A comprehensive search was conducted for relevant studies in the following electronic databases: Cochrane Library; CINAHL; EMBASE and MEDLINE. These electronic databases were selected because they are continuously updated with new publications. Databases were searched between 28th July 2022 and 8th September 2022. The search strategy included terms relating to the following concepts: cancer; mAPP; and LMICs. Medical subject headings, keywords and free text terms were combined using “AND” or “OR” Boolean operators. Hand search through Google and tracing of references of all articles included were searched for additional studies. The search strategy was first developed for MEDLINE (Ovid) and modified in other databases when needed (see Supplementary Material S1).

The inclusion criteria limited admission to those studies that: reported primary research findings about mAPP interventions to control cancer; conducted in LMICs; and were published in the English language, with no date limit. Editorials, commentaries, non-original articles, studies without full texts, studies not focusing on cancer control and care, studies conducted in HICs and articles published in languages other than English were excluded. The World Bank Group categorises LMICs into low-income countries (those with a gross national income (GNI) per capita of $1,085 or less), low-middle income countries (those with a GNI per capita between $1,086 and $4,255), and upper-middle-income countries (those with a GNI per capita between $4,256 and $13,205). A total of 136 countries are combined to represent LMICs, with 28 low-middle countries, 54 low-middle income countries and 54 upper-middle-income countries [30].

Following the search of the electronic databases, all citations of the identified records were collated and uploaded into the EndNote Version 20 reference manager for removal of duplicated files and storage. The titles and abstracts of the articles were screened by two independent review authors (JA and DK-M) for relevance. The two reviewers then reconciled the outcome of the screening. Full-text of potential articles assessed as relevant on the abstract review were retrieved and screened by the same independent reviewer authors against the inclusion criteria. Full-text articles that did not meet the inclusion criteria were excluded and reasons for exclusion were justified. Any disagreements that occurred between the two review authors were resolved through mutual discussion and where no consensus was reached, a third reviewer (AD) was involved. The review articles selection process is further detailed in the PRISMA flow diagram (Figure 1).

Three review authors (EO, JA, and AD) independently extracted data such as study characteristics (first authors, publication year, country, study aim and design, age group, participants, cancer type and sample size), key study findings, mAPP name, purpose of the mAPP, main features of the mAPP, privacy and security, platform and easy-to-use. According to the mAPP intervention’s purpose, studies were categorised as providing screening, prevention, early detection, diagnosis, treatment, follow-up, survivorship, supportive and palliative care.

Two review authors (DK-M and AD) and a further researcher (SO) assessed the quality of the included studies. Quantitative studies were assessed according to the appropriate Joanna Briggs Institute Critical Appraisal Checklist, such as cross-sectional studies and cohort studies [31, 32]. Qualitative studies were assessed by using the Joanna Briggs Institute Critical Appraisal Checklist for Qualitative Research [32]. Discrepancies were resolved by discussion. Adapted

Due to the range of study designs and outcomes involved, a narrative synthesis using approaches described by Popay et al. was performed without meta-analysis [33]. The Mhealth Index and Navigation Database (MIND) framework has the following criteria: Accessibility; privacy and security; clinical foundation; engagement style; and therapeutic goal [34]. Although MIND focuses on mental health mAPP interventions, its considerations and categories are transferable to health mAPP interventions more broadly [35]. The modified MIND framework was approved as the analytical framework for this review because it offers a useful tool to appraise cancer mAPP interventions, with criteria such as: privacy and security; input and outputs; evidence base and clinical foundation; interoperability and data sharing; mAPP origin and functionality; features and engagement; and ease of use [35]. Privacy and security relate safety systems available and a description of how the mAPP uses protected health information. Interoperability and data sharing relate to the mAPP ability to connect and communicate with different systems and devices in a coordinated way, without effort from the user. Evidence base and clinical foundation evaluate if the mAPP has been researched and if there is a theory driving the mAPP development. The origin and functionality relate to the mAPP availability across platforms, cost and offline functionality. Ease of use indicates how easily users can use a the mAPP [35]. The modified MIND framework criteria are generic and applicable to cancer mAPP interventions. Information from the included studies was independently coded by two reviewers (AD and EO) and mapped against the modified MIND framework. The extracted information from the studies was read and reread to identify common themes. Any discrepancies were resolved through discussion.

Out of the 20 articles included in this review published between 2014 and 2022, 13 used quantitative [37, 38, 40–50], six used mixed-methods study designs [36, 51–55] and one used qualitative [39]. Participants included cancer patients, family members, community members and healthcare professionals, with a cumulative sample size of 33,615 ranging from 13 [39] to 22,337 [50]. The mean age was 52.57 ± 9.37. The studies were conducted across 10 countries, namely: China [38, 39, 41, 45, 47]; Iran [40, 42, 51, 53]; Brazil [43, 48, 54]; Malaysia [36, 37]; Bangladesh [50]; India [52]; Indonesia [44]; Tanzania [46]; Taiwan [55]; and Madagascar [49] (see Table 1). Figure 2 shows the geographical distribution of the identified studies.

