Identifying Generative Mechanisms in a Mobile Health (mHealth) Project in Sierra Leone: A Critical Realist Framework for Retroduction P.J. Wall [0000-0002-5859-4425] , Dave Lewis [0000-0002-3503-4644] & Lucy Hederman [0000-0001-6073-4063] ADAPT Centre, School of Computer Science and Statistics, Trinity College Dublin, Ireland wallp2@tcd.ie dave.lewis@adaptcentre.ie hederman@tcd.ie Abstract. This paper examines an ongoing mobile health (mHealth) initiative in Sierra Leone. The objective is to describe the ontological perspective and method- ological approach used which relies on critical realism and Archer’s morphogenetic approach. Such a critical realist-based approach addresses many of the concerns asso- ciated with the interpretivist and positivist philosophical duopoly that has dominated the ICT4D field for many years. The approach allows the mechanisms that deter- mined what caused the events associated with the mHealth case to be hypothesized, thus making clear why this mHealth case has sustained through many challenges. Keywords: ICT4D, mHealth, critical realism, morphogenetic approach 1 Introduction Over the last decade, information and communications technology (ICT) has become more important and ubiquitous across developing countries. This is particularly true of mobile phones which are pervasive in many developing countries. Mobile tech- nologies are also key to meeting the Sustainable Development Goals, and have thus become an integral part of many projects in a variety of fields such as healthcare, emergency management, and food and water security (Masiero 2018). This makes it easy to understand why the number of mHealth implementations around the world has been increasing dramatically (e.g. Purkayastha, Manda, and Sanner 2013, Cameron, Ramaprasad, and Syn 2017, White et al. 2016, Sundin, Callan, and Mehta 2016). However, while mHealth has the potential to transform health services and to increase access to healthcare (e.g. Hurt et al. 2016, Latif et al. 2017, Agarwal et al. 2016, Be- ratarrechea et al. 2017), the design, implementation and adoption of mHealth in devel- oping countries is beset with wide-ranging challenges and many mHealth projects have been unable to sustain or demonstrate any significant impact at scale (e.g. Chig- ona, Nyemba, and Metfula 2012, Manda and Msosa 2012, Kahn, Yang, and Kahn 2010, Latif et al. 2017, Sundin, Callan, and Mehta 2016). This is because any mHealth implementation in such environments is bound to be shaped by a variety of