METHODS FOR DEVELOPING MOBILE APPS IN HEALTH : AN INTEGRATIVE REVIEW OF THE LITERATURE

Objective: to identify, in the Brazilian and international publications indexed in the databases, the main methods adopted by researchers for developing mobile apps in health. Method: integrative review of the literature, of studies published in the following databases: MEDLINE/PubMed, Scopus, Web of Science, CINAHL and SciELO, in 2012 – 2016. A total of 21 articles were selected for analysis. Results: the main methods for developing mobile apps in the area of health, described in the articles, were: systematic design of instruction, contextualized design of instruction, user-centered design and systems development life cycle. Conclusion: regardless of the method of development selected, the stages must be well-defined and structured, so that the mobile app developed may be useful to the end-user. DESCRIPTORS: Mobile applications. Medical informatics applications. Nursing informatics. Medical informatics. Information technology. MÉTODOS PARA DESENVOLVIMENTO DE APLICATIVOS MÓVEIS EM SAÚDE: REVISÃO INTEGRATIVA DA LITERATURA


INTRODUCTION
2][3][4][5] These technologies are considered a global resource, which connects various computers, creating a network of information, and which makes it possible to contribute to the development and improvement of the health professions. 15 It is also emphasized that accessing information in real time and/or remotely contributes to resolving health problems/needs in different geographical regions, promoting broad coverage in terms of specialized healthcare, undertaken in the major urban centers. 1 In this scenario, emphasis is placed on the phenomenon of the mobile technologies (tablets, smartphones etc.) -in particular, the use of worldwide mobile applications (also known as apps).Apps are conceptualized as a set of tools designed to undertake specific tasks and jobs. 6obile devices, and mobile apps in particular, aim to enable people to access information and knowledge, unrestricted by time and space.8][9] Such characteristics add strategic value to the new society in the Information Era. 10 It is currently possible to observe a proliferation of mobile technologies and apps (mHealth), which is contributing to the construction of a new mode of healthcare, in which the information referent to people's health is readily available and ubiquitous. 6][13][14][15][16] Specifically in the area of nursing, it is considered that the tools made available by ICT, associated with clinical, educational and management practice, require nurses to make efforts to achieve a definition of their role regarding ICT in nursing.There is evidence of the urgent need for these professionals to undertake reflection on this subject -and also to include themselves in the technological environment of the mobile apps, which are so strongly present in the Brazilian cultural, social and economic contexts. 17n the light of these considerations, the following research question arose: what are the main methods adopted by researchers for developing mobile apps in health?As a result, this integrative literature review aims to identify, in the Brazilian and international publications indexed in the databases, the main methods adopted by researchers for developing mobile apps in health.

METHOD
This is an integrative literature review of scientific studies published in 2012 -2016.The stages of this review were based in a previously-established protocol, the aim being to maintain scientific and methodological rigor, namely: 1) elaboration of the research question; 2) definition of the inclusion criteria for studies, and selection of the sample (search or sampling in the literature); 3) representation of the study selected in the format of tables, considering all the characteristics in common (data collection); 4) critical analysis of the studies included, identifying differences and conflicts; 5) interpretation/discussion of the results; 6) presentation of the integrative review in a clear and objective way including the evidence/data found. 18n order to respond to the review's guiding question, a bibliographic search was undertaken in the publications indexed in the following databases: MEDLINE/PubMed, Scopus, CINAHL, Web of Science and SciELO.The MeSH descriptors adopted were: Mobile applications; Medical informatics; Medical informatics applications; Public health informatics; Nursing informatics; Information technology; Telemedicine; and Technology.Due to the variety of terms used in the area of ICT, the decision was also made to include the following keywords in the search: mobile technology; eHealth; mHealth; telehealth; healthcare application; and cybercare.It is emphasized that the Boolean expressions AND and OR were the resources adopted for the study, with the aim of obtaining the largest number of studies possible regarding the topic to be reviewed.
The inclusion criteria for the studies were: original research, literature reviews (systematic, integrative or narrative) and experience reports, published between January 2012 and December 2016; in English, Portuguese or Spanish; available in full, and which clearly expressed the stages/methods (framework) for the development of the mobile app in the area of health -specifically, nursing, medicine and public health.The exclusion criteria considered were: duplicated articles, editorials, conference proceedings, case studies and reflection articles.
It should be emphasized that the selection of studies was undertaken in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology. 19The researchers were divided into 2 groups, each one with 2 members.Group 1 was responsible for carrying out the search in the databases, using the descriptors and keywords selected.In this stage, a total of 1,984 articles were found.After the search, the articles were shared with Group 2, and the researchers proceeded to the remaining stages of the review independently and simultaneously.Figure 1 presents the summary of the results obtained in each stage.The scientific method of reading was adopted for carrying out the data analysis.This method is undertaken in three stages: 1) syncretic visionreading for general recognition, aiming to familiarize oneself with the topic being studied, and selective reading, seeking information about the study objective; 2) analytical vision: reflexive and critical reading of the articles selected, and choice of the main content related to the topic; and 3) synthetic vision: reading for interpretation of the data/results presented in the studies. 20t is highlighted that an instrument was developed for collecting and analyzing the data from the studies included.This instrument contains the following information: authorship, country, language, category of publication, year of publication, journal, study objective, area of health attended by the mobile app, and theoretical framework and method adopted for the development of the mobile app.Grouped by similarity of content, two categories for analysis were formulated: 'Description of the characteristics of the studies', and 'Methods used for developing mobile apps in health'.