Supplementary Table S1 shows the characteristics of identified cancer control and care mAPP interventions. Majority of the studies (n = 15) did not have a theoretical framework guiding the mAPP intervention design, implementation and/or evaluation. Identified mAPP interventions focused on breast cancer (n = 6) [38, 42, 44, 50, 53, 55], colorectal cancer (n = 3) [37, 40, 51], cervical cancer (n = 2) [49, 52], skin cancer (n = 2) [46, 48], oesophageal cancer (n = 1) [47], liver cancer (n = 1) [41], prostate cancer (n = 1) [45], paediatric cancers (n = 1) [43] and general cancers (n = 1) [54]. More than half (n = 10) of the identified 18 mAPP interventions aimed at optimizing prevention, screening and early detection of cancer [36, 37, 42–46, 48–50, 52]. Almost one-third (n = 5) of the mAPP interventions focused on improving supportive and palliative care for adults with cancer [38, 40, 47, 53, 55] and two mAPP interventions were designed to help people with cancer make informed treatment decisions and maintain autonomy [41, 54]. Only one mAPP intervention was developed and implemented to facilitate follow-up and survivorship care for people with colorectal cancer in Iran [51].

Nine of the mAPP interventions were compatible with only Android [37, 40–44, 50, 53, 55] and two were compatible with both Android and iPhone operating systems (iOS) [38, 54]. Four of the mAPP interventions combined the features of both web-based and Android/iOS [46, 48–50]. However, four studies did not report whether their mAPP interventions were built on a single or cross-platform to engage users [45, 47, 51, 52].

Two breast cancer prevention and early detection mAPP interventions included information about breast cancer (e.g., risk factors, awareness of early signs and symptoms and prevention strategies), breast self-examination training videos and a monthly breast self-exam reminder system [42, 44]. One quasi-experimental study assessed the effect of an mAPP (n = 65) compared to no intervention (n = 75) on participants’ performance and health beliefs about breast self-examination. The mean differences of the scores of perceived susceptibility (1.03 ± 2.65 vs. 0.01 ± 0.42, p = 0.001), breast self-examination barriers (2.80 ± 5.32 vs. 0.04 ± 1.43, p = 0.001), self-efficacy (10.75 ± 7.63 vs. −2.75 ± 2.44, p = 0.001) and health motivation (2.77 ± 3.70 vs. −0.29 ± 0.63, p = 0.001) were significantly higher in the mAPP group compared to the control group [42].

The most common features of two skin cancer screening and early detection mAPP interventions were: camera for digital imaging of skin lesions; clinical history and characterization of skin lesions; and web portal for a dermatologist to provide diagnostic opinion [46, 48]. One feasibility study reported that 81% of 77 lesions from 69 participants captured by an mAPP were considered malignant and referred for biopsy and histologic diagnosis [46]. The histopathological findings showed that 85% of the 55 lesions biopsied were confirmed as skin malignancies [46].

Key features of mAPP interventions to support people with cancer navigate treatment decision-making included: epidemiological information; laboratory examination information; treatment-related information; and patient-clinician interactive platform [41, 54]. Findings from a quasi-experimental study from China revealed that participants in an mAPP intervention group named “Shared Decision-Making Assistant” had significantly higher decision preparation score (80.73 ± 8.16 vs. 63.84 ± 7.38; p = 0.001), decision self-efficacy score (87.75 ± 6.87 vs. 76.89 ± 13.46; p = 0.001), decision satisfaction score (25.68 ± 2.10 vs. 23.12 ± 3.91; p = 0.001) and knowledge of primary liver cancer treatment score (14.52 ± 1.91 vs. 12.72 ± 2.13; p = 0.001) than those of the control group [41]. The study also reported significant lower decision conflict scores for participants in the intervention group compared to those in the control group (16.89 ± 8.80 vs. control group: 26.75 ± 9.79; p < 0.05) after 3 months [41].

Supportive and palliative care mAPP interventions had features such as: psychological support [40, 47, 55]; nutrition/diet guidelines [40, 47, 55]; physical activity [40, 47]; patient educational content [38, 40]; online community for people with cancer and their families for social support [38, 47]; self-care platform for monitoring physical symptoms (e.g., pain) and sharing experiences [38, 40, 55]; and medication chart to enable safe and accountable medication management [40]. It was identified that psychological techniques including mental support, music therapy, meditation, good sleep and relaxation exercises help to ease anxiety, distress or sadness the person with cancer may feel [40, 47, 55].

