Midwives in Ghana provide the majority of rural primary and maternal healthcare services, but have limited access to data for decision making and knowledge work. Few mobile health (mHealth) applications have been designed for midwives. The study's purpose was to design and test an mHealth application (mClinic) that can improve data access and reduce the reporting burden for midwives at the Millennium Villages Project site in Ghana.
From the design science field, the Information Systems Research Framework guided this study through two research cycles: 1) Relevance, and 2) Design. The first phase of the Relevance Cycle took a user-centered approach to assess the people, organizations, and technology of the midwives’ environment through participant observation, contextual inquiry, and interviews. In the second phase, structured requirements specification was used to categorize the data into goals, system qualities, and constraints. From the categorized data, use cases were developed for patient registration, antenatal care, malaria, family planning, and referrals. Use cases then informed the development of functional requirements. In the Design Cycle, we first used functional requirements for patient registration and malaria to develop the mClinic prototype as part of a coded-in-country initiative. Next, we examined usability of the mClinic prototype by conducting field testing, heuristic evaluation, and usability surveys. Additionally, low-fidelity prototyping was used to determine applicability of the other use cases to the midwives’ environment.
Midwives reported inability to access critical data, high patient loads, and extensive reporting requirements. Low technical self-efficacy and inadequate infrastructure were identified as barriers to implementation. Heuristic evaluation noted issues related to hardware selection, workflow, and security. Midwives ranked the tool as useful in the usability survey; however, ease-of-use rankings were neutral. Interviews indicated this was related to low technical self-efficacy. Applicability checks found support for touch-entry prototypes over those that included lengthy forms or text-entry.
