Towards Better Supporting Human Aspects in Mobile eHealth Applications
eHealth apps raise awareness of health to ensure they are more accessible during emergencies, usable in home-based disease management and effectively support personalized care through sensing/interaction. However, despite the huge number of these apps available, many do not adopt end-user human aspects in development and deployment – causing them to be ineffective and non-inclusive of diverse users. This project focuses on creating a more integrated approach for eHealth app development that addresses the human aspects of users.
The project aims to determine how different aspects are being addressed by developers now, which ones are missing or poorly handled, and which are the most important for different end user groups. It then explores different ways to improve support for users, developers and other stakeholders of eHealth apps in the form of improved actionable guidelines, work-flow framework, best practice examples and evaluation techniques.
The team initially carried out a preliminary background study to identify the challenges in app development and related future needs. It was uncovered that:
- Metadata based app classification methods are beneficial, but these do not work if the app description is unavailable.
- The Model Driven Development (MDD) technique has the potential to address the challenges of non-trivial mobile app development because it simplifies the process, reduces complexity, increases abstraction level, helps achieve scalable solutions and maximizes cost-effectiveness and productivity.
In the second phase, this projects investigated a Reverse Engineering-based Approach to Classify mobile apps based on the data that exists inside the app (REACT) to overcome the limitations of meta-data based app classification approaches.
To replicate and validate the proposed REACT method, our researchers used an extensive set of Android apps (24,652 apps in total). Although our analysis shows some limitations in REACT procedure and implementations, we identified the root cause of failures and marked future enhancement ideas. Moreover, the REACT approach can identify the eHealth app development patterns, especially when mined with our extracted hundred thousand keywords from medical dictionaries.
In the third phase, the team investigated MDD-based mobile app development schemes via a Systematic Literature Review (SLR) to identify key benefits, limitations, gaps and future research potential in this domain.
The formal search protocol identified a total of 1,042 peer-reviewed academic research papers from four major software engineering databases, i.e., ACM digital library, IEEE Xplore, Springer link, and Science direct. These papers were subsequently filtered, and 55 high-quality relevant studies were selected for analysis, synthesis and reporting.
Our researchers identified the popularity of different applied MDD approaches, supporting tools, artefacts and evaluation techniques. The analysis found that architecture, domain model and code generation are the most crucial purposes in MDD-based app development. Three qualities – productivity, scalability and reliability – can benefit from these modelling strategies. We also summarized the key collective strengths, limitations and gaps from the studies, and made several future recommendations to guide future researchers, developers and stakeholders.
Throughout the research and SLR review process, we realized a much greater need for improved approaches supporting HCIs in eHealth app development, not only in MDD-based approaches but also in overall existing methods. Thus, we reviewed a set of existing eHealth apps, recent literature, current development procedures and evaluation guidelines. Outcomes of these analyses led us to narrow down the topic via a viable gap analysis.
In the third stage, we examined current user and developer engagement to support human aspects in the eHealth app domain. We collected and analysed data from 240 usable survey responses and 25 detailed interviews to better understand how app developers support human aspects in eHealth apps, their corresponding challenges, and impacts on end-users.
Our analysis highlighted that human aspects dominate practical and effective eHealth app usage. However, several factors hinder appropriate aspect identification, addressing, and management. These factors include less research and education, poor practice culture, restrictions from vendors, a shortage of experts, insufficient technical support for developers, and user disengagement from the app development process. We sorted identified aspects, their impacts, and best practice examples into four categories considering mobile app development life-cycle phases, then provided a set of findings to help app developers better address these aspects in eHealth apps.
Finally, we triangulated all these findings and developed a set of new and enhanced actionable guidelines, best practices, and evaluation methods for better supporting human aspects in eHealth apps. In this process, we explored gaps in the current guidelines and standards and integrated our findings from previous research tasks to develop these new and enhanced sets of guidelines with best practice examples. These measures collectively enhance the applicability of the proposed guidelines in practice.
We then gathered feedback from app developers, software engineers, and other relevant stakeholders to evaluate and verify the proposed guidelines' strengths, limitations, practicality, and feasibility in real-world scenarios, aiming to produce more effective eHealth apps. Our validation revealed that over 85% of developers found the guidelines useful and appreciated the focus on real-life examples. They acknowledged that the proposed guidelines promote user-centred design, usability, accessibility, reliability, validity, and diverse user issues in the eHealth app development process with ease, consequently offering more human-centric and effective eHealth apps for their end-users. Some constructive critiques suggested transforming these guidelines into a toolset to facilitate their daily usage in both academia and industry, an initiative we are currently working on.
An interesting future direction for this research project is to explore establishing a collaborative platform for co-design and evaluation of eHealth apps that unites researchers, app developers, healthcare professionals, users, and related stakeholders. This collaboration will facilitate a more comprehensive understanding of user trends, needs, preferences, and concerns. It will also lead to the production of eHealth apps that better incorporate stakeholder requirements, support seamless user-developer knowledge exchange, and facilitate easy sharing of best practices. Ultimately, this will contribute to the advancement of human-centric eHealth app research and development.
- Project Lead: Md. Shamsujjoha
- Principal Supervisor: Prof John Grundy
- Associate Supervisors: Prof Li Li, Dr Hourieh Khalajzadeh, Dr Qinghua Lu (Data61)
Publications from PhD Projects (to date)
- Md. Shamsujjoha, John Grundy, Li Li, Hourieh Khalajzadeh, and Qinghua Lu, “Developing Mobile Applications via Model Driven Development: A Systematic Literature Review”, Information and Software Technology, Eelsevier, vol: 140, no: 106693, pp:1-24, year: Apr-2021, ISSN: 0950-5849, DOI: 10.1016/j.infsof.2021.106693, (Based on PhD research works), [PDF File]
- Md. Shamsujjoha, John Grundy, Li Li, Hourieh Khalajzadeh, and Qinghua Lu, “Human-Centric Issues in eHealth App Development and Usage: A Preliminary Assessment”, 28th IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER), IEEE, pp:506-510, year: 2021, e-ISBN: 978-0-7695-4889-0, DOI: 10.1109/SANER50967.2021.00055 (Based on PhD research works), [PDF File]
- Md. Shamsujjoha, John Grundy, Li Li, Hourieh Khalajzadeh, and Qinghua Lu, “Checking App Behavior Against App Descriptions: What If There are No App Descriptions?”, 29th IEEE/ACM International Conference on Program Comprehension (ICPC), IEEE/ACM, ISSN-2643-7171, pp:422-432, year: 2021, DOI: 10.1109/ICPC52881.2021.00050 (Based on PhD research works) [PDF File]