Globally, Indigenous populations share experiences of devastating disruptions to their development due to violent settler colonial practices that persisted well into the twentieth century and continue to impact them today [1]. This historical suffering and the consequent vestiges are inextricably linked to the current overrepresentation of Indigenous people in population statistics relating to avoidable mortality, disease burden, and social and economic disadvantage [2]. To note, in this review, the term Indigenous is used for Indigenous populations across the globe.

Adverse health outcomes including low birth weight, preterm birth, stillbirths, and perinatal mortality rates disproportionately affect Indigenous families [3,4]. Indigenous women are more likely to experience domestic violence, have high levels of psychological distress, and exhibit a higher prevalence of behavioral risk factors including tobacco, alcohol, and other drug use in comparison with non-Indigenous women [5-7]. Disparities between Indigenous and non-Indigenous maternal and child health outcomes highlight the need for urgent and specialized support [3]. The perinatal period offers a unique opportunity for women to adopt healthy behaviors for the benefit of both themselves and their babies [8]. In addition, experiences during the first 2000 days of a child’s life have a significant, lasting impact on cognitive, physical, social, and emotional health [9,10]. Therefore, interventions applied during the perinatal period and early childhood are more successful than those implemented at a later stage of life [9].

Of relevance is the geographic impact on health inequities experienced by Indigenous populations. Compared with urban or regional populations, communities in rural and remote regions experience higher mortality [11]. The disparity is attributed to barriers associated with high cost, poor access, and culturally inappropriate service provision, as well as employment, education, and income disadvantages [12].

Mobile health (mHealth) interventions offer an alternative option for the provision and dissemination of health information, particularly for communities outside major cities where health services reach is reduced. Health service providers are increasingly using digital technology for health promotion due to its practicality, reach, and accessibility [13]. mHealth is being used for education, behavior modification, data collection and tracking, point-of-care diagnostics, health decision support, and record keeping [14,15]. This coincides with the increased use of mobile technology and specifically, its use for health and social and emotional well-being purposes by Indigenous women [16,17].

It has been established that Indigenous women are likely to successfully engage in services during the antenatal period when they feel empowered through education, family and community support, and cultural connectedness [4,6,18,19]. Effective health promotion is accessible, empowering, respects the values of Indigenous family structures and kinship systems, and most critically, is Indigenous-led [20]. Like any other intervention, mHealth when used for Indigenous populations requires the intervention to be culturally appropriate and acceptable [21].

There are very few studies that have reported the benefits of culturally responsive mHealth interventions specifically developed to support healthy lifestyle choices for Indigenous women [17,22,23]. Research that fails to consult with Indigenous communities has been shown to provide little benefit [10]. Applying homogenized, mainstream health interventions risks perpetuating inequities, hindering engagement, and discounting the legacy of historical atrocities [24]. It is critical that mHealth is culturally safe and acceptable and that evaluations consider differences in participant characteristics, features of the intervention, and the context of delivery [25].

The primary objective of this systematic review was to assess the effectiveness of mHealth interventions designed for use by Indigenous mothers (from conception to 5 years post partum) who seek to promote healthy behaviors. The secondary aims of this review included examining the observed differences in the effectiveness of the interventions based on participant engagement, the intervention design, and the programmatic and community context in which the intervention was implemented. This review also explored whether the mHealth intervention was designed or adapted specifically for Indigenous populations.

A systematic review protocol outlining the search strategy, methods for title and abstract and full-text screening, data extraction template, quality assessment methods, data analysis methods, synthesis, statistical issues, publication bias, and any conflicts of interest was developed and registered with PROSPERO (CRD42023395710) [26]. The review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for the conduct and reporting [27] (Multimedia Appendix 1). The protocol was followed fully except that the third aim, to explore if the identified interventions were co-designed, was included after the protocol was submitted for registration to PROSPERO based on experts’ advice.

