An investigation into software requirements for systems supporting task-shifted interventions

Background: There is a considerable shortfall in specialized health professionals worldwide to deliver health services and this shortfall is especially pronounced in low-middle income countries. This has led to the implementation of task-shifted interventions, in which specific tasks are moved away from highly qualified health workers to health workers with less training. The World Health Organisation has published recommendations for such interventions, but guidelines for software and systems supporting such interventions are not included. Objective: The objective of this paper is to formulate a number of software requirements for computer systems supporting task-shifted interventions. As the treatment of mental health problems is generally considered to be a task for highly trained health professionls, it poses interesting case studies for task-shifted interventions. Therefore, we will illustrate the use of the identified software requirements in a mobile system created for a task-shifted depression intervention to be provided to older adults in deprived areas of São Paulo, Brazil. Methods: Using a set of recommendations based on the World Health Organisation’s guidance documentation for task-shifted interventions, we identified nine software requirements that aim to support health workers in management and supervision, training, good relationship with other health workers, and community embeddedness of the intervention. These nine software requirements were used to implement a system for the provision of a psychosocial depression intervention with mobile Android interfaces to structure interventions and collect data, and web interfaces for supervision and support of the healthcare workers delivering the intervention. The system was tested in a two-arm pilot study with 33 patients and 11 health workers. Eight of these 11 health workers participated in a ‘usability’ study subsequent to the pilot. Results: Qualitative and quantitative feedback obtained with the System Usability Scale suggest the system was deemed to have a usability of between ‘OK’ and ‘Good’. Nevertheless, some participants’ responses indicate they felt they needed technical assistance to use the system. This was reinforced by answers obtained with ‘Perceived Usefulness’ and ‘Ease of Use’ questionnaires, which indicated some users felt they had issues around correct use of the system and perceived ability to becoming skilful at using the system. Conclusions: Overall, these high-level requirements adequately captured the functionality required to enable the health workers to provide the intervention successfully. Nevertheless, the analysis of results indicated some improvements are required for the system to be useable in a task-shifted intervention. The most important of these are better access to a training environment, access for supervisors to meta-data such as duration of sessions or exercises to identify issues, and a more robust and human-error proof approach to availability of patient data on the mobile devices used during the intervention.