features may help improve uptake and engagement of the mHealth apps; however, it is important to consider that mHealth apps may need to use a variety of features to support different types of learners.

Previous studies support the need to use mHealth apps as part of a suite of measures, including face-to-face and telehealth care [54]. It is proposed that mHealth apps should aim to incorporate functions to increase patient access to health care services and interaction with health care professionals, such as the ability to schedule appointments via the app as well as support the coordination of health care through patient data sharing. This is particularly relevant to geographically remote regions where in-person care and care across many health professions requires assistance or is not possible [55]. Inputting clinical data in mHealth apps also supports information sharing, improves efficiency by reducing redundant assessments, and supports continuity of care within and among health care professions to help ensure a patient-centered approach to the assessment and management of chronic health conditions such as diabetes [56]. In addition, there is a growing body of evidence supporting the use of nonintrusive advanced analytics to evaluate user engagement with the mHealth app [57]. This is because limited engagement has been identified as a key reason for the poor performance of mHealth apps for the management of other medical conditions [57].

Our systematic review demonstrated improved self-management of foot health and increased self-care knowledge in non-Indigenous populations with diabetes after the use of an mHealth app. These outcomes support the use of mHealth apps as an effective health promotion strategy that can increase knowledge and self-care capacity. As demonstrated by international guidelines, evidence-based care is effective for reducing the complications associated with DFD, such as foot ulcer and amputation [11,58]. On the basis of the findings of our review, mHealth apps that incorporate evidence-based health promotion and prevention education, encourage patient engagement through patient-driven upload of data, and support the coordination of care delivery among health services are required to address the growing impact of DFD, particularly in priority populations and those living in rural and remote areas, including First Nations Communities. Such mHealth apps offer an innovative, sustainable, and long-term strategy that will work with, and for, discrete populations to help address health care delivery barriers relating to geographic distance and the coordination of care delivery from multiple health care providers.

The results of this systematic review need to be considered in the context of a number of limitations. Although systematic searching included a number of First Nations databases and resources, it is important to note that First Nations research may be published in alternative forms to academic publications and will not have been identified in this search. The differences in study designs and interventions as well as the small number of included studies precluded a meta-analysis. Furthermore, the small number of participants in some of the included studies (4/6, 67%) as well as the homogeneous nature of participants (majority community-dwelling patients categorized as low risk from metropolitan or regional areas) reduce the ability of the findings of this review to be generalized to all people with diabetes and support the need for further research in other subpopulations with diabetes, including patients with DFD complications such as foot ulceration and amputation, as well as First Nations Peoples from Australia. Future research that aims to develop mHealth apps should consider using the assessment framework for mHealth apps [59] and learn from the findings from previous studies that have investigated barriers and facilitators to the use of digital health technologies, including mHealth apps, in other conditions in First Nations populations in geographically diverse locations [25,47,60]. In addition, an evaluation of internet access as a potential barrier to end-user engagement needs to be undertaken. In the absence of internet access, efforts must be made to provide content that is accessible offline and functionality that enables communication without the internet, including via SMS text messages. This may be the case in remote First Nations Communities in Australia where, despite the mobile phone use rate being as high as 43% [61], limitations with internet connectivity remain [47]. Although there were 3 (50%) randomized controlled trials included in the 6 studies in this systematic review, there were some concerns with the quality of their methodology, including a lack of blinding of participants to the intervention, but this is likely a limitation of the nature of this type of intervention. The knowledge and self-care management outcomes reported in these randomized controlled trials are therefore potentially biased and should be interpreted in the context of this limitation. The self-report nature of the knowledge and self-care management practice outcomes across all studies may be subject to recall bias. However, the impact of this may be limited owing to the integration of features such as on-demand education resources in all studies [26-28,42-44] and the ability to input data instantly within the mHealth apps in 2 (33%) [28,44] of the 6 included studies.

This systematic review did not identify any studies that evaluated mHealth apps for the assessment and management of DFD in First Nations Peoples in Australia. The included studies (n=6) conducted in the general population in adults with relatively low risk with diabetes support the use of mHealth apps to improve diabetes-related knowledge and self-care management. No data evaluating DFD outcomes were retrieved. Future research needs to focus on mHealth apps for populations for whom there is inadequate or ineffective service delivery and aim to evaluate the direct effects of mHealth apps on DFD outcomes. Future research needs to be for First Nations Peoples, without cross-cultural generalization, and for those living in geographically remote areas.