Easy-to-use mAPP was the most commonly identified usability dimension, which was assessed in eight studies [36, 40, 42–44, 51–53]. The review identified that highly educated participants perceived mAPP interventions as easy and effortless to use. For example, a study among Brazilian nurses to evaluate the usability of an mAPP for early detection of paediatric cancer reported:

Studies reported benefits that older adults with cancer derived from mAPP interventions. These benefits included: medication reminders [40]; high-quality remote consultation [51]; and timely provision of psychosocial support [40, 55]. However, the review recognized that older adults with cancer face challenges in learning and utilizing mAPP interventions. Family members needed to support and guide the elderly cancer survivors to learn and use mAPP interventions. For example, an Iranian participant narrated:

Other identified dimensions of cancer mAPP usability included: layout and interface quality [36, 40, 43, 53]; learnability [36, 40, 53]; system quality [36, 43, 53]; simplicity and understandable [36, 43, 53]; user-friendly [36, 53]; tailored communication [36, 53]; coordination [36, 53]; technical support [36, 53]; terminology [40]; and task execution [43].

All the identified mAPP interventions collected and tracked users’ data to help improve early detection, enhance cancer treatment and optimize palliative care. More than half (n = 10) of the included studies stored the data on users’ devices [37, 38, 40–44, 50, 53–55]. Four studies stored them on both the server and users’ devices [38, 50, 53, 54] and one study stored all the data on the server [49]. The review identified that storing users’ information behind password-protected encrypted channels was imperative to ensure privacy and security. For example, one study that used mAPP intervention to monitor women participating in cervical cancer screening campaign in Madagascar reported:

Sixteen validated assessment tools were identified in the included studies. One study used the Questionnaire for User Interaction Satisfaction version 5.5 to evaluate the usability of a smartphone-based application for the self-management of people with colorectal cancer [40]. The System Usability Scale was employed to evaluate the usability of an mAPP developed for early detection of paediatric cancer in Brazil [43], the quality of an educational mAPP aimed at improving the resilience of people with breast cancer in Iran [53] and an mAPP focused on community education on colorectal cancer [36]. Tools that were adopted to evaluate the feasibility, safety and efficacy of mAPP intervention to improve the quality of life of people with oesophageal cancer included: European Organization for Research and Treatment of Cancer-Quality of life Question-Core; European Organization for Research and Treatment of Cancer-Quality of life Question-Oesophageal Cancer Module; Chinese versions of Patient Health Questionnaire-9; General Anxiety Disorder-7; and Perceived Stress Scale-10 [47]. The study that developed the mAPP intervention to assist the decision-making needs of people with primary liver cancer used six tools: Decisional Conflict Scale; Preparation Decision Making Scale; O’Connor’s 11-item Decision Self-Efficacy Scale; Satisfaction with Decision Scale; Breast Cancer Knowledge Scale; and Decision Regret Scale [41]. The Champion’s Health Belief Model Scale was used by one study to determine the effect of an mAPP intervention on women’s performance and health beliefs regarding breast self-examination [42].

Six of the quantitative studies adopted instruments with acceptable validity and reliability to measure outcomes [37, 40–43, 47]. All the studies described the settings and eligible participants. Two of the included studies did not report whether or not they secured ethics approval [51, 54]. Five of the included studies explicitly indicated that extensive literature review was conducted to iteratively develop the cancer mAPP interventions [39–41, 51, 55]. The included qualitative studies did not have a statement locating the researchers culturally or theoretically. The outputs of the integration of qualitative and quantitative components of three of the mixed method studies were not adequately interpreted [51, 53, 54] (see Supplementary Material S2).

To the best of our knowledge, this is the first systematic review of the available literature on mAPP interventions implemented to help control and manage cancer in LMICs. The review was conducted using rigorous systematic literature search methods. The limitations are that the review did not search app stores, so useful cancer control and care mAPP interventions not published in peer-review journals may have been overlooked. At the same time, however, other reviews have found that mAPP interventions that have not been published are unlikely to have been extensively user-tested and often have limited usability [69]. Perhaps most importantly, while several of the mAPP interventions we identified had satisfactory usability, our review could not identify robust evidence for the effectiveness of mAPP interventions in achieving positive cancer control and care outcomes in LMICs. Poor access to mobile technology, low digital skills and low health and written literacy levels were missing from the MIND framework. Finally, the available literature does not provide detailed information on data management of cancer control and care mAPPs such as the costs and number of downloads, making it difficult to ascertain how widely used the interventions are.

Cancer mAPP interventions are gradually gaining attention in LMICs because they provide unique resources to empower and strengthen the participation of people with an interest in cancer or to involve those with cancer in their own care. To enhance cancer control, a focus on prevention and early detection is important; however, mAPP interventions related to cancer treatment, follow-up and survivorship are also needed to enable more cost-effective cancer care. It is recommended that future cancer control and care mAPP development, implementation and evaluation should be informed by the existing knowledge available and stakeholder engagement to improve acceptability. This research team will proceed with a study to explore application functionalities, features and potential barriers for the development of an mAPP suited to enhancing the supportive care of patients receiving radiotherapy.