Description of the Characteristics of the Studies
The publications selected for the identification of the main methods adopted by the researchers for developing mobile apps in health are described in Table 1.In the MEDLINE/PubMed databases, 12 articles were selected for analysis; in Scopus and Web of Science, three studies in each; in CINAHL, two studies; and in SciELO, one study, totaling 21 articles analyzed.It is emphasized that 86% (18) of the studies were published in international periodicals, and 14% (03) in Brazilian periodicals.
In relation to the studies' country of origin, distribution was as follows: United States of America, 33.5% (07); Brazil, 14% (03); Sweden, 9.5% (02); and Argentina, Australia, Chile, South Korea, Spain, Holland, Indonesia, Italy and Taiwan, 43% (09) -one article for each nation.Regarding the languages of the articles analyzed, 86% (18) were published in English, and 14% (03) in Portuguese; no publication in Spanish was selected for analysis.
Figure 2 presents the distribution of the study selected by year of publication.The articles were classified by their category of publication, as specified by the periodicals: 86% (18) original studies; 9.5% (2) short communiqués and; 4.5% (1) protocol.

Methods used for developing mobile apps in health
Based on the review undertaken, the following descriptors were evidenced: Mobile applications; Medical informatics; Medical informatics applications; Public health informatics; Nursing informatics; Information technology; Telemedicine; and Technology -as well as the keywords 'mobile technology', 'eHealth', 'mHealth', 'telehealth' and 'healthcare application'.These encompassed studies which addressed the main methods for developing mobile apps in the area of health.The keyword 'cybercare', researched in isolation and/or in association with other descriptors, was not found in studies on the central topic of this integrative literature review.
Another method used for developing mobile apps in health is Contextualized Design of Instruction (CDI). 35][46] The method of User Centered Design (UCD) was used to develop apps in the area of Geriatric Nursing 32 and Primary Care (prevention of chronic diseases). 29][49][50] The Systems Development Life Cycle (SDLC) method was adopted in a study which developed an app for consulting on chemotherapy medications, the aim being to assist with the auditing -in nursing -of hospital accounts. 21This method has three approaches for developing systems: the classic lifecycle, the spiral lifecycle, and the prototyping life cycle.

