Hypertension Diagnosis and Management in Africa Using Mobile Phones: A Scoping Review

Target 3.4 of the third Sustainable Development Goal (SDG) of the United Nations (UN) General Assembly proposes to reduce premature mortality from non-communicable diseases (NCDs) by one-third. Epidemiological data presented by the World Health Organization (WHO) in 2016 show that out of a total of 57 million deaths worldwide, approximately 41 million deaths occurred due to NCDs, with 78% of such deaths occurring in low-and-middle-income countries (LMICs). The majority of investigations on NCDs agree that the leading risk factor for mortality worldwide is hypertension. Over 75% of the world's mobile phone subscriptions reside in LMICs, hence making the mobile phone particularly relevant to mHealth deployment in Africa. This study is aimed at determining the scope of the literature available on hypertension diagnosis and management in Africa, with particular emphasis on determining the feasibility, acceptability and effectiveness of interventions based on the use of mobile phones. The bulk of the evidence considered overwhelmingly shows that SMS technology is yet the most used medium for executing interventions in Africa. Consequently, the need to define novel and superior ways of providing effective and low-cost monitoring, diagnosis, and management of hypertension-related NCDs delivered through artificial intelligence and machine learning techniques is clear.

source of concern within the past decade.A non-communicable disease is usually a chronic (long-lasting) disease which cannot be transmitted directly from one individual to another.Statistics presented by the World Health Organization (WHO) [1] show that in 2016, out of a total of 57 million deaths worldwide, approximately 41 million (71%) deaths occurred due to NCDs.Four major NCDs-specifically, cardiovascular diseases (CVDs), chronic respiratory diseases, cancer, and diabetescause the death of about 15 million men and women each year with age distribution between 30 and 70 years [1]- [3].

A. Significance to Low-and Middle-Income Countries
Even though many low-and middle-income countries (LMICs) do not have systems in place for collecting information on causes of death, estimates from incomplete data gathered by the WHO show that 78% of global NCD deaths occurred in LMICs in 2016 [2]- [4].According to a World Bank classification conducted using the Atlas method [5], low-to upper-middle-income economies are defined as those with a gross national income (GNI) per capita ranging from $1,045 or less to just about $12,695 in a fiscal year [6].Unfortunately, most countries in Africa fall within the low-and lower-middleincome groups and are faced with limited human, financial, and infrastructural resources.This situation provides a breeding ground for an impending NCD epidemic [7]- [10] that is mostly being driven by health service problems, population growth, acute poverty, rapid urbanization, as well as a proliferation of globalization strategies in marketing leading to consumption of products harmful to health [11]- [13].Consequently, novel ways of providing effective and low-cost diagnosis and management of NCDs that can deliver superior results must be engaged.

B. NCD Risk Factors
NCD risk factors can be broadly categorized into modifiable behavioural and metabolic risk factors.Modifiable behaviours include, among other factors, physical inactivity and unhealthy diet, tobacco use, and the harmful use of alcohol [14], [15]- [17], while four key metabolic changes that can increase the risk of NCDs include raised blood pressure, obesity, hyperglycemia, and hyperlipidemia.According to [18], the leading risk factor worldwide leading to death is elevated blood pressure, which, if left uncontrolled, eventually leads to hypertension and finally, death through a number of complications.This finding is also confirmed by various studies done on major NCDs [19]- [21].Outcomes of a randomized controlled trial (RCT) carried out  [29] on stroke patients in both Ghana and Nigeria [22] show that the risk of hypertension for stroke occurrence is the highest in the world.Results obtained from these studies reasonably lead us to conclude that controlling the incidence of hypertension in Africa will significantly reduce the burden of disease.Hypertension is progressively assuming epidemic proportions in Africa [23], [24].The most alarming aspect is that it now features in younger populations in many African countries [25], and many cases go undetected because of a poor health culture, thus making it difficult to initiate control strategies on time.
Historically, control and management of hypertension rely on antihypertensive drug use [26] and hospital visits, but more recently, hybrid therapy regimens are emerging, which include early detection and prevention strategies, lifestyle changes [27], and self-monitoring.This changing outlook in therapeutic approaches and public perceptions presents a wide vista of opportunities for Africa to achieve hypertension control.It is imperative to combat challenges such as poor attitudes towards basic personal health maintenance which are usually fueled by unfavorable cultural, religious and superstitious beliefs [28], low levels of awareness [27], inadequate medical facilities, poorly established health systems, poverty, and shortages of suitably qualified medical personnel [29] as shown in Table I.