This systematic review demonstrates that there is no evidence of the effectiveness of any currently available mHealth intervention designed for Indigenous women with young children (aged 0-5 years) to assist with behavior change for making healthy choices for themselves and their children. Only 3 articles found were eligible to be included in this systematic review. These studies aimed to change maternal and child behaviors in the prenatal period and preschool age group by provision of an mHealth intervention. The 3 mHealth interventions that were identified included a hybrid face-to-face and online prenatal education program; an automated, individualized, 3-times-per-week text message program giving healthy lifestyle advice for targeting preschool children; and a web-based self-assessment followed by tailored advice related to maternal alcohol drinking. All interventions were applied in a real-life community setting. A statistically significant change was reported in some behavioral outcomes. However, in the absence of any reported sample size or power calculation; it was not possible to ascertain if significant positive results were due to an inflated type I error, showing positive results in favor of intervention when in fact the interventions only worked due to very large sample sizes.

A recent review on barriers and facilitators of engagement of Indigenous peoples with mHealth interventions revealed themes that echo the need for effective and timely co-design practices in this field [39]. Researchers found that barriers to engagement included poor digital design, unreliable technology and internet, repetitive content with limited user input, language barriers, poor cultural representation, nonspecific health information, and privacy concerns [39]. Co-design has been identified as one of the most important factors determining the uptake of web-based therapeutic interventions for Indigenous populations [40]. In our review, it was deemed important to investigate if the digital interventions used were either co-designed in partnership with Indigenous communities or were adapted in consultation with the target population to ensure cultural appropriateness, or if they were selected from a pool of previous literature on culturally appropriate interventions for the target population. That information as reported in the included studies is presented in Table 3.

Of the 3 identified mHealth interventions, 2 were co-designed with an Indigenous population and showed some positive outcomes in favor of the interventions. Citations of their authors’ previous work on co-designing the interventions with Indigenous populations were reported for these 2 interventions in these studies [34,35]. Although the current evidence on their effectiveness is limited, the authors of these studies are following the iterative process of evidence gathering as suggested in the guideline on the evaluation of complex interventions [41-43]. Therefore, it will be important to see future results on the use of these interventions. Also, due to the diligent process of co-design and the assurance of cultural safety, these interventions may provide guidance on the planning and development of work with Indigenous populations on the provision of maternal behavior support. The consultation and partnership with communities reported by the 2 interventions are in line with the international guidelines on research with Indigenous peoples [44].

During the study selection process, 8 articles on the development of mHealth interventions for Indigenous women were identified that could not be included in this review [17,23,45-50]. Upon a further search with the names of the digital interventions reported in the studies, no follow-up studies on the evaluation of their effectiveness could be retrieved. Further to this, the corresponding authors of these articles (n=8) were contacted to inquire about any ongoing studies to evaluate their effectiveness. Except for 1, all authors replied to the initial email (7/8). Four authors (4/7) mentioned further evaluation work on the identified digital interventions, of which only 1 has already been published [51]. However, the focus of the publication was not an Indigenous population, therefore the study was not included in this review [51]. The other authors (n=3), with ongoing trials, were not able to share the unpublished data. Evidently, nearly 40% (3/8) of these mHealth interventions did not evaluate their effectiveness and are no longer active [17,23,48]. Although, in response to personal communication, one of the authors [23] mentioned the collection of qualitative data on the use of an app designed for Indigenous women [23], the need for an RCT on the evaluation of effectiveness was also identified. Consistent with the findings of our review, the evidence on the effectiveness of digital health interventions in other populations, such as children and adolescents [52], patients with heart failure [53], self-management of chronic obstructive pulmonary disease [54], and sexual health promotion [55] is limited due to the lack of published findings of well-designed and well-conducted effectiveness evaluation research. There are some health conditions for which the use of mHealth interventions has been well studied, such as asthma [56,57] and consumption of hazardous amounts of alcohol [58], in which the use of mHealth interventions has been shown to produce positive health outcomes in favor of the interventions [57,58]. Also, the use of mHealth to promote behavioral change during prenatal care in the general population has been reported to result in significant improvement in behavior risks, and improved healthy behaviors among pregnant women [59]. A current scoping review on the use and uptake of web-based therapeutic interventions among Indigenous populations across Australia, Canada, New Zealand, and the United States reported the use of these interventions for conditions such as cardiac care, diabetes, nutrition, mental health, asthma, neonatal care, otitis media, smoking cessation, and substance misuse [40]. For that review, a web-based therapeutic intervention was defined as a therapeutic intervention that was self guided, or a clinician-assisted program delivered via the internet to provide guidance, support, and treatment for health conditions. While the evaluation of effectiveness was beyond the scope of that review, it was concluded that these interventions have the potential to improve health, overcome treatment barriers, and reduce inequalities for Indigenous populations [40].