DISCUSSION
A range of methods/procedures involved in SDI has been developed in recent decades, with a view to constructing new technological tools (namely, the mobile apps), which improve the teaching-learning process and the performance of the users in the most widely-differing of contexts.It is observed that the developer/researchers have adopted a series of instructional means/resources in order to achieve their objectives but that, however, regardless of the method chosen and the different stages which make up each one of these, the majority of the methods include the stages of analysis, conception, development, implementation and evaluation. 42,44,46,53It is also noted that, in the instructional and technological design, besides the stages mentioned, one can include the stage of management as a whole. 53n a general way, SDI covers planning, preparation, production and publication of texts, images, sounds and movements, simulations and activities supported by virtual tools made available by ICT. 45 This method has some contextual characteristics and mechanisms, including: individualization of rhythms of learning; it is adaptable to institutional, national, regional and/or local characteristics; it is possible for it to be constantly updated by users' feedback/ opinions; the information can be accessed in real-time and/or remotely; there is the possibility for communication between users and developers/researchers; and it is possible to monitor the individual and collective construction of the users' knowledge.
Different modalities of SDI, which have similar objectives but which differ in their modelling and characteristics, are described in the literature. 44ccordingly, the SDI model is divided according to how it is presented: Fixed Instructional Design (FID), Open Instructional Design (OID), CDI and the Integrative Learning Design Framework (ILDF) model.
The CDI method considers human activity to be central, and seeks a balance between the automation of processes of planning, personalization and contextualization of the teaching/content and the technological tools available.] CDI is made up of the following stages: analysis, design, development, implementation and evaluation.However, while the conventional instructional design models incorporate specific stages in each stage, this model adopts interlacing between its stages throughout the whole of the process of development of the teaching/content.CDI firstly characterizes the users, identifies the needs of the same, and carries out a survey of the limitations, improving and updating these data and information in parallel with the new requirements and participation of the users.][46] In order to exemplify the use of the CDI method, one can cite a study which developed and assessed a mobile app for teaching the measuring of central venous pressure, which was designed for students of nursing (end-users).In this study, the researchers considered all the stages described by CDI, specified as following: I) Analysis: "identifying the needs, the characterizing of the targetpublic, the collection of bibliographic references, the definition of the educational objectives, the definition of the content, the analysis of the technological infrastructure, and the creation of a diagram to guide the construction of the tool"; 35:109 II) Design: "planning and production of the didactic content, the definition of the topics and editing of the modules, the selection of the media, and the design of the interface (layout).:109 The SDI method, also termed the Dick and Carey model, has a systems approach aiming for effective instruction in order to support the teachinglearning process in a way which has been highly successful.54] The UCD method has a broadened approach.This is a general term, characterized by a philosophy and by stages which describe the processes of a project, centered on its creation and the users' involvement in the elaboration of computer-based systems. 47][57] It is highlighted that in the UCD method, the user's participation can vary in intensity. 47At one end of the technological tool's construction, the involve-Barra DCC, Paim SMS, Dal Sasso GTM, Colla GW 8/12 ment can be relatively low -that is, the users may be consulted and observed regarding their needs, and maybe invited to participate in usability tests with the system being developed.At the other end, the users may be involved intensely, actively participating in all of the stages of the project, including the conception of the technological tool itself.
It is worth emphasizing that the International Organization for Standardization (ISO) regulation 13,407: Human-centered design processes for interactive systems describes three design solutions for the UCD method: I) Cooperative design (users and developers/researchers are involved in all stages); II) Participative design (users occasionally participate in the development process) and; III) Contextual design (based in the current context). 55arious mechanisms support the UCD method, including: usability tests, systems usability engineering, heuristic assessment and rapid evaluations/pilot-tests.The rapid evaluations (pilot-tests) are considered important so that the users will be able to give their feedback from the very beginning of the project, as well as to bring the same closer to the developer/researchers and the technological tool itself. 47s an example of the application of the UCD method, one can mention the study which developed a mobile app for monitoring and reacting in order to encourage physical activity in people with chronic illnesses in primary care.In that study, the researchers recruited people from national associations of patients who had Chronic Obstructive Pulmonary Disease (COPD) and diabetes, so that they could participate in the research team.These patients described and reflected on the needs, requirements and restrictions of their conditions as persons with chronic diseases.They also provided the researchers with feedback on their understanding of the issues addressed in the interviews and of the content/documents created for the end-users.Through adopting the UCD method, the researchers managed to gather requirements, which were then translated into technical solutions by the project's engineering team, allowing continuous interaction between all those involved. 29he SDLC method covers the following stages: analysis (collecting data on the needs, and identifying the needs of the institution/user); project (detailed specifications of the project); development (which includes the development or acquisition of the software); implementation (after undergoing evaluative tests); and maintenance (continuously maintaining and updating the system).] The classic lifecycle, also termed the Cascade Model, follows all the stages of the SDLC described above.It has a linear structure, is sequential, and there is an absence of revision for each stage, that is, the system's development takes place in a single direction, it being considered to be an inflexible model/method.] The spiral lifecycle is based in the concept of the user's greatest need.This method develops and hands over the system/technological tool in versions.It is emphasized that each version follows all the stages of the SDLC, with the following exceptions: I) the implementation stage may be adopted in some or all versions; II) the stage of maintenance will be applied only to the last version made available.] The prototyping life cycle addresses the gradual and evolving discovery of the system under development by the users and developers/researchers.] This model has three categories: I) paper prototyping, or computational prototype (demonstrates the human/machine interaction, and the clear understanding of this interaction which is present; II) working prototypes (implements some functions required by the user which may be improved during the development of the system); and III) functional prototype (which allows the user to store data and undertake operations using the data).It is stressed that the prototypes make it possible to visualize uncertain aspects of the system under development.They also make it possible to check and test hypotheses regarding these aspects.] One study developed an app for consulting regarding chemotherapy medications in the auditing of hospital accounts. 21:180 One systematic review study summarized the current knowledge regarding the factors which influence the adoption of mobile health apps (mHealth) by health professionals.Based on the 33 studies selected for analysis, 179 elements recognized as facilitating (54.7%), or as barriers (45.3%) to, the adoption of these technological tools.Among the main elements identified, characterized as individual, organizational and contextual, emphasis is placed on the following: use/purpose and ease of use; design and technical concerns; security, privacy, cost and time; familiarity with the technology; and interaction with colleagues, patients and managers. 16Stress is placed, therefore, on the relevance of the appropriate choice of method for the development of mobile apps in health, as issues related to ease-of-use, design and technical components of the systems may be either factors directly related to the successful adoption of these technological tools, or may be barriers to this adoption.

CONCLUSION
In this integrative review of the literature, the decision was made to analyze only articles which contained the description of the method used, associated with the detailing of their respective stages, for the development of mobile apps in health.However, it is worth emphasizing that as a result of adopting the "Scientific Reading Method", in the syncretic vision phase, for the analysis of the 1,984 articles, it was possible for the authors to evidence that, in a general way, the researchers/developers/ designers used the main methods for developing apps described in the literature, namely: instructional design, systematized design instruction, contextualized instructional design, user-centered design, and systems development lifecycle.
This integrative review made it possible to identify the stages described in each method, highlighting the characteristics of each one and also making it possible to understand that, regardless of the method chosen, the stages must be well defined and appropriately structured, so that the mobile app developed will be useful for the end-user.

Figure 1 -
Figure 1 -Flowchart of the process for selecting the studies, based on the PRISMA methodology.Florianópolis, State of Santa Catarina (SC), 2017