C. ICT and Mhealth Strategies
An emerging trend in the management of NCDs is the use of information and communications technology (ICT), where management systems are deployed using mobile technologies and tools.These systems are also referred to as mobile health (mHealth) systems, and they present a promising area for providing reliable, efficient and cost-effective access to health services [7], [30].One of the more popular devices employed for mHealth strategies is the mobile phone.Smartphones represent a subset of regular mobile phones because they possess more features and computing power.Their penetration within the past decade has significantly grown worldwide, although at different speeds and assimilation levels in advanced and emerging economies [31].In this review, we shall be using both terms (i.e., smartphones and mobile phones) interchangeably to provide a broader context where relevant since both technologies can be used to make calls and send text messages.Documented evidence as far back as 2012 shows that over 75% of the world's mobile phone subscriptions reside in LMICs [32].Current statistics obtained from the Nigerian Communications Commission (NCC) as of January 2019 show that the number of active lines for Global System for Mobile communications (GSM) services totals over 173 million, thus making Nigeria the African country with the highest smartphone penetration [33].Similarly, about 83% of the total population in Kenya is in possession of a mobile phone [7].The mobile phone is particularly relevant to mHealth deployment in Africa [34], [35] because of its prevalent use.Some of the characteristics that make its use so desirable include portability, ease of use, relatively low cost, personalized effects, and constant connectivity.Although several works have proposed some form of mHealth deployment, often, the peculiar dynamics of low resource settings (LRS) are never fully considered [36], hence the need to design, develop, and adapt smartphone interfaces and applications for expanded, frugal [37]- [40], and resilient functions in hypertension detection, diagnosis, and management.This study therefore attempts to show the practicability (or lack thereof) of hypertension diagnosis and management using mHealth technologies in the African region.

II. METHODS
Since this review is aimed at determining the scope of the bulk of literature available on hypertension diagnosis and management in Africa, with particular emphasis on interventions based on the use of the mobile/smart phone, key factors relating to hypertension diagnosis and management were highlighted with the intention of identifying and analyzing knowledge gaps in the body of research.For this review, Arksey's and O'Malley's method [41] for scoping reviews was employed, with reference to some other published works [42].This method enabled easy determination of the inclusion criteria, identification of pertinent studies, study selection, and data charting.

A. Data Sources and Search Strategy
PubMed database served as the main information source, and it was queried for the following search terms: "(hypertension) AND (Africa) AND ((phone OR phones) OR (mobile phones OR smartphones) OR (smartphone OR mobile phone))".The search turned up only publications written in English and was filtered to reflect articles published within the past decade.Thirty publications were retrieved using this process.

B. Study Selection
Records retrieved from the procedure described above were scanned by their titles for a preliminary assessment of relevance.Thereafter, a detailed perusal of the abstracts of the remaining articles was carried out to identify articles related to the topic of interest.Studies were included in the selection if they: a) were carried out in African countries, or in a pool of LMICs that included at least one African country; b) focused on hypertension or elevated blood pressure, or comorbidities that included hypertension; and c) the technology used for such intervention was mobile phone based.In the first instance, precedence was given to review articles since they typically gather and summarize information from a large pool of articles related to the area of interest.Secondarily, RCTs as well as other germane studies were also included.After excluding select studies based on the inclusion criteria, 18 articles remained which formed the core of the scoping review.Furthermore, additional articles were retrieved from both the reference lists of the final selection for the scoping review and the discarded articles.

C. Data Extraction, Categorization, and Charting
A data extraction process was initiated based on examples obtained in [42]- [45].Applicable data extracted were categorized and charted from the selected studies.The data extraction chart is presented in Table III in the Appendix.This table presents an overview of the characteristics of selected studies relating to paper cited, author, publication year, risk factor or disease, study population, duration of study, sample size, technology used, study/intervention, methodology, and finally, outcomes/results obtained for each article.Papers were subsequently chronicled by classification of intervention/study using a descriptive method.

III. RESULTS
Overall, a total of 30 articles were retrieved from the search done on PubMed.Out of these, four were considered to be irrelevant to the topic of interest after reviewing only the titles.Additionally, eight more articles were classified as being ineligible for inclusion after reviewing the abstracts.Even though the interventions were directly or indirectly related to hypertension, either they were carried out in other LMICs outside of the African region, or the study deployment was not achieved by using any of a number of features offered by a mobile phone.Fig. 1 gives a graphical illustration of the entire procedure.

A. Study Characteristics and Distribution
A search period of ten years (spanning 2009 to 2019) was specified for records retrieval.The rationale behind this was to recover as much relevant information as possible while suppressing obsolete material.The earliest records available were dated   2014 [43], [46], [47], while the latest were published in [48]- [51], showing that adaptations of mHealth strategies for addressing NCD-related risk factors/conditions is a relatively new concept in the African region.The final selection of articles included four reviews, three study protocols for RCTs, one documentation of development and evaluation of a RCT, six studies of varying depth and breadth, and four RCTs.Out of the four reviews, three were systematic reviews.All systematic reviews featured articles from a mix of LMICs, and each had only one documented intervention coming from Africa.In one of the systematic reviews, the documented intervention from Africa [52] had already been identified in the retrieval.All reviews reported hypertension as being a major risk factor for NCDs.One of the systematic reviews combined hyperlipidemia as a risk factor, while others discussed NCDs such as diabetes and atherosclerotic cardiovascular disease.Of the three study protocols documented in the retrieval, two were targeted at West African countries [47], [53].The third was directed in particular at South Africa [46], and the documentation of the development and evaluation of the RCT in [46] was presented in [51].
The largest segment of publications included studies covering different aspects of mHealth deployment and management.Two studies focused on deaf people in South Africa: the former dealt with improving their knowledge on hypertension [52], while the latter [54] evaluated the cost of health technologies deployed for hypertension in deaf patients.Other documented studies include a 10-year CVD risk identification tool deployed amongst Kenyans [50], an exploratory investigation of challenges surrounding hypertension management amongst stroke survivors [55], and the investigation of the level of awareness and management of elevated blood pressure amongst human immunodeficiency virus (HIV)-infected adults in Zambia [56].The last study focused on the assessment of issues surrounding the use of private sector drug retail outlets for screening hypertension in Tanzania [57].
The RCTs featured in this review were implemented in South Africa and Ghana, respectively.Two separate interventions were implemented in South Africa, and both focused on supporting and improving treatment adherence for lowering blood pressure in adults [58], [59].The intervention delivered in Ghana was implemented in multiple stages [48], [49], and it also had a similar objective of lowering blood pressure, specifically among stroke survivors, to reduce the risk of recurrent vascular events.Even though a study protocol for a RCT to be administered in Nigeria [47] was presented, no documentation concerning its actual implementation was uncovered in the search.Overall, the most represented country in terms of documentation in the African region was South Africa, followed by Ghana.

B. mHealth Technology Adaptations
The frequency chart depicted in Table II gives us an overall picture about the feasibility of employing ICT and the technological features of the mobile phone as diagnostic and management tools for detecting risk factors associated with NCDs.Studies carried out in [48], [53], and [60] are centered on the same intervention: Phone-based Intervention under Nurse Guidance after Stroke (PINGS).PINGS used interactive short message service (SMS) in combination with a Bluetooth blood pressure measuring device (UA-767Plus BT) to monitor and control blood pressure levels.Similarly, studies in [46], [51], [58], and [59] also reported different aspects of another intervention, SMS Text-message Adherence suppoRt (StAR), which combined interactive text messages with an open-source webbased electronic medical record system to test for the efficacy of the system in improving blood pressure control and treatment adherence.
Out of all the interventions documented in Table II (n = 10), a total of eight adapted the use of text messages (SMS) using mobile phones at the client end; two employed phone surveys/interviews (in the "Mobile/Smartphone/Tablet without App" column), while only two used a mobile app in synergy with a medical device.Only five studies reported using internet-based technologies and tools.All three systematic reviews included in the categorization chart reported a bevy of interventions on the use of mHealth technologies to address health-related issues [43]- [45].However, only one intervention in each case was reported for Africa, and all studies used SMS to implement the intervention.Findings in [43] showed that most of the earlier interventions reported (i.e., before 2014) employed email and online websites, while more recent studies made use of smartphone technologies, applications and biosensors.