There are several factors that could lead to the lack of effective evaluation research and in turn scarcity of evidence on any mHealth interventions [60]. Some of these challenges pertain to interventions themselves, whereas some are related to designing a powered RCT with complex interventions, which is the common approach to effectiveness evaluation. Some of the challenges related to mHealth interventions are rapidly changing technology and the need for an intervention to evolve quickly to merely remain functional, let alone be upgraded; accessibility of the intervention across various operating systems and devices; acceptability and usability of the intervention by the target population; real-world use of any intervention or implementation; the burden on various stakeholders; adaptation of mHealth the intervention to context and population; and integration of an intervention into existing health care systems. There are also challenges specific to the design of an RCT for mHealth intervention evaluation, such as selecting a suitable context for trial so that the findings are generalizable, choice between conducting a trial with high external versus internal validity, specification, and explanation of the components of the mHealth intervention under evaluation, choice of an appropriate comparator group, and data collection methods from mHealth interventions. Given all these, it is arduous to conduct powered RCTs of mHealth interventions [41,60].

The included studies in this review were appraised for their quality with the CREATE and MMAT [36] tools. The CREATE tool was designed to ensure that the research conducted with Indigenous populations is conducted from the perspective of Indigenous peoples [30]. There are certain items that authors need to report for the assessors to mark the CREATE tool. While appraising the included studies in this review, we found that many of these items had not been reported and a further inquiry with the corresponding authors was necessary to access that information. This could have been due to the lack of any guidance on reporting of Indigenous research. Recently, a guideline was published on reporting of observational studies in Indigenous populations [61]; however, the studies included in this review were of different methodological design. Another reporting guideline, the Consolidated Criteria (CONSIDER) statement, developed with a review of available literature and a meta-synthesis, was published in 2019. The guideline provides 8 research domains and 17 criteria for the reporting of research involving Indigenous peoples. However, except for one, the included studies in the review were published before this date, so they cannot be discredited for not following the guideline.

Single data extraction was one limitation of this systematic review. Due to the paucity of available information, the data were extracted by 1 reviewer only. However, once extracted, the data were checked for completeness and errors by a second reviewer with more experience in conducting systematic reviews. This limitation is less likely to have changed the results and conclusions made, since the main concern is the lack of relevant literature and research in the field. Another limitation was the relatively small sample sizes in the included studies (n=420, n=17, and n=263), meaning the total number of women included in all studies was only 700. Furthermore, the lack of formal power calculations to calculate sample size and formally assess the statistical significance of the interventions was problematic. Also, the title and abstract screening for eligibility was performed by one reviewer. This can be perceived as a limitation as there is a risk of missing relevant studies, as they could be overlooked by a single reviewer. We, however, kept our decision-making more inclusive and the reviewer performing the title and abstract screening was instructed to include any study they were unsure about for the full-text screening stage. Another potential limitation may be the use of the CREATE tool for risk of bias assessment. The CREATE tool was developed with Indigenous and non-Indigenous experts and uses a 14-question checklist that assesses community engagement, consultation, research governance, and intellectual property to evaluate the methodology of studies conducted in the Australian Aboriginal and/or Torres Strait Islander population groups [30]. However, the tool has been widely used in research involving global Indigenous populations [62-67]. Therefore, the use of the CREATE tool provides additional information about the methodological suitability of the included studies from the Indigenous viewpoint. The addition of a third aim to the systematic review after the registration of the review protocol may be considered a potential weakness. This was, however, not included as a primary aim of the review and did not impact the conclusion made in the review. In addition, it was not possible to evaluate differences in the effectiveness of the identified interventions based on participant engagement, intervention nature, and context due to the small number of studies included. Also, each study used a different intervention and none of them reported power and sample size calculations. This may be perceived as a weakness of this systematic review. However, it is reflective of the nature of the research and literature available and demonstrates that more research is required in the field.