IV. DISCUSSION
Large-scale mobile health technology adaptation for mHealth deployment in Africa is increasingly becoming a viable and notable option to combat the rising menace of NCDs.The intent is to deploy these technologies maximally in the long run to solve challenges arising from sub-optimal medical facilities, lack of specialized medical knowledge, shortage of medical practitioners, economic constraints (e.g., out-of-pocket expenses), poor supply chain, location issues, and behavioural concerns, both on the side of users/patients and health providers/managers.This scoping review highlights and maps the body of work on the diagnosis and management of hypertension in Africa using mobile phones or smartphone devices.
In all, eleven studies carried out exclusively on the African continent were revealed which employed mHealth tools and technologies with varying degree of evidence.Studies done in [48], [53], and [60] in Table II represent different submissions on the same study (PINGS), while [46], [51], [58], and [59] focused on different stages of another study (StAR).Relative to the above, the three systematic reviews [43]- [45], also featured in the chart (Fig. 2), provide a broader coverage of interventions done across LMICs, albeit with different emphases.It is worthy of note, though, that only one intervention from Africa was reflected in each of the studies.Altogether, four major themes were extracted from the body of work consulted: namely, hypertension management and control (n = 7), cost evaluation of mHealth technologies (n = 1), screening for risk factor detection (n = 2), and health promotion using mobile phone technologies (n = 3).
Interestingly enough, only one intervention [49] reported use of some form of mobile application on a smartphone, which was combined with a Bluetooth-enabled blood pressure measuring device, to monitor blood pressure measurements and medication.Outcomes reported include a very positive attitude shown by a majority of participants to the mHealth technologies they had been exposed to.Many expressed their satisfaction with the intervention, while some even regretted not having been exposed to such earlier.Equally, a number of participants expressed the Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.
desire to continue with the new knowledge and attitudes they had acquired through the interventions [48], [49].
The bulk of the evidence considered overwhelmingly shows that SMS technology is yet the most used medium for executing interventions in Africa.However, it would seem that text messages appear more suitable for programs dealing with medication adherence, health education, and interventions to promote physical activity and healthy diets than early detection, diagnosis and prognosis of disease.Despite the encouraging results recorded in the studies presented above which used text messages for deploying mHealth interventions, the general applicability of such interventions to every sector of the African populace is doubtful because of reported low levels of literacy that still exists in many parts of Africa.Given this limitation, it seems imperative that novel ways of providing effective and low-cost monitoring, diagnosis, and management of NCDs which can yet deliver superior results are to be engaged, either in combination with the use of text messaging, or with an entirely new focus.The use of Apps employing rich pictorials and graphical illustrations as well as frugal design principles may well compensate for the low levels of literacy and lack of specialized knowledge.Of particular note is the recommendation by the WHO and the UN of the use of infographics and iconography [61], [62] in conveying health messages to the public.These graphic visual representations of information, data, or knowledge can be color characterized to convey information quickly and clearly.Furthermore, the feasibility of horizontal models engaging quantitative approaches based on machine learning methods and decision support algorithms was underreported.Local manufacturing of phone accessories using 3D printing techniques was also little represented in the gathered evidence.In terms of detection and diagnosis of NCDs, integration of biosensors, either available on the mobile phone or attached as an external sensor, could potentially provide optimized capabilities for early detection and prevention, while exploiting communication capabilities (ICT) on the smartphone may improve linkage to and quality of care [7] with health care providers.
The negligible amount of evidence on interventions using mobile phone technologies beyond SMS in Africa to tackle the NCD burden suggests that further research is needed.Locally developed AI, ML, and 3D printing technologies can be successfully implemented through higher quality evidence-based research in the African health sector.This submission is supported by the fact that all eleven interventions included in this review report the feasibility and potential of carrying out mHealth interventions in Africa, particularly as regards NCDs and their risk factors, of which hypertension is the leading risk factor.Levels of evidence could vary from experimental RCTs and mixed methods research to systematic literature review of qualitative studies/quasi-experimental studies/RCTs (with or without meta-analysis), and also field studies.For example, Stokes et al. [63] report the design of an interpretable machine learning model from data obtained from a LMIC, for distinguishing pneumonia patients from bronchitis patients based on symptoms and signs alone.Pending further validation, this system could be suitable for incorporation into a diagnostic tool such as an application for use on a mobile phone.Similarly, Ellis et al. [64] presented the redesign of the inlet filter of an oxygen concentrator.The redesign is based on a reverse engineering approach, which makes use of activated charcoal coupled with 3D-printing methods, to accommodate environmental conditions that present in many African countries.In another work, Piaggio et al. [65] describe the design, prototyping and validation of a mobile app for testing the photopupillary reflex using an Android smartphone, specifically for low-resource settings.This is proposed for use in the screening of eye pathologies and brain trauma without requiring expensive settings or complex procedures.These are recent exemplars of evidence-based AI research focused on LRS, which can easily be adapted to the African context.
A worrying issue endemic to most African countries is the lack of infrastructural and organizational [66] healthcare capacity.The systematic reviews earlier mentioned highlighted some core concerns impacting negatively on the successful integration of mHealth technologies into the healthcare sector on the African continent.Though there were no lengthy discussions on these crucial aspects, common regional challenges identified include low systems and workforce capacity for managing data and digital technology and inadequate financing to support digital health; especially when considering levels of AI expertise, available infrastructure, existing AI financing arrangements, and logisticsrelated follow-up care.In [67] for example, Luxon argues that every aspect of healthcare infrastructure and organization should promote improved standards of care and wellbeing for patients and healthcare personnel.The need to include local populations in AI-based research that allows adaptions to local cultures, community needs, and languages cannot be overemphasized [68], [69].Likewise, applying a systems-based approach to implementing national mHealth strategies [70], institution of strong regulatory and policy frameworks, and appropriate sourcing and monitoring of financing structures can ultimately ensure sustainability and reliability.

V. CONCLUSION
By carefully examining the scope of work presented in the results, it becomes evident that Africa is still at an infant stage in developing mHealth solutions and adaptations that can fast-track the process of providing reliable and effective solutions, applicable on a large scale, to counter and reduce the NCD threat and its resulting effects on the African populace.Culturally relevant solutions resilient to extant conditions in the African region are needed to provide lasting results.These solutions should mitigate the effects of poor responsiveness and availability of public and private sector health services and infrastructure, counter discordant beliefs about maintaining personal fitness [71], and reduce costs and challenges associated with accessing health facilities situated at great distances from patient locations.The outcomes from this scoping review suggest that implementing mHealth solutions to combat NCDs and related diseases exhibiting hypertension as their major risk factor would find large-scale acceptance among prospective users.This scoping review has served to highlight research gaps with respect to diagnosis and management of hypertension in Africa, and our findings show that with a greater adoption of mHealth technologies in Africa, there can be potential improvement of the NCD burden.

Data Access Statement
This review is a re-analysis of data extracted from studies publicly available on PubMed.Data derived through the reanalysis undertaken in this study as well as all supporting data are available in the appendix of this paper and in the supplementary information accompanying the paper.Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.

Fig. 1 .
Fig. 1.Flow chart for search strategy and study selection.

Fig. 3 .
Fig. 3. Article distribution in relation to country location.
Fig. shows the percentage distribution of articles.

TABLE II FREQUENCY
DISTRIBUTION OF TECHNOLOGY USED

TABLE III DATA CATEGORIZATION CHART
Authorized licensed use limited to the terms of the applicable license agreement with IEEE.Restrictions apply.*-Same study identified in two different papers(9 * , 